* [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
@ 2016-10-30 21:23 Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 01/47] virtio/migration: Add VMStateDescription to VirtioDeviceClass Michael S. Tsirkin
` (49 more replies)
0 siblings, 50 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
are available in the git repository at:
git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
----------------------------------------------------------------
virtio, pc: fixes and features
nvdimm hotplug support
virtio migration and ioeventfd rework
virtio crypto device
ipmi fixes
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
----------------------------------------------------------------
Corey Minyard (5):
ipmi: Remove hotplug from IPMI BMCs
ipmi_bmc_sim: Remove an unnecessary mutex
ipmi: Implement shutdown via ACPI overtemp
ipmi: Add graceful shutdown handling to the external BMC
acpi/ipmi: Initialize the fwinfo before fetching it
Cédric Le Goater (1):
ipmi: chassis poweroff should use qemu_system_shutdown_request()
Daniel P. Berrange (1):
ipmi: fix build config variable name for ipmi_bmc_extern.o
Dr. David Alan Gilbert (2):
virtio/migration: Add VMStateDescription to VirtioDeviceClass
virtio/migration: Migrate balloon to VMState
Gonglei (12):
cryptodev: introduce cryptodev backend interface
cryptodev: add symmetric algorithm operation stuff
virtio-crypto: introduce virtio_crypto.h
cryptodev: introduce a new cryptodev backend
virtio-crypto: add virtio crypto device emulation
virtio-crypto-pci: add virtio crypto pci support
virtio-crypto: set capacity of algorithms supported
virtio-crypto: add control queue handler
virtio-crypto: add data queue processing handler
cryptodev: introduce an unified wrapper for crypto operation
virtio-crypto: using bh to handle dataq's requests
virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
Haozhong Zhang (1):
acpi: fix assert failure caused by commit 35c5a52d
Paolo Bonzini (13):
virtio: disable ioeventfd as early as possible
virtio: move ioeventfd_disabled flag to VirtioBusState
virtio: move ioeventfd_started flag to VirtioBusState
virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
virtio: introduce virtio_device_ioeventfd_enabled
virtio-blk: always use dataplane path if ioeventfd is active
virtio-scsi: always use dataplane path if ioeventfd is active
Revert "virtio: Introduce virtio_add_queue_aio"
virtio: remove set_handler argument from set_host_notifier_internal
virtio: remove ioeventfd_disabled altogether
virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
virtio: inline virtio_queue_set_host_notifier_fd_handler
virtio: inline set_host_notifier_internal
Xiao Guangrong (12):
acpi nvdimm: fix wrong buffer size returned by DSM method
acpi nvdimm: fix OperationRegion definition
acpi nvdimm: fix device physical address base
acpi nvdimm: fix ARG3 conflict
acpi nvdimm: fix Arg6 usage
nvdimm acpi: compile nvdimm acpi code arch-independently
acpi nvdimm: rename result_size to dsm_out_buf_siz
nvdimm acpi: use common macros instead of magic names
nvdimm acpi: prebuild nvdimm devices for available slots
nvdimm acpi: introduce fit buffer
nvdimm acpi: introduce _FIT
pc: memhp: enable nvdimm device hotplug
hw/block/dataplane/virtio-blk.h | 6 +-
hw/s390x/virtio-ccw.h | 2 -
hw/virtio/virtio-pci.h | 17 +-
include/hw/acpi/acpi_dev_interface.h | 1 +
include/hw/hotplug.h | 10 +
include/hw/mem/nvdimm.h | 27 +-
include/hw/virtio/virtio-bus.h | 27 +-
include/hw/virtio/virtio-crypto.h | 101 +++
include/hw/virtio/virtio-scsi.h | 6 +-
include/hw/virtio/virtio.h | 15 +-
include/standard-headers/linux/virtio_crypto.h | 429 ++++++++++++
include/standard-headers/linux/virtio_ids.h | 2 +-
include/sysemu/cryptodev.h | 298 ++++++++
backends/cryptodev-builtin.c | 361 ++++++++++
backends/cryptodev.c | 245 +++++++
hw/acpi/ipmi.c | 1 +
hw/acpi/memory_hotplug.c | 31 +-
hw/acpi/nvdimm.c | 468 ++++++++++---
hw/block/dataplane/virtio-blk.c | 73 +-
hw/block/virtio-blk.c | 15 +-
hw/core/hotplug.c | 11 +
hw/core/qdev.c | 20 +-
hw/i386/acpi-build.c | 9 +-
hw/i386/pc.c | 31 +
hw/ipmi/ipmi.c | 10 +-
hw/ipmi/ipmi_bmc_extern.c | 12 +-
hw/ipmi/ipmi_bmc_sim.c | 7 +-
hw/mem/nvdimm.c | 4 -
hw/s390x/virtio-ccw.c | 44 +-
hw/scsi/virtio-scsi-dataplane.c | 56 +-
hw/scsi/virtio-scsi.c | 24 +-
hw/virtio/vhost.c | 5 +-
hw/virtio/virtio-balloon.c | 31 +-
hw/virtio/virtio-bus.c | 154 ++---
hw/virtio/virtio-crypto-pci.c | 77 +++
hw/virtio/virtio-crypto.c | 898 +++++++++++++++++++++++++
hw/virtio/virtio-mmio.c | 35 +-
hw/virtio/virtio-pci.c | 40 +-
hw/virtio/virtio.c | 153 +++--
tests/ipmi-bt-test.c | 2 +-
MAINTAINERS | 13 +
backends/Makefile.objs | 3 +
docs/specs/acpi_mem_hotplug.txt | 3 +
docs/specs/acpi_nvdimm.txt | 58 +-
hw/acpi/Makefile.objs | 2 +-
hw/ipmi/Makefile.objs | 2 +-
hw/virtio/Makefile.objs | 2 +
qemu-options.hx | 18 +
48 files changed, 3352 insertions(+), 507 deletions(-)
create mode 100644 include/hw/virtio/virtio-crypto.h
create mode 100644 include/standard-headers/linux/virtio_crypto.h
create mode 100644 include/sysemu/cryptodev.h
create mode 100644 backends/cryptodev-builtin.c
create mode 100644 backends/cryptodev.c
create mode 100644 hw/virtio/virtio-crypto-pci.c
create mode 100644 hw/virtio/virtio-crypto.c
^ permalink raw reply [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 01/47] virtio/migration: Add VMStateDescription to VirtioDeviceClass
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 02/47] virtio/migration: Migrate balloon to VMState Michael S. Tsirkin
` (48 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Dr. David Alan Gilbert
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Provide a vmsd pointer for VirtIO devices to use instead of the
load/save methods.
We'll eventually kill off the load/save methods.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio.h | 4 ++++
hw/virtio/virtio.c | 14 ++++++++++++++
2 files changed, 18 insertions(+)
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index b913aac..52d4b55 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -125,8 +125,12 @@ typedef struct VirtioDeviceClass {
* must mask in frontend instead.
*/
void (*guest_notifier_mask)(VirtIODevice *vdev, int n, bool mask);
+ /* Saving and loading of a device; trying to deprecate save/load
+ * use vmsd for new devices.
+ */
void (*save)(VirtIODevice *vdev, QEMUFile *f);
int (*load)(VirtIODevice *vdev, QEMUFile *f, int version_id);
+ const VMStateDescription *vmsd;
} VirtioDeviceClass;
void virtio_instance_init_common(Object *proxy_obj, void *data,
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index d48d1a9..3e318e4 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -1635,6 +1635,10 @@ void virtio_save(VirtIODevice *vdev, QEMUFile *f)
vdc->save(vdev, f);
}
+ if (vdc->vmsd) {
+ vmstate_save_state(f, vdc->vmsd, vdev, NULL);
+ }
+
/* Subsections */
vmstate_save_state(f, &vmstate_virtio, vdev, NULL);
}
@@ -1781,6 +1785,13 @@ int virtio_load(VirtIODevice *vdev, QEMUFile *f, int version_id)
}
}
+ if (vdc->vmsd) {
+ ret = vmstate_load_state(f, vdc->vmsd, vdev, version_id);
+ if (ret) {
+ return ret;
+ }
+ }
+
/* Subsections */
ret = vmstate_load_state(f, &vmstate_virtio, vdev, 1);
if (ret) {
@@ -2118,6 +2129,9 @@ static void virtio_device_realize(DeviceState *dev, Error **errp)
VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(dev);
Error *err = NULL;
+ /* Devices should either use vmsd or the load/save methods */
+ assert(!vdc->vmsd || !vdc->load);
+
if (vdc->realize != NULL) {
vdc->realize(dev, &err);
if (err != NULL) {
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 02/47] virtio/migration: Migrate balloon to VMState
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 01/47] virtio/migration: Add VMStateDescription to VirtioDeviceClass Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 03/47] virtio: disable ioeventfd as early as possible Michael S. Tsirkin
` (47 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Dr. David Alan Gilbert
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Replace the load/save with a vmsd.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/virtio-balloon.c | 31 +++++++++++++++----------------
1 file changed, 15 insertions(+), 16 deletions(-)
diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
index 1d77028..cfba053 100644
--- a/hw/virtio/virtio-balloon.c
+++ b/hw/virtio/virtio-balloon.c
@@ -394,21 +394,9 @@ static void virtio_balloon_to_target(void *opaque, ram_addr_t target)
trace_virtio_balloon_to_target(target, dev->num_pages);
}
-static void virtio_balloon_save_device(VirtIODevice *vdev, QEMUFile *f)
+static int virtio_balloon_post_load_device(void *opaque, int version_id)
{
- VirtIOBalloon *s = VIRTIO_BALLOON(vdev);
-
- qemu_put_be32(f, s->num_pages);
- qemu_put_be32(f, s->actual);
-}
-
-static int virtio_balloon_load_device(VirtIODevice *vdev, QEMUFile *f,
- int version_id)
-{
- VirtIOBalloon *s = VIRTIO_BALLOON(vdev);
-
- s->num_pages = qemu_get_be32(f);
- s->actual = qemu_get_be32(f);
+ VirtIOBalloon *s = VIRTIO_BALLOON(opaque);
if (balloon_stats_enabled(s)) {
balloon_stats_change_timer(s, s->stats_poll_interval);
@@ -416,6 +404,18 @@ static int virtio_balloon_load_device(VirtIODevice *vdev, QEMUFile *f,
return 0;
}
+static const VMStateDescription vmstate_virtio_balloon_device = {
+ .name = "virtio-balloon-device",
+ .version_id = 1,
+ .minimum_version_id = 1,
+ .post_load = virtio_balloon_post_load_device,
+ .fields = (VMStateField[]) {
+ VMSTATE_UINT32(num_pages, VirtIOBalloon),
+ VMSTATE_UINT32(actual, VirtIOBalloon),
+ VMSTATE_END_OF_LIST()
+ },
+};
+
static void virtio_balloon_device_realize(DeviceState *dev, Error **errp)
{
VirtIODevice *vdev = VIRTIO_DEVICE(dev);
@@ -517,9 +517,8 @@ static void virtio_balloon_class_init(ObjectClass *klass, void *data)
vdc->get_config = virtio_balloon_get_config;
vdc->set_config = virtio_balloon_set_config;
vdc->get_features = virtio_balloon_get_features;
- vdc->save = virtio_balloon_save_device;
- vdc->load = virtio_balloon_load_device;
vdc->set_status = virtio_balloon_set_status;
+ vdc->vmsd = &vmstate_virtio_balloon_device;
}
static const TypeInfo virtio_balloon_info = {
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 03/47] virtio: disable ioeventfd as early as possible
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 01/47] virtio/migration: Add VMStateDescription to VirtioDeviceClass Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 02/47] virtio/migration: Migrate balloon to VMState Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 04/47] virtio: move ioeventfd_disabled flag to VirtioBusState Michael S. Tsirkin
` (46 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Paolo Bonzini, Cornelia Huck,
Christian Borntraeger, Richard Henderson, Alexander Graf
From: Paolo Bonzini <pbonzini@redhat.com>
Avoid "tricking" virtio-blk-dataplane into thinking that ioeventfd will be
available when it is not. This bug has always been there, but it will break
TCG+ioeventfd=on once the dataplane code will be always used when ioeventfd=on.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/s390x/virtio-ccw.c | 8 ++++----
hw/virtio/virtio-pci.c | 8 ++++----
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
index ee136ab..31304fe 100644
--- a/hw/s390x/virtio-ccw.c
+++ b/hw/s390x/virtio-ccw.c
@@ -709,6 +709,10 @@ static void virtio_ccw_device_realize(VirtioCcwDevice *dev, Error **errp)
sch->cssid, sch->ssid, sch->schid, sch->devno,
ccw_dev->bus_id.valid ? "user-configured" : "auto-configured");
+ if (!kvm_eventfds_enabled()) {
+ dev->flags &= ~VIRTIO_CCW_FLAG_USE_IOEVENTFD;
+ }
+
if (k->realize) {
k->realize(dev, &err);
}
@@ -1311,10 +1315,6 @@ static void virtio_ccw_device_plugged(DeviceState *d, Error **errp)
return;
}
- if (!kvm_eventfds_enabled()) {
- dev->flags &= ~VIRTIO_CCW_FLAG_USE_IOEVENTFD;
- }
-
sch->id.cu_model = virtio_bus_get_vdev_id(&dev->bus);
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index 06831de..29896d8 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -1719,10 +1719,6 @@ static void virtio_pci_device_plugged(DeviceState *d, Error **errp)
pci_register_bar(&proxy->pci_dev, proxy->legacy_io_bar_idx,
PCI_BASE_ADDRESS_SPACE_IO, &proxy->bar);
}
-
- if (!kvm_has_many_ioeventfds()) {
- proxy->flags &= ~VIRTIO_PCI_FLAG_USE_IOEVENTFD;
- }
}
static void virtio_pci_device_unplugged(DeviceState *d)
@@ -1751,6 +1747,10 @@ static void virtio_pci_realize(PCIDevice *pci_dev, Error **errp)
bool pcie_port = pci_bus_is_express(pci_dev->bus) &&
!pci_bus_is_root(pci_dev->bus);
+ if (!kvm_has_many_ioeventfds()) {
+ proxy->flags &= ~VIRTIO_PCI_FLAG_USE_IOEVENTFD;
+ }
+
/*
* virtio pci bar layout used by default.
* subclasses can re-arrange things if needed.
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 04/47] virtio: move ioeventfd_disabled flag to VirtioBusState
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (2 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 03/47] virtio: disable ioeventfd as early as possible Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 05/47] virtio: move ioeventfd_started " Michael S. Tsirkin
` (45 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck,
Christian Borntraeger, Alexander Graf, Richard Henderson
From: Paolo Bonzini <pbonzini@redhat.com>
This simplifies the code and removes the ioeventfd_set_disabled
callback.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/s390x/virtio-ccw.h | 1 -
hw/virtio/virtio-pci.h | 1 -
include/hw/virtio/virtio-bus.h | 8 ++++++--
hw/s390x/virtio-ccw.c | 11 +----------
hw/virtio/virtio-bus.c | 4 ++--
hw/virtio/virtio-mmio.c | 13 +------------
hw/virtio/virtio-pci.c | 11 +----------
7 files changed, 11 insertions(+), 38 deletions(-)
diff --git a/hw/s390x/virtio-ccw.h b/hw/s390x/virtio-ccw.h
index 565094e..327e75a 100644
--- a/hw/s390x/virtio-ccw.h
+++ b/hw/s390x/virtio-ccw.h
@@ -87,7 +87,6 @@ struct VirtioCcwDevice {
uint32_t max_rev;
VirtioBusState bus;
bool ioeventfd_started;
- bool ioeventfd_disabled;
uint32_t flags;
uint8_t thinint_isc;
AdapterRoutes routes;
diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h
index b4edea6..b13d28d 100644
--- a/hw/virtio/virtio-pci.h
+++ b/hw/virtio/virtio-pci.h
@@ -158,7 +158,6 @@ struct VirtIOPCIProxy {
uint32_t guest_features[2];
VirtIOPCIQueue vqs[VIRTIO_QUEUE_MAX];
- bool ioeventfd_disabled;
bool ioeventfd_started;
VirtIOIRQFD *vector_irqfd;
int nvqs_with_notifiers;
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index 2e4b67e..4aabec4 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -83,8 +83,6 @@ typedef struct VirtioBusClass {
void (*ioeventfd_set_started)(DeviceState *d, bool started, bool err);
/* Returns true if the ioeventfd has been disabled for the device. */
bool (*ioeventfd_disabled)(DeviceState *d);
- /* Sets the 'ioeventfd disabled' state for the device. */
- void (*ioeventfd_set_disabled)(DeviceState *d, bool disabled);
/*
* Assigns/deassigns the ioeventfd backing for the transport on
* the device for queue number n. Returns an error value on
@@ -102,6 +100,12 @@ typedef struct VirtioBusClass {
struct VirtioBusState {
BusState parent_obj;
+
+ /*
+ * Set if the default ioeventfd handlers are disabled by vhost
+ * or dataplane.
+ */
+ bool ioeventfd_disabled;
};
void virtio_bus_device_plugged(VirtIODevice *vdev, Error **errp);
diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
index 31304fe..672e6c4 100644
--- a/hw/s390x/virtio-ccw.c
+++ b/hw/s390x/virtio-ccw.c
@@ -82,15 +82,7 @@ static bool virtio_ccw_ioeventfd_disabled(DeviceState *d)
{
VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
- return dev->ioeventfd_disabled ||
- !(dev->flags & VIRTIO_CCW_FLAG_USE_IOEVENTFD);
-}
-
-static void virtio_ccw_ioeventfd_set_disabled(DeviceState *d, bool disabled)
-{
- VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
-
- dev->ioeventfd_disabled = disabled;
+ return !(dev->flags & VIRTIO_CCW_FLAG_USE_IOEVENTFD);
}
static int virtio_ccw_ioeventfd_assign(DeviceState *d, EventNotifier *notifier,
@@ -1619,7 +1611,6 @@ static void virtio_ccw_bus_class_init(ObjectClass *klass, void *data)
k->ioeventfd_started = virtio_ccw_ioeventfd_started;
k->ioeventfd_set_started = virtio_ccw_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_ccw_ioeventfd_disabled;
- k->ioeventfd_set_disabled = virtio_ccw_ioeventfd_set_disabled;
k->ioeventfd_assign = virtio_ccw_ioeventfd_assign;
}
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index 11f65bd..3607c29 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -195,7 +195,7 @@ void virtio_bus_start_ioeventfd(VirtioBusState *bus)
if (!k->ioeventfd_started || k->ioeventfd_started(proxy)) {
return;
}
- if (k->ioeventfd_disabled(proxy)) {
+ if (bus->ioeventfd_disabled || k->ioeventfd_disabled(proxy)) {
return;
}
vdev = virtio_bus_get_device(bus);
@@ -257,7 +257,7 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
if (!k->ioeventfd_started) {
return -ENOSYS;
}
- k->ioeventfd_set_disabled(proxy, assign);
+ bus->ioeventfd_disabled = assign;
if (assign) {
/*
* Stop using the generic ioeventfd, we are doing eventfd handling
diff --git a/hw/virtio/virtio-mmio.c b/hw/virtio/virtio-mmio.c
index 13798b3..12a9e79 100644
--- a/hw/virtio/virtio-mmio.c
+++ b/hw/virtio/virtio-mmio.c
@@ -89,7 +89,6 @@ typedef struct {
uint32_t guest_page_shift;
/* virtio-bus */
VirtioBusState bus;
- bool ioeventfd_disabled;
bool ioeventfd_started;
bool format_transport_address;
} VirtIOMMIOProxy;
@@ -111,16 +110,7 @@ static void virtio_mmio_ioeventfd_set_started(DeviceState *d, bool started,
static bool virtio_mmio_ioeventfd_disabled(DeviceState *d)
{
- VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
-
- return !kvm_eventfds_enabled() || proxy->ioeventfd_disabled;
-}
-
-static void virtio_mmio_ioeventfd_set_disabled(DeviceState *d, bool disabled)
-{
- VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
-
- proxy->ioeventfd_disabled = disabled;
+ return !kvm_eventfds_enabled();
}
static int virtio_mmio_ioeventfd_assign(DeviceState *d,
@@ -560,7 +550,6 @@ static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
k->ioeventfd_started = virtio_mmio_ioeventfd_started;
k->ioeventfd_set_started = virtio_mmio_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_mmio_ioeventfd_disabled;
- k->ioeventfd_set_disabled = virtio_mmio_ioeventfd_set_disabled;
k->ioeventfd_assign = virtio_mmio_ioeventfd_assign;
k->has_variable_vring_alignment = true;
bus_class->max_dev = 1;
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index 29896d8..a8cf1e5 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -281,15 +281,7 @@ static bool virtio_pci_ioeventfd_disabled(DeviceState *d)
{
VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
- return proxy->ioeventfd_disabled ||
- !(proxy->flags & VIRTIO_PCI_FLAG_USE_IOEVENTFD);
-}
-
-static void virtio_pci_ioeventfd_set_disabled(DeviceState *d, bool disabled)
-{
- VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
-
- proxy->ioeventfd_disabled = disabled;
+ return !(proxy->flags & VIRTIO_PCI_FLAG_USE_IOEVENTFD);
}
#define QEMU_VIRTIO_PCI_QUEUE_MEM_MULT 0x1000
@@ -2542,7 +2534,6 @@ static void virtio_pci_bus_class_init(ObjectClass *klass, void *data)
k->ioeventfd_started = virtio_pci_ioeventfd_started;
k->ioeventfd_set_started = virtio_pci_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_pci_ioeventfd_disabled;
- k->ioeventfd_set_disabled = virtio_pci_ioeventfd_set_disabled;
k->ioeventfd_assign = virtio_pci_ioeventfd_assign;
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 05/47] virtio: move ioeventfd_started flag to VirtioBusState
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (3 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 04/47] virtio: move ioeventfd_disabled flag to VirtioBusState Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 06/47] virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass Michael S. Tsirkin
` (44 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck,
Kevin Wolf, Max Reitz, Christian Borntraeger, Richard Henderson,
Alexander Graf, qemu-block
From: Paolo Bonzini <pbonzini@redhat.com>
This simplifies the code and removes the ioeventfd_started
and ioeventfd_set_started callback. The only difference is
in how virtio-ccw handles an error---it doesn't disable
ioeventfd forever anymore. It was the only backend to do
so, and if desired this behavior should be implemented in
virtio-bus.c.
Instead of ioeventfd_started, the ioeventfd_assign callback now
determines whether the virtio bus supports host notifiers.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/s390x/virtio-ccw.h | 1 -
hw/virtio/virtio-pci.h | 1 -
include/hw/virtio/virtio-bus.h | 17 +++++++----------
hw/block/dataplane/virtio-blk.c | 2 +-
hw/s390x/virtio-ccw.c | 21 ---------------------
hw/scsi/virtio-scsi-dataplane.c | 2 +-
hw/virtio/vhost.c | 2 +-
hw/virtio/virtio-bus.c | 14 ++++++--------
hw/virtio/virtio-mmio.c | 18 ------------------
hw/virtio/virtio-pci.c | 17 -----------------
10 files changed, 16 insertions(+), 79 deletions(-)
diff --git a/hw/s390x/virtio-ccw.h b/hw/s390x/virtio-ccw.h
index 327e75a..77d10f1 100644
--- a/hw/s390x/virtio-ccw.h
+++ b/hw/s390x/virtio-ccw.h
@@ -86,7 +86,6 @@ struct VirtioCcwDevice {
int revision;
uint32_t max_rev;
VirtioBusState bus;
- bool ioeventfd_started;
uint32_t flags;
uint8_t thinint_isc;
AdapterRoutes routes;
diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h
index b13d28d..4d206c8 100644
--- a/hw/virtio/virtio-pci.h
+++ b/hw/virtio/virtio-pci.h
@@ -158,7 +158,6 @@ struct VirtIOPCIProxy {
uint32_t guest_features[2];
VirtIOPCIQueue vqs[VIRTIO_QUEUE_MAX];
- bool ioeventfd_started;
VirtIOIRQFD *vector_irqfd;
int nvqs_with_notifiers;
VirtioBusState bus;
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index 4aabec4..f74ef1e 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -70,17 +70,9 @@ typedef struct VirtioBusClass {
void (*device_unplugged)(DeviceState *d);
int (*query_nvectors)(DeviceState *d);
/*
- * ioeventfd handling: if the transport implements ioeventfd_started,
- * it must implement the other ioeventfd callbacks as well
+ * ioeventfd handling: if the transport implements ioeventfd_assign,
+ * it must implement ioeventfd_disabled as well.
*/
- /* Returns true if the ioeventfd has been started for the device. */
- bool (*ioeventfd_started)(DeviceState *d);
- /*
- * Sets the 'ioeventfd started' state after the ioeventfd has been
- * started/stopped for the device. err signifies whether an error
- * had occurred.
- */
- void (*ioeventfd_set_started)(DeviceState *d, bool started, bool err);
/* Returns true if the ioeventfd has been disabled for the device. */
bool (*ioeventfd_disabled)(DeviceState *d);
/*
@@ -106,6 +98,11 @@ struct VirtioBusState {
* or dataplane.
*/
bool ioeventfd_disabled;
+
+ /*
+ * Set if ioeventfd has been started.
+ */
+ bool ioeventfd_started;
};
void virtio_bus_device_plugged(VirtIODevice *vdev, Error **errp);
diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c
index 704a763..9b268f4 100644
--- a/hw/block/dataplane/virtio-blk.c
+++ b/hw/block/dataplane/virtio-blk.c
@@ -93,7 +93,7 @@ void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *conf,
}
/* Don't try if transport does not support notifiers. */
- if (!k->set_guest_notifiers || !k->ioeventfd_started) {
+ if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
error_setg(errp,
"device is incompatible with dataplane "
"(transport does not support notifiers)");
diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
index 672e6c4..bf5670c 100644
--- a/hw/s390x/virtio-ccw.c
+++ b/hw/s390x/virtio-ccw.c
@@ -59,25 +59,6 @@ static void virtio_ccw_stop_ioeventfd(VirtioCcwDevice *dev)
virtio_bus_stop_ioeventfd(&dev->bus);
}
-static bool virtio_ccw_ioeventfd_started(DeviceState *d)
-{
- VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
-
- return dev->ioeventfd_started;
-}
-
-static void virtio_ccw_ioeventfd_set_started(DeviceState *d, bool started,
- bool err)
-{
- VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
-
- dev->ioeventfd_started = started;
- if (err) {
- /* Disable ioeventfd for this device. */
- dev->flags &= ~VIRTIO_CCW_FLAG_USE_IOEVENTFD;
- }
-}
-
static bool virtio_ccw_ioeventfd_disabled(DeviceState *d)
{
VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
@@ -1608,8 +1589,6 @@ static void virtio_ccw_bus_class_init(ObjectClass *klass, void *data)
k->pre_plugged = virtio_ccw_pre_plugged;
k->device_plugged = virtio_ccw_device_plugged;
k->device_unplugged = virtio_ccw_device_unplugged;
- k->ioeventfd_started = virtio_ccw_ioeventfd_started;
- k->ioeventfd_set_started = virtio_ccw_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_ccw_ioeventfd_disabled;
k->ioeventfd_assign = virtio_ccw_ioeventfd_assign;
}
diff --git a/hw/scsi/virtio-scsi-dataplane.c b/hw/scsi/virtio-scsi-dataplane.c
index b173b94..f537b4e 100644
--- a/hw/scsi/virtio-scsi-dataplane.c
+++ b/hw/scsi/virtio-scsi-dataplane.c
@@ -31,7 +31,7 @@ void virtio_scsi_set_iothread(VirtIOSCSI *s, IOThread *iothread)
s->ctx = iothread_get_aio_context(vs->conf.iothread);
/* Don't try if transport does not support notifiers. */
- if (!k->set_guest_notifiers || !k->ioeventfd_started) {
+ if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
fprintf(stderr, "virtio-scsi: Failed to set iothread "
"(transport does not support notifiers)");
exit(1);
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index bd051ab..501a5f4 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -1190,7 +1190,7 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
int i, r, e;
- if (!k->ioeventfd_started) {
+ if (!k->ioeventfd_assign) {
error_report("binding does not support host notifiers");
r = -ENOSYS;
goto fail;
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index 3607c29..97cdb11 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -192,10 +192,10 @@ void virtio_bus_start_ioeventfd(VirtioBusState *bus)
VirtIODevice *vdev;
int n, r;
- if (!k->ioeventfd_started || k->ioeventfd_started(proxy)) {
+ if (!k->ioeventfd_assign || k->ioeventfd_disabled(proxy)) {
return;
}
- if (bus->ioeventfd_disabled || k->ioeventfd_disabled(proxy)) {
+ if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
return;
}
vdev = virtio_bus_get_device(bus);
@@ -208,7 +208,7 @@ void virtio_bus_start_ioeventfd(VirtioBusState *bus)
goto assign_error;
}
}
- k->ioeventfd_set_started(proxy, true, false);
+ bus->ioeventfd_started = true;
return;
assign_error:
@@ -220,18 +220,16 @@ assign_error:
r = set_host_notifier_internal(proxy, bus, n, false, false);
assert(r >= 0);
}
- k->ioeventfd_set_started(proxy, false, true);
error_report("%s: failed. Fallback to userspace (slower).", __func__);
}
void virtio_bus_stop_ioeventfd(VirtioBusState *bus)
{
- VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
DeviceState *proxy = DEVICE(BUS(bus)->parent);
VirtIODevice *vdev;
int n, r;
- if (!k->ioeventfd_started || !k->ioeventfd_started(proxy)) {
+ if (!bus->ioeventfd_started) {
return;
}
vdev = virtio_bus_get_device(bus);
@@ -242,7 +240,7 @@ void virtio_bus_stop_ioeventfd(VirtioBusState *bus)
r = set_host_notifier_internal(proxy, bus, n, false, false);
assert(r >= 0);
}
- k->ioeventfd_set_started(proxy, false, false);
+ bus->ioeventfd_started = false;
}
/*
@@ -254,7 +252,7 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
DeviceState *proxy = DEVICE(BUS(bus)->parent);
- if (!k->ioeventfd_started) {
+ if (!k->ioeventfd_assign) {
return -ENOSYS;
}
bus->ioeventfd_disabled = assign;
diff --git a/hw/virtio/virtio-mmio.c b/hw/virtio/virtio-mmio.c
index 12a9e79..04a9655 100644
--- a/hw/virtio/virtio-mmio.c
+++ b/hw/virtio/virtio-mmio.c
@@ -89,25 +89,9 @@ typedef struct {
uint32_t guest_page_shift;
/* virtio-bus */
VirtioBusState bus;
- bool ioeventfd_started;
bool format_transport_address;
} VirtIOMMIOProxy;
-static bool virtio_mmio_ioeventfd_started(DeviceState *d)
-{
- VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
-
- return proxy->ioeventfd_started;
-}
-
-static void virtio_mmio_ioeventfd_set_started(DeviceState *d, bool started,
- bool err)
-{
- VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
-
- proxy->ioeventfd_started = started;
-}
-
static bool virtio_mmio_ioeventfd_disabled(DeviceState *d)
{
return !kvm_eventfds_enabled();
@@ -547,8 +531,6 @@ static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
k->save_config = virtio_mmio_save_config;
k->load_config = virtio_mmio_load_config;
k->set_guest_notifiers = virtio_mmio_set_guest_notifiers;
- k->ioeventfd_started = virtio_mmio_ioeventfd_started;
- k->ioeventfd_set_started = virtio_mmio_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_mmio_ioeventfd_disabled;
k->ioeventfd_assign = virtio_mmio_ioeventfd_assign;
k->has_variable_vring_alignment = true;
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index a8cf1e5..f00fc25 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -262,21 +262,6 @@ static int virtio_pci_load_queue(DeviceState *d, int n, QEMUFile *f)
return 0;
}
-static bool virtio_pci_ioeventfd_started(DeviceState *d)
-{
- VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
-
- return proxy->ioeventfd_started;
-}
-
-static void virtio_pci_ioeventfd_set_started(DeviceState *d, bool started,
- bool err)
-{
- VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
-
- proxy->ioeventfd_started = started;
-}
-
static bool virtio_pci_ioeventfd_disabled(DeviceState *d)
{
VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
@@ -2531,8 +2516,6 @@ static void virtio_pci_bus_class_init(ObjectClass *klass, void *data)
k->device_plugged = virtio_pci_device_plugged;
k->device_unplugged = virtio_pci_device_unplugged;
k->query_nvectors = virtio_pci_query_nvectors;
- k->ioeventfd_started = virtio_pci_ioeventfd_started;
- k->ioeventfd_set_started = virtio_pci_ioeventfd_set_started;
k->ioeventfd_disabled = virtio_pci_ioeventfd_disabled;
k->ioeventfd_assign = virtio_pci_ioeventfd_assign;
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 06/47] virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (4 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 05/47] virtio: move ioeventfd_started " Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 07/47] virtio: introduce virtio_device_ioeventfd_enabled Michael S. Tsirkin
` (43 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
Allow customization of the start and stop of ioeventfd. This will
allow direct start of dataplane without passing through the default
ioeventfd handlers, which in turn allows using the dataplane logic
instead of virtio_add_queue_aio. It will also enable some code
simplification, because the sole entry point to ioeventfd setup
will be virtio_bus_set_host_notifier.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-bus.h | 7 ++++-
include/hw/virtio/virtio.h | 4 +++
hw/virtio/virtio-bus.c | 54 +++++++++++------------------------
hw/virtio/virtio.c | 64 ++++++++++++++++++++++++++++++++++++++++++
4 files changed, 91 insertions(+), 38 deletions(-)
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index f74ef1e..aa326e1 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -132,10 +132,15 @@ static inline VirtIODevice *virtio_bus_get_device(VirtioBusState *bus)
}
/* Start the ioeventfd. */
-void virtio_bus_start_ioeventfd(VirtioBusState *bus);
+int virtio_bus_start_ioeventfd(VirtioBusState *bus);
/* Stop the ioeventfd. */
void virtio_bus_stop_ioeventfd(VirtioBusState *bus);
/* Switch from/to the generic ioeventfd handler */
int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign);
+/* This is temporary. It is only needed because virtio_bus_set_host_notifier
+ * sets ioeventfd_disabled but we will shortly get rid of it. */
+int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
+ int n, bool assign, bool set_handler);
+
#endif /* VIRTIO_BUS_H */
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 52d4b55..27e5151 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -125,6 +125,8 @@ typedef struct VirtioDeviceClass {
* must mask in frontend instead.
*/
void (*guest_notifier_mask)(VirtIODevice *vdev, int n, bool mask);
+ int (*start_ioeventfd)(VirtIODevice *vdev);
+ void (*stop_ioeventfd)(VirtIODevice *vdev);
/* Saving and loading of a device; trying to deprecate save/load
* use vmsd for new devices.
*/
@@ -269,6 +271,8 @@ uint16_t virtio_get_queue_index(VirtQueue *vq);
EventNotifier *virtio_queue_get_guest_notifier(VirtQueue *vq);
void virtio_queue_set_guest_notifier_fd_handler(VirtQueue *vq, bool assign,
bool with_irqfd);
+int virtio_device_start_ioeventfd(VirtIODevice *vdev);
+void virtio_device_stop_ioeventfd(VirtIODevice *vdev);
EventNotifier *virtio_queue_get_host_notifier(VirtQueue *vq);
void virtio_queue_set_host_notifier_fd_handler(VirtQueue *vq, bool assign,
bool set_handler);
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index 97cdb11..a8d2bd8 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -153,8 +153,8 @@ void virtio_bus_set_vdev_config(VirtioBusState *bus, uint8_t *config)
* assign: register/deregister ioeventfd with the kernel
* set_handler: use the generic ioeventfd handler
*/
-static int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign, bool set_handler)
+int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
+ int n, bool assign, bool set_handler)
{
VirtIODevice *vdev = virtio_bus_get_device(bus);
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
@@ -185,61 +185,41 @@ static int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
return r;
}
-void virtio_bus_start_ioeventfd(VirtioBusState *bus)
+int virtio_bus_start_ioeventfd(VirtioBusState *bus)
{
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
DeviceState *proxy = DEVICE(BUS(bus)->parent);
- VirtIODevice *vdev;
- int n, r;
+ VirtIODevice *vdev = virtio_bus_get_device(bus);
+ VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
+ int r;
if (!k->ioeventfd_assign || k->ioeventfd_disabled(proxy)) {
- return;
+ return -ENOSYS;
}
if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
- return;
+ return 0;
}
- vdev = virtio_bus_get_device(bus);
- for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
- if (!virtio_queue_get_num(vdev, n)) {
- continue;
- }
- r = set_host_notifier_internal(proxy, bus, n, true, true);
- if (r < 0) {
- goto assign_error;
- }
+ r = vdc->start_ioeventfd(vdev);
+ if (r < 0) {
+ error_report("%s: failed. Fallback to userspace (slower).", __func__);
+ return r;
}
bus->ioeventfd_started = true;
- return;
-
-assign_error:
- while (--n >= 0) {
- if (!virtio_queue_get_num(vdev, n)) {
- continue;
- }
-
- r = set_host_notifier_internal(proxy, bus, n, false, false);
- assert(r >= 0);
- }
- error_report("%s: failed. Fallback to userspace (slower).", __func__);
+ return 0;
}
void virtio_bus_stop_ioeventfd(VirtioBusState *bus)
{
- DeviceState *proxy = DEVICE(BUS(bus)->parent);
VirtIODevice *vdev;
- int n, r;
+ VirtioDeviceClass *vdc;
if (!bus->ioeventfd_started) {
return;
}
+
vdev = virtio_bus_get_device(bus);
- for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
- if (!virtio_queue_get_num(vdev, n)) {
- continue;
- }
- r = set_host_notifier_internal(proxy, bus, n, false, false);
- assert(r >= 0);
- }
+ vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
+ vdc->stop_ioeventfd(vdev);
bus->ioeventfd_started = false;
}
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 3e318e4..e228518 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2172,15 +2172,79 @@ static Property virtio_properties[] = {
DEFINE_PROP_END_OF_LIST(),
};
+static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
+{
+ VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
+ DeviceState *proxy = DEVICE(BUS(qbus)->parent);
+ int n, r, err;
+
+ for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+ if (!virtio_queue_get_num(vdev, n)) {
+ continue;
+ }
+ r = set_host_notifier_internal(proxy, qbus, n, true, true);
+ if (r < 0) {
+ err = r;
+ goto assign_error;
+ }
+ }
+ return 0;
+
+assign_error:
+ while (--n >= 0) {
+ if (!virtio_queue_get_num(vdev, n)) {
+ continue;
+ }
+
+ r = set_host_notifier_internal(proxy, qbus, n, false, false);
+ assert(r >= 0);
+ }
+ return err;
+}
+
+int virtio_device_start_ioeventfd(VirtIODevice *vdev)
+{
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
+ VirtioBusState *vbus = VIRTIO_BUS(qbus);
+
+ return virtio_bus_start_ioeventfd(vbus);
+}
+
+static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
+{
+ VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
+ DeviceState *proxy = DEVICE(BUS(qbus)->parent);
+ int n, r;
+
+ for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+ if (!virtio_queue_get_num(vdev, n)) {
+ continue;
+ }
+ r = set_host_notifier_internal(proxy, qbus, n, false, false);
+ assert(r >= 0);
+ }
+}
+
+void virtio_device_stop_ioeventfd(VirtIODevice *vdev)
+{
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
+ VirtioBusState *vbus = VIRTIO_BUS(qbus);
+
+ virtio_bus_stop_ioeventfd(vbus);
+}
+
static void virtio_device_class_init(ObjectClass *klass, void *data)
{
/* Set the default value here. */
+ VirtioDeviceClass *vdc = VIRTIO_DEVICE_CLASS(klass);
DeviceClass *dc = DEVICE_CLASS(klass);
dc->realize = virtio_device_realize;
dc->unrealize = virtio_device_unrealize;
dc->bus_type = TYPE_VIRTIO_BUS;
dc->props = virtio_properties;
+ vdc->start_ioeventfd = virtio_device_start_ioeventfd_impl;
+ vdc->stop_ioeventfd = virtio_device_stop_ioeventfd_impl;
}
static const TypeInfo virtio_device_info = {
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 07/47] virtio: introduce virtio_device_ioeventfd_enabled
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (5 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 06/47] virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 08/47] virtio-blk: always use dataplane path if ioeventfd is active Michael S. Tsirkin
` (42 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck,
Christian Borntraeger, Alexander Graf, Richard Henderson
From: Paolo Bonzini <pbonzini@redhat.com>
This will be used to forbid iothread configuration when the
proxy does not allow using ioeventfd. To simplify the implementation,
change the direction of the ioeventfd_disabled callback too.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-bus.h | 8 +++++---
include/hw/virtio/virtio.h | 1 +
hw/s390x/virtio-ccw.c | 6 +++---
hw/virtio/virtio-bus.c | 10 +++++++++-
hw/virtio/virtio-mmio.c | 6 +++---
hw/virtio/virtio-pci.c | 6 +++---
hw/virtio/virtio.c | 8 ++++++++
7 files changed, 32 insertions(+), 13 deletions(-)
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index aa326e1..521fac7 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -71,10 +71,10 @@ typedef struct VirtioBusClass {
int (*query_nvectors)(DeviceState *d);
/*
* ioeventfd handling: if the transport implements ioeventfd_assign,
- * it must implement ioeventfd_disabled as well.
+ * it must implement ioeventfd_enabled as well.
*/
- /* Returns true if the ioeventfd has been disabled for the device. */
- bool (*ioeventfd_disabled)(DeviceState *d);
+ /* Returns true if the ioeventfd is enabled for the device. */
+ bool (*ioeventfd_enabled)(DeviceState *d);
/*
* Assigns/deassigns the ioeventfd backing for the transport on
* the device for queue number n. Returns an error value on
@@ -131,6 +131,8 @@ static inline VirtIODevice *virtio_bus_get_device(VirtioBusState *bus)
return (VirtIODevice *)qdev;
}
+/* Return whether the proxy allows ioeventfd. */
+bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus);
/* Start the ioeventfd. */
int virtio_bus_start_ioeventfd(VirtioBusState *bus);
/* Stop the ioeventfd. */
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 27e5151..12292bb 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -273,6 +273,7 @@ void virtio_queue_set_guest_notifier_fd_handler(VirtQueue *vq, bool assign,
bool with_irqfd);
int virtio_device_start_ioeventfd(VirtIODevice *vdev);
void virtio_device_stop_ioeventfd(VirtIODevice *vdev);
+bool virtio_device_ioeventfd_enabled(VirtIODevice *vdev);
EventNotifier *virtio_queue_get_host_notifier(VirtQueue *vq);
void virtio_queue_set_host_notifier_fd_handler(VirtQueue *vq, bool assign,
bool set_handler);
diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
index bf5670c..7d7f8f6 100644
--- a/hw/s390x/virtio-ccw.c
+++ b/hw/s390x/virtio-ccw.c
@@ -59,11 +59,11 @@ static void virtio_ccw_stop_ioeventfd(VirtioCcwDevice *dev)
virtio_bus_stop_ioeventfd(&dev->bus);
}
-static bool virtio_ccw_ioeventfd_disabled(DeviceState *d)
+static bool virtio_ccw_ioeventfd_enabled(DeviceState *d)
{
VirtioCcwDevice *dev = VIRTIO_CCW_DEVICE(d);
- return !(dev->flags & VIRTIO_CCW_FLAG_USE_IOEVENTFD);
+ return (dev->flags & VIRTIO_CCW_FLAG_USE_IOEVENTFD) != 0;
}
static int virtio_ccw_ioeventfd_assign(DeviceState *d, EventNotifier *notifier,
@@ -1589,7 +1589,7 @@ static void virtio_ccw_bus_class_init(ObjectClass *klass, void *data)
k->pre_plugged = virtio_ccw_pre_plugged;
k->device_plugged = virtio_ccw_device_plugged;
k->device_unplugged = virtio_ccw_device_unplugged;
- k->ioeventfd_disabled = virtio_ccw_ioeventfd_disabled;
+ k->ioeventfd_enabled = virtio_ccw_ioeventfd_enabled;
k->ioeventfd_assign = virtio_ccw_ioeventfd_assign;
}
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index a8d2bd8..3ffec66 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -193,7 +193,7 @@ int virtio_bus_start_ioeventfd(VirtioBusState *bus)
VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
int r;
- if (!k->ioeventfd_assign || k->ioeventfd_disabled(proxy)) {
+ if (!k->ioeventfd_assign || !k->ioeventfd_enabled(proxy)) {
return -ENOSYS;
}
if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
@@ -223,6 +223,14 @@ void virtio_bus_stop_ioeventfd(VirtioBusState *bus)
bus->ioeventfd_started = false;
}
+bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus)
+{
+ VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
+ DeviceState *proxy = DEVICE(BUS(bus)->parent);
+
+ return k->ioeventfd_assign && k->ioeventfd_enabled(proxy);
+}
+
/*
* This function switches from/to the generic ioeventfd handler.
* assign==false means 'use generic ioeventfd handler'.
diff --git a/hw/virtio/virtio-mmio.c b/hw/virtio/virtio-mmio.c
index 04a9655..a30270f 100644
--- a/hw/virtio/virtio-mmio.c
+++ b/hw/virtio/virtio-mmio.c
@@ -92,9 +92,9 @@ typedef struct {
bool format_transport_address;
} VirtIOMMIOProxy;
-static bool virtio_mmio_ioeventfd_disabled(DeviceState *d)
+static bool virtio_mmio_ioeventfd_enabled(DeviceState *d)
{
- return !kvm_eventfds_enabled();
+ return kvm_eventfds_enabled();
}
static int virtio_mmio_ioeventfd_assign(DeviceState *d,
@@ -531,7 +531,7 @@ static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
k->save_config = virtio_mmio_save_config;
k->load_config = virtio_mmio_load_config;
k->set_guest_notifiers = virtio_mmio_set_guest_notifiers;
- k->ioeventfd_disabled = virtio_mmio_ioeventfd_disabled;
+ k->ioeventfd_enabled = virtio_mmio_ioeventfd_enabled;
k->ioeventfd_assign = virtio_mmio_ioeventfd_assign;
k->has_variable_vring_alignment = true;
bus_class->max_dev = 1;
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index f00fc25..62001b4 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -262,11 +262,11 @@ static int virtio_pci_load_queue(DeviceState *d, int n, QEMUFile *f)
return 0;
}
-static bool virtio_pci_ioeventfd_disabled(DeviceState *d)
+static bool virtio_pci_ioeventfd_enabled(DeviceState *d)
{
VirtIOPCIProxy *proxy = to_virtio_pci_proxy(d);
- return !(proxy->flags & VIRTIO_PCI_FLAG_USE_IOEVENTFD);
+ return (proxy->flags & VIRTIO_PCI_FLAG_USE_IOEVENTFD) != 0;
}
#define QEMU_VIRTIO_PCI_QUEUE_MEM_MULT 0x1000
@@ -2516,7 +2516,7 @@ static void virtio_pci_bus_class_init(ObjectClass *klass, void *data)
k->device_plugged = virtio_pci_device_plugged;
k->device_unplugged = virtio_pci_device_unplugged;
k->query_nvectors = virtio_pci_query_nvectors;
- k->ioeventfd_disabled = virtio_pci_ioeventfd_disabled;
+ k->ioeventfd_enabled = virtio_pci_ioeventfd_enabled;
k->ioeventfd_assign = virtio_pci_ioeventfd_assign;
}
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index e228518..138a644 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2247,6 +2247,14 @@ static void virtio_device_class_init(ObjectClass *klass, void *data)
vdc->stop_ioeventfd = virtio_device_stop_ioeventfd_impl;
}
+bool virtio_device_ioeventfd_enabled(VirtIODevice *vdev)
+{
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
+ VirtioBusState *vbus = VIRTIO_BUS(qbus);
+
+ return virtio_bus_ioeventfd_enabled(vbus);
+}
+
static const TypeInfo virtio_device_info = {
.name = TYPE_VIRTIO_DEVICE,
.parent = TYPE_DEVICE,
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 08/47] virtio-blk: always use dataplane path if ioeventfd is active
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (6 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 07/47] virtio: introduce virtio_device_ioeventfd_enabled Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 09/47] virtio-scsi: " Michael S. Tsirkin
` (41 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Paolo Bonzini, Cornelia Huck, Stefan Hajnoczi,
Kevin Wolf, Max Reitz, qemu-block
From: Paolo Bonzini <pbonzini@redhat.com>
Override start_ioeventfd and stop_ioeventfd to start/stop the
whole dataplane logic. This has some positive side effects:
- no need anymore for virtio_add_queue_aio (i.e. a revert of
commit 0ff841f6d138904d514efa1d885bcaf54583852d)
- no need anymore to switch from generic ioeventfd handlers to
dataplane
It detects some errors better:
$ qemu-system-x86_64 -object iothread,id=io \
-drive id=null,file=null-aio://,if=none,format=raw \
-device virtio-blk-pci,ioeventfd=off,iothread=io,drive=null
qemu-system-x86_64: -device virtio-blk-pci,ioeventfd=off,iothread=io,drive=null:
ioeventfd is required for iothread
while previously it would have started just fine.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/block/dataplane/virtio-blk.h | 6 ++--
hw/block/dataplane/virtio-blk.c | 73 +++++++++++++++++++++++++----------------
hw/block/virtio-blk.c | 15 ++++-----
3 files changed, 55 insertions(+), 39 deletions(-)
diff --git a/hw/block/dataplane/virtio-blk.h b/hw/block/dataplane/virtio-blk.h
index b1f0b95..db3f47b 100644
--- a/hw/block/dataplane/virtio-blk.h
+++ b/hw/block/dataplane/virtio-blk.h
@@ -23,9 +23,9 @@ void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *conf,
VirtIOBlockDataPlane **dataplane,
Error **errp);
void virtio_blk_data_plane_destroy(VirtIOBlockDataPlane *s);
-void virtio_blk_data_plane_start(VirtIOBlockDataPlane *s);
-void virtio_blk_data_plane_stop(VirtIOBlockDataPlane *s);
-void virtio_blk_data_plane_drain(VirtIOBlockDataPlane *s);
void virtio_blk_data_plane_notify(VirtIOBlockDataPlane *s, VirtQueue *vq);
+int virtio_blk_data_plane_start(VirtIODevice *vdev);
+void virtio_blk_data_plane_stop(VirtIODevice *vdev);
+
#endif /* HW_DATAPLANE_VIRTIO_BLK_H */
diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c
index 9b268f4..90ef557 100644
--- a/hw/block/dataplane/virtio-blk.c
+++ b/hw/block/dataplane/virtio-blk.c
@@ -88,23 +88,28 @@ void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *conf,
*dataplane = NULL;
- if (!conf->iothread) {
- return;
- }
+ if (conf->iothread) {
+ if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
+ error_setg(errp,
+ "device is incompatible with iothread "
+ "(transport does not support notifiers)");
+ return;
+ }
+ if (!virtio_device_ioeventfd_enabled(vdev)) {
+ error_setg(errp, "ioeventfd is required for iothread");
+ return;
+ }
- /* Don't try if transport does not support notifiers. */
- if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
- error_setg(errp,
- "device is incompatible with dataplane "
- "(transport does not support notifiers)");
- return;
+ /* If dataplane is (re-)enabled while the guest is running there could
+ * be block jobs that can conflict.
+ */
+ if (blk_op_is_blocked(conf->conf.blk, BLOCK_OP_TYPE_DATAPLANE, errp)) {
+ error_prepend(errp, "cannot start virtio-blk dataplane: ");
+ return;
+ }
}
-
- /* If dataplane is (re-)enabled while the guest is running there could be
- * block jobs that can conflict.
- */
- if (blk_op_is_blocked(conf->conf.blk, BLOCK_OP_TYPE_DATAPLANE, errp)) {
- error_prepend(errp, "cannot start dataplane thread: ");
+ /* Don't try if transport does not support notifiers. */
+ if (!virtio_device_ioeventfd_enabled(vdev)) {
return;
}
@@ -112,9 +117,13 @@ void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *conf,
s->vdev = vdev;
s->conf = conf;
- s->iothread = conf->iothread;
- object_ref(OBJECT(s->iothread));
- s->ctx = iothread_get_aio_context(s->iothread);
+ if (conf->iothread) {
+ s->iothread = conf->iothread;
+ object_ref(OBJECT(s->iothread));
+ s->ctx = iothread_get_aio_context(s->iothread);
+ } else {
+ s->ctx = qemu_get_aio_context();
+ }
s->bh = aio_bh_new(s->ctx, notify_guest_bh, s);
s->batch_notify_vqs = bitmap_new(conf->num_queues);
@@ -124,14 +133,19 @@ void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *conf,
/* Context: QEMU global mutex held */
void virtio_blk_data_plane_destroy(VirtIOBlockDataPlane *s)
{
+ VirtIOBlock *vblk;
+
if (!s) {
return;
}
- virtio_blk_data_plane_stop(s);
+ vblk = VIRTIO_BLK(s->vdev);
+ assert(!vblk->dataplane_started);
g_free(s->batch_notify_vqs);
qemu_bh_delete(s->bh);
- object_unref(OBJECT(s->iothread));
+ if (s->iothread) {
+ object_unref(OBJECT(s->iothread));
+ }
g_free(s);
}
@@ -147,17 +161,18 @@ static void virtio_blk_data_plane_handle_output(VirtIODevice *vdev,
}
/* Context: QEMU global mutex held */
-void virtio_blk_data_plane_start(VirtIOBlockDataPlane *s)
+int virtio_blk_data_plane_start(VirtIODevice *vdev)
{
- BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(s->vdev)));
+ VirtIOBlock *vblk = VIRTIO_BLK(vdev);
+ VirtIOBlockDataPlane *s = vblk->dataplane;
+ BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(vblk)));
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
- VirtIOBlock *vblk = VIRTIO_BLK(s->vdev);
unsigned i;
unsigned nvqs = s->conf->num_queues;
int r;
if (vblk->dataplane_started || s->starting) {
- return;
+ return 0;
}
s->starting = true;
@@ -204,20 +219,22 @@ void virtio_blk_data_plane_start(VirtIOBlockDataPlane *s)
virtio_blk_data_plane_handle_output);
}
aio_context_release(s->ctx);
- return;
+ return 0;
fail_guest_notifiers:
vblk->dataplane_disabled = true;
s->starting = false;
vblk->dataplane_started = true;
+ return -ENOSYS;
}
/* Context: QEMU global mutex held */
-void virtio_blk_data_plane_stop(VirtIOBlockDataPlane *s)
+void virtio_blk_data_plane_stop(VirtIODevice *vdev)
{
- BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(s->vdev)));
+ VirtIOBlock *vblk = VIRTIO_BLK(vdev);
+ VirtIOBlockDataPlane *s = vblk->dataplane;
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vblk));
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
- VirtIOBlock *vblk = VIRTIO_BLK(s->vdev);
unsigned i;
unsigned nvqs = s->conf->num_queues;
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index 37fe72b..0c5fd27 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -611,7 +611,7 @@ static void virtio_blk_handle_output(VirtIODevice *vdev, VirtQueue *vq)
/* Some guests kick before setting VIRTIO_CONFIG_S_DRIVER_OK so start
* dataplane here instead of waiting for .set_status().
*/
- virtio_blk_data_plane_start(s->dataplane);
+ virtio_device_start_ioeventfd(vdev);
if (!s->dataplane_disabled) {
return;
}
@@ -687,11 +687,9 @@ static void virtio_blk_reset(VirtIODevice *vdev)
virtio_blk_free_request(req);
}
- if (s->dataplane) {
- virtio_blk_data_plane_stop(s->dataplane);
- }
aio_context_release(ctx);
+ assert(!s->dataplane_started);
blk_set_enable_write_cache(s->blk, s->original_wce);
}
@@ -789,9 +787,8 @@ static void virtio_blk_set_status(VirtIODevice *vdev, uint8_t status)
{
VirtIOBlock *s = VIRTIO_BLK(vdev);
- if (s->dataplane && !(status & (VIRTIO_CONFIG_S_DRIVER |
- VIRTIO_CONFIG_S_DRIVER_OK))) {
- virtio_blk_data_plane_stop(s->dataplane);
+ if (!(status & (VIRTIO_CONFIG_S_DRIVER | VIRTIO_CONFIG_S_DRIVER_OK))) {
+ assert(!s->dataplane_started);
}
if (!(status & VIRTIO_CONFIG_S_DRIVER_OK)) {
@@ -919,7 +916,7 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp)
s->sector_mask = (s->conf.conf.logical_block_size / BDRV_SECTOR_SIZE) - 1;
for (i = 0; i < conf->num_queues; i++) {
- virtio_add_queue_aio(vdev, 128, virtio_blk_handle_output);
+ virtio_add_queue(vdev, 128, virtio_blk_handle_output);
}
virtio_blk_data_plane_create(vdev, conf, &s->dataplane, &err);
if (err != NULL) {
@@ -1002,6 +999,8 @@ static void virtio_blk_class_init(ObjectClass *klass, void *data)
vdc->reset = virtio_blk_reset;
vdc->save = virtio_blk_save_device;
vdc->load = virtio_blk_load_device;
+ vdc->start_ioeventfd = virtio_blk_data_plane_start;
+ vdc->stop_ioeventfd = virtio_blk_data_plane_stop;
}
static const TypeInfo virtio_blk_info = {
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 09/47] virtio-scsi: always use dataplane path if ioeventfd is active
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (7 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 08/47] virtio-blk: always use dataplane path if ioeventfd is active Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 10/47] Revert "virtio: Introduce virtio_add_queue_aio" Michael S. Tsirkin
` (40 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
Override start_ioeventfd and stop_ioeventfd to start/stop the
whole dataplane logic. This has some positive side effects:
- no need anymore for virtio_add_queue_aio (i.e. a revert of
commit 1c627137c10ee2dcf59e0383ade8a9abfa2d4355)
- no need anymore to switch from generic ioeventfd handlers to
dataplane
It detects some errors better:
$ qemu-system-x86_64 -object iothread,id=io \
-device virtio-scsi-pci,ioeventfd=off,iothread=io
qemu-system-x86_64: -device virtio-scsi-pci,ioeventfd=off,iothread=io:
ioeventfd is required for iothread
while previously it would have started just fine.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-scsi.h | 6 ++---
hw/scsi/virtio-scsi-dataplane.c | 56 +++++++++++++++++++++++++----------------
hw/scsi/virtio-scsi.c | 24 ++++++++----------
3 files changed, 48 insertions(+), 38 deletions(-)
diff --git a/include/hw/virtio/virtio-scsi.h b/include/hw/virtio/virtio-scsi.h
index a1e0cfb..9fbc7d7 100644
--- a/include/hw/virtio/virtio-scsi.h
+++ b/include/hw/virtio/virtio-scsi.h
@@ -134,9 +134,9 @@ void virtio_scsi_free_req(VirtIOSCSIReq *req);
void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
uint32_t event, uint32_t reason);
-void virtio_scsi_set_iothread(VirtIOSCSI *s, IOThread *iothread);
-void virtio_scsi_dataplane_start(VirtIOSCSI *s);
-void virtio_scsi_dataplane_stop(VirtIOSCSI *s);
+void virtio_scsi_dataplane_setup(VirtIOSCSI *s, Error **errp);
+int virtio_scsi_dataplane_start(VirtIODevice *s);
+void virtio_scsi_dataplane_stop(VirtIODevice *s);
void virtio_scsi_dataplane_notify(VirtIODevice *vdev, VirtIOSCSIReq *req);
#endif /* QEMU_VIRTIO_SCSI_H */
diff --git a/hw/scsi/virtio-scsi-dataplane.c b/hw/scsi/virtio-scsi-dataplane.c
index f537b4e..aa6be54 100644
--- a/hw/scsi/virtio-scsi-dataplane.c
+++ b/hw/scsi/virtio-scsi-dataplane.c
@@ -12,6 +12,7 @@
*/
#include "qemu/osdep.h"
+#include "qapi/error.h"
#include "hw/virtio/virtio-scsi.h"
#include "qemu/error-report.h"
#include "sysemu/block-backend.h"
@@ -21,20 +22,30 @@
#include "hw/virtio/virtio-access.h"
/* Context: QEMU global mutex held */
-void virtio_scsi_set_iothread(VirtIOSCSI *s, IOThread *iothread)
+void virtio_scsi_dataplane_setup(VirtIOSCSI *s, Error **errp)
{
- BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(s)));
- VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(s);
+ VirtIODevice *vdev = VIRTIO_DEVICE(s);
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
+ VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
- assert(!s->ctx);
- s->ctx = iothread_get_aio_context(vs->conf.iothread);
-
- /* Don't try if transport does not support notifiers. */
- if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
- fprintf(stderr, "virtio-scsi: Failed to set iothread "
- "(transport does not support notifiers)");
- exit(1);
+ if (vs->conf.iothread) {
+ if (!k->set_guest_notifiers || !k->ioeventfd_assign) {
+ error_setg(errp,
+ "device is incompatible with iothread "
+ "(transport does not support notifiers)");
+ return;
+ }
+ if (!virtio_device_ioeventfd_enabled(vdev)) {
+ error_setg(errp, "ioeventfd is required for iothread");
+ return;
+ }
+ s->ctx = iothread_get_aio_context(vs->conf.iothread);
+ } else {
+ if (!virtio_device_ioeventfd_enabled(vdev)) {
+ return;
+ }
+ s->ctx = qemu_get_aio_context();
}
}
@@ -105,19 +116,19 @@ static void virtio_scsi_clear_aio(VirtIOSCSI *s)
}
/* Context: QEMU global mutex held */
-void virtio_scsi_dataplane_start(VirtIOSCSI *s)
+int virtio_scsi_dataplane_start(VirtIODevice *vdev)
{
int i;
int rc;
- BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(s)));
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
- VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(s);
+ VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev);
+ VirtIOSCSI *s = VIRTIO_SCSI(vdev);
if (s->dataplane_started ||
s->dataplane_starting ||
- s->dataplane_fenced ||
- s->ctx != iothread_get_aio_context(vs->conf.iothread)) {
- return;
+ s->dataplane_fenced) {
+ return 0;
}
s->dataplane_starting = true;
@@ -152,7 +163,7 @@ void virtio_scsi_dataplane_start(VirtIOSCSI *s)
s->dataplane_starting = false;
s->dataplane_started = true;
aio_context_release(s->ctx);
- return;
+ return 0;
fail_vrings:
virtio_scsi_clear_aio(s);
@@ -165,14 +176,16 @@ fail_guest_notifiers:
s->dataplane_fenced = true;
s->dataplane_starting = false;
s->dataplane_started = true;
+ return -ENOSYS;
}
/* Context: QEMU global mutex held */
-void virtio_scsi_dataplane_stop(VirtIOSCSI *s)
+void virtio_scsi_dataplane_stop(VirtIODevice *vdev)
{
- BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(s)));
+ BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
- VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(s);
+ VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev);
+ VirtIOSCSI *s = VIRTIO_SCSI(vdev);
int i;
if (!s->dataplane_started || s->dataplane_stopping) {
@@ -186,7 +199,6 @@ void virtio_scsi_dataplane_stop(VirtIOSCSI *s)
return;
}
s->dataplane_stopping = true;
- assert(s->ctx == iothread_get_aio_context(vs->conf.iothread));
aio_context_acquire(s->ctx);
diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
index 4762f05..3e5ae6a 100644
--- a/hw/scsi/virtio-scsi.c
+++ b/hw/scsi/virtio-scsi.c
@@ -434,7 +434,7 @@ static void virtio_scsi_handle_ctrl(VirtIODevice *vdev, VirtQueue *vq)
VirtIOSCSI *s = (VirtIOSCSI *)vdev;
if (s->ctx) {
- virtio_scsi_dataplane_start(s);
+ virtio_device_start_ioeventfd(vdev);
if (!s->dataplane_fenced) {
return;
}
@@ -610,7 +610,7 @@ static void virtio_scsi_handle_cmd(VirtIODevice *vdev, VirtQueue *vq)
VirtIOSCSI *s = (VirtIOSCSI *)vdev;
if (s->ctx) {
- virtio_scsi_dataplane_start(s);
+ virtio_device_start_ioeventfd(vdev);
if (!s->dataplane_fenced) {
return;
}
@@ -669,9 +669,7 @@ static void virtio_scsi_reset(VirtIODevice *vdev)
VirtIOSCSI *s = VIRTIO_SCSI(vdev);
VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev);
- if (s->ctx) {
- virtio_scsi_dataplane_stop(s);
- }
+ assert(!s->dataplane_started);
s->resetting++;
qbus_reset_all(&s->bus.qbus);
s->resetting--;
@@ -749,7 +747,7 @@ static void virtio_scsi_handle_event(VirtIODevice *vdev, VirtQueue *vq)
VirtIOSCSI *s = VIRTIO_SCSI(vdev);
if (s->ctx) {
- virtio_scsi_dataplane_start(s);
+ virtio_device_start_ioeventfd(vdev);
if (!s->dataplane_fenced) {
return;
}
@@ -848,14 +846,10 @@ void virtio_scsi_common_realize(DeviceState *dev, Error **errp,
s->sense_size = VIRTIO_SCSI_SENSE_DEFAULT_SIZE;
s->cdb_size = VIRTIO_SCSI_CDB_DEFAULT_SIZE;
- s->ctrl_vq = virtio_add_queue_aio(vdev, VIRTIO_SCSI_VQ_SIZE, ctrl);
- s->event_vq = virtio_add_queue_aio(vdev, VIRTIO_SCSI_VQ_SIZE, evt);
+ s->ctrl_vq = virtio_add_queue(vdev, VIRTIO_SCSI_VQ_SIZE, ctrl);
+ s->event_vq = virtio_add_queue(vdev, VIRTIO_SCSI_VQ_SIZE, evt);
for (i = 0; i < s->conf.num_queues; i++) {
- s->cmd_vqs[i] = virtio_add_queue_aio(vdev, VIRTIO_SCSI_VQ_SIZE, cmd);
- }
-
- if (s->conf.iothread) {
- virtio_scsi_set_iothread(VIRTIO_SCSI(s), s->conf.iothread);
+ s->cmd_vqs[i] = virtio_add_queue(vdev, VIRTIO_SCSI_VQ_SIZE, cmd);
}
}
@@ -885,6 +879,8 @@ static void virtio_scsi_device_realize(DeviceState *dev, Error **errp)
return;
}
}
+
+ virtio_scsi_dataplane_setup(s, errp);
}
static void virtio_scsi_instance_init(Object *obj)
@@ -957,6 +953,8 @@ static void virtio_scsi_class_init(ObjectClass *klass, void *data)
vdc->set_config = virtio_scsi_set_config;
vdc->get_features = virtio_scsi_get_features;
vdc->reset = virtio_scsi_reset;
+ vdc->start_ioeventfd = virtio_scsi_dataplane_start;
+ vdc->stop_ioeventfd = virtio_scsi_dataplane_stop;
hc->plug = virtio_scsi_hotplug;
hc->unplug = virtio_scsi_hotunplug;
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 10/47] Revert "virtio: Introduce virtio_add_queue_aio"
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (8 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 09/47] virtio-scsi: " Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 11/47] virtio: remove set_handler argument from set_host_notifier_internal Michael S. Tsirkin
` (39 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
This reverts commit 872dd82c83745a603d2e07a03d34313eb6467ae4.
virtio_add_queue_aio is unused.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio.h | 3 ---
hw/virtio/virtio.c | 38 ++++----------------------------------
2 files changed, 4 insertions(+), 37 deletions(-)
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 12292bb..2fcd23c 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -152,9 +152,6 @@ typedef void (*VirtIOHandleOutput)(VirtIODevice *, VirtQueue *);
VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
VirtIOHandleOutput handle_output);
-VirtQueue *virtio_add_queue_aio(VirtIODevice *vdev, int queue_size,
- VirtIOHandleOutput handle_output);
-
void virtio_del_queue(VirtIODevice *vdev, int n);
void *virtqueue_alloc_element(size_t sz, unsigned out_num, unsigned in_num);
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 138a644..5221aba 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -97,7 +97,6 @@ struct VirtQueue
uint16_t vector;
VirtIOHandleOutput handle_output;
VirtIOHandleOutput handle_aio_output;
- bool use_aio;
VirtIODevice *vdev;
EventNotifier guest_notifier;
EventNotifier host_notifier;
@@ -1287,9 +1286,8 @@ void virtio_queue_set_vector(VirtIODevice *vdev, int n, uint16_t vector)
}
}
-static VirtQueue *virtio_add_queue_internal(VirtIODevice *vdev, int queue_size,
- VirtIOHandleOutput handle_output,
- bool use_aio)
+VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
+ VirtIOHandleOutput handle_output)
{
int i;
@@ -1306,28 +1304,10 @@ static VirtQueue *virtio_add_queue_internal(VirtIODevice *vdev, int queue_size,
vdev->vq[i].vring.align = VIRTIO_PCI_VRING_ALIGN;
vdev->vq[i].handle_output = handle_output;
vdev->vq[i].handle_aio_output = NULL;
- vdev->vq[i].use_aio = use_aio;
return &vdev->vq[i];
}
-/* Add a virt queue and mark AIO.
- * An AIO queue will use the AioContext based event interface instead of the
- * default IOHandler and EventNotifier interface.
- */
-VirtQueue *virtio_add_queue_aio(VirtIODevice *vdev, int queue_size,
- VirtIOHandleOutput handle_output)
-{
- return virtio_add_queue_internal(vdev, queue_size, handle_output, true);
-}
-
-/* Add a normal virt queue (on the contrary to the AIO version above. */
-VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
- VirtIOHandleOutput handle_output)
-{
- return virtio_add_queue_internal(vdev, queue_size, handle_output, false);
-}
-
void virtio_del_queue(VirtIODevice *vdev, int n)
{
if (n < 0 || n >= VIRTIO_QUEUE_MAX) {
@@ -2073,21 +2053,11 @@ static void virtio_queue_host_notifier_read(EventNotifier *n)
void virtio_queue_set_host_notifier_fd_handler(VirtQueue *vq, bool assign,
bool set_handler)
{
- AioContext *ctx = qemu_get_aio_context();
if (assign && set_handler) {
- if (vq->use_aio) {
- aio_set_event_notifier(ctx, &vq->host_notifier, true,
+ event_notifier_set_handler(&vq->host_notifier, true,
virtio_queue_host_notifier_read);
- } else {
- event_notifier_set_handler(&vq->host_notifier, true,
- virtio_queue_host_notifier_read);
- }
} else {
- if (vq->use_aio) {
- aio_set_event_notifier(ctx, &vq->host_notifier, true, NULL);
- } else {
- event_notifier_set_handler(&vq->host_notifier, true, NULL);
- }
+ event_notifier_set_handler(&vq->host_notifier, true, NULL);
}
if (!assign) {
/* Test and clear notifier before after disabling event,
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 11/47] virtio: remove set_handler argument from set_host_notifier_internal
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (9 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 10/47] Revert "virtio: Introduce virtio_add_queue_aio" Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether Michael S. Tsirkin
` (38 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
Make virtio_device_start_ioeventfd_impl use the same logic as
dataplane to set up the host notifier. This removes the need
for the set_handler argument in set_host_notifier_internal.
This is a first step towards using virtio_bus_set_host_notifier
as the sole entry point to set up ioeventfds. At least now
the functions have the same interface, but they still differ
in that virtio_bus_set_host_notifier sets ioeventfd_disabled.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-bus.h | 2 +-
hw/virtio/virtio-bus.c | 12 +++---------
hw/virtio/virtio.c | 16 +++++++++++++---
3 files changed, 17 insertions(+), 13 deletions(-)
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index 521fac7..af6b5c4 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -143,6 +143,6 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign);
/* This is temporary. It is only needed because virtio_bus_set_host_notifier
* sets ioeventfd_disabled but we will shortly get rid of it. */
int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign, bool set_handler);
+ int n, bool assign);
#endif /* VIRTIO_BUS_H */
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index 3ffec66..a619445 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -147,14 +147,8 @@ void virtio_bus_set_vdev_config(VirtioBusState *bus, uint8_t *config)
}
}
-/*
- * This function handles both assigning the ioeventfd handler and
- * registering it with the kernel.
- * assign: register/deregister ioeventfd with the kernel
- * set_handler: use the generic ioeventfd handler
- */
int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign, bool set_handler)
+ int n, bool assign)
{
VirtIODevice *vdev = virtio_bus_get_device(bus);
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
@@ -169,7 +163,7 @@ int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
__func__, strerror(-r), r);
return r;
}
- virtio_queue_set_host_notifier_fd_handler(vq, true, set_handler);
+ virtio_queue_set_host_notifier_fd_handler(vq, true, false);
r = k->ioeventfd_assign(proxy, notifier, n, assign);
if (r < 0) {
error_report("%s: unable to assign ioeventfd: %d", __func__, r);
@@ -257,7 +251,7 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
*/
virtio_bus_stop_ioeventfd(bus);
}
- return set_host_notifier_internal(proxy, bus, n, assign, false);
+ return set_host_notifier_internal(proxy, bus, n, assign);
}
static char *virtio_bus_get_dev_path(DeviceState *dev)
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 5221aba..2ece690 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2152,11 +2152,21 @@ static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, true, true);
+ r = set_host_notifier_internal(proxy, qbus, n, true);
if (r < 0) {
err = r;
goto assign_error;
}
+ virtio_queue_set_host_notifier_fd_handler(&vdev->vq[n], true, true);
+ }
+
+ for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+ /* Kick right away to begin processing requests already in vring */
+ VirtQueue *vq = &vdev->vq[n];
+ if (!vq->vring.num) {
+ continue;
+ }
+ event_notifier_set(&vq->host_notifier);
}
return 0;
@@ -2166,7 +2176,7 @@ assign_error:
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, false, false);
+ r = set_host_notifier_internal(proxy, qbus, n, false);
assert(r >= 0);
}
return err;
@@ -2190,7 +2200,7 @@ static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, false, false);
+ r = set_host_notifier_internal(proxy, qbus, n, false);
assert(r >= 0);
}
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (10 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 11/47] virtio: remove set_handler argument from set_host_notifier_internal Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-11-10 14:35 ` Christian Borntraeger
2016-10-30 21:23 ` [Qemu-devel] [PULL 13/47] virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd Michael S. Tsirkin
` (37 subsequent siblings)
49 siblings, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
Now that there is not anymore a switch from the generic ioeventfd handler
to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
does nothing in this case. Move the invocation to vhost.c, which is the
only place that needs it.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-bus.h | 6 ------
hw/virtio/vhost.c | 3 +++
hw/virtio/virtio-bus.c | 23 ++++++++---------------
3 files changed, 11 insertions(+), 21 deletions(-)
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index af6b5c4..cbdf745 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -94,12 +94,6 @@ struct VirtioBusState {
BusState parent_obj;
/*
- * Set if the default ioeventfd handlers are disabled by vhost
- * or dataplane.
- */
- bool ioeventfd_disabled;
-
- /*
* Set if ioeventfd has been started.
*/
bool ioeventfd_started;
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index 501a5f4..131f164 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -1196,6 +1196,7 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
goto fail;
}
+ virtio_device_stop_ioeventfd(vdev);
for (i = 0; i < hdev->nvqs; ++i) {
r = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
true);
@@ -1215,6 +1216,7 @@ fail_vq:
}
assert (e >= 0);
}
+ virtio_device_start_ioeventfd(vdev);
fail:
return r;
}
@@ -1237,6 +1239,7 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
}
assert (r >= 0);
}
+ virtio_device_start_ioeventfd(vdev);
}
/* Test and clear event pending status.
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index a619445..b0e4544 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -190,7 +190,7 @@ int virtio_bus_start_ioeventfd(VirtioBusState *bus)
if (!k->ioeventfd_assign || !k->ioeventfd_enabled(proxy)) {
return -ENOSYS;
}
- if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
+ if (bus->ioeventfd_started) {
return 0;
}
r = vdc->start_ioeventfd(vdev);
@@ -226,8 +226,8 @@ bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus)
}
/*
- * This function switches from/to the generic ioeventfd handler.
- * assign==false means 'use generic ioeventfd handler'.
+ * This function switches ioeventfd on/off in the device.
+ * The caller must set or clear the handlers for the EventNotifier.
*/
int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
{
@@ -237,19 +237,12 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
if (!k->ioeventfd_assign) {
return -ENOSYS;
}
- bus->ioeventfd_disabled = assign;
if (assign) {
- /*
- * Stop using the generic ioeventfd, we are doing eventfd handling
- * ourselves below
- *
- * FIXME: We should just switch the handler and not deassign the
- * ioeventfd.
- * Otherwise, there's a window where we don't have an
- * ioeventfd and we may end up with a notification where
- * we don't expect one.
- */
- virtio_bus_stop_ioeventfd(bus);
+ assert(!bus->ioeventfd_started);
+ } else {
+ if (!bus->ioeventfd_started) {
+ return 0;
+ }
}
return set_host_notifier_internal(proxy, bus, n, assign);
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 13/47] virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (11 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 14/47] virtio: inline virtio_queue_set_host_notifier_fd_handler Michael S. Tsirkin
` (36 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
ioeventfd_disabled was the only reason for the default
implementation of virtio_device_start_ioeventfd not to use
virtio_bus_set_host_notifier. This is now fixed, and the sole entry
point to set up ioeventfd can be virtio_bus_set_host_notifier.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-bus.h | 5 -----
hw/virtio/virtio-bus.c | 4 ++--
hw/virtio/virtio.c | 8 +++-----
3 files changed, 5 insertions(+), 12 deletions(-)
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index cbdf745..fdf7fda 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -134,9 +134,4 @@ void virtio_bus_stop_ioeventfd(VirtioBusState *bus);
/* Switch from/to the generic ioeventfd handler */
int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign);
-/* This is temporary. It is only needed because virtio_bus_set_host_notifier
- * sets ioeventfd_disabled but we will shortly get rid of it. */
-int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign);
-
#endif /* VIRTIO_BUS_H */
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index b0e4544..fd105b8 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -147,8 +147,8 @@ void virtio_bus_set_vdev_config(VirtioBusState *bus, uint8_t *config)
}
}
-int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign)
+static int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
+ int n, bool assign)
{
VirtIODevice *vdev = virtio_bus_get_device(bus);
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 2ece690..b738163 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2145,14 +2145,13 @@ static Property virtio_properties[] = {
static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
{
VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
- DeviceState *proxy = DEVICE(BUS(qbus)->parent);
int n, r, err;
for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, true);
+ r = virtio_bus_set_host_notifier(qbus, n, true);
if (r < 0) {
err = r;
goto assign_error;
@@ -2176,7 +2175,7 @@ assign_error:
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, false);
+ r = virtio_bus_set_host_notifier(qbus, n, false);
assert(r >= 0);
}
return err;
@@ -2193,14 +2192,13 @@ int virtio_device_start_ioeventfd(VirtIODevice *vdev)
static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
{
VirtioBusState *qbus = VIRTIO_BUS(qdev_get_parent_bus(DEVICE(vdev)));
- DeviceState *proxy = DEVICE(BUS(qbus)->parent);
int n, r;
for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
- r = set_host_notifier_internal(proxy, qbus, n, false);
+ r = virtio_bus_set_host_notifier(qbus, n, false);
assert(r >= 0);
}
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 14/47] virtio: inline virtio_queue_set_host_notifier_fd_handler
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (12 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 13/47] virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd Michael S. Tsirkin
@ 2016-10-30 21:23 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 15/47] virtio: inline set_host_notifier_internal Michael S. Tsirkin
` (35 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:23 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
Of the three possible parameter combinations for
virtio_queue_set_host_notifier_fd_handler:
- assign=true/set_handler=true is only called from
virtio_device_start_ioeventfd
- assign=false/set_handler=false is called from
set_host_notifier_internal but it only does something when
reached from virtio_device_stop_ioeventfd_impl; otherwise
there is no EventNotifier set on qemu_get_aio_context().
- assign=true/set_handler=false is called from
set_host_notifier_internal, but it is not doing anything:
with the new start_ioeventfd and stop_ioeventfd methods,
there is never an EventNotifier set on qemu_get_aio_context()
at this point. This is enforced by the assertion in
virtio_bus_set_host_notifier.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio.h | 3 +--
hw/virtio/virtio-bus.c | 19 +++++++++++--------
hw/virtio/virtio.c | 27 +++++++++------------------
3 files changed, 21 insertions(+), 28 deletions(-)
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 2fcd23c..ac65d6a 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -272,8 +272,7 @@ int virtio_device_start_ioeventfd(VirtIODevice *vdev);
void virtio_device_stop_ioeventfd(VirtIODevice *vdev);
bool virtio_device_ioeventfd_enabled(VirtIODevice *vdev);
EventNotifier *virtio_queue_get_host_notifier(VirtQueue *vq);
-void virtio_queue_set_host_notifier_fd_handler(VirtQueue *vq, bool assign,
- bool set_handler);
+void virtio_queue_host_notifier_read(EventNotifier *n);
void virtio_queue_aio_set_host_notifier_handler(VirtQueue *vq, AioContext *ctx,
void (*fn)(VirtIODevice *,
VirtQueue *));
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index fd105b8..670746c 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -163,19 +163,22 @@ static int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
__func__, strerror(-r), r);
return r;
}
- virtio_queue_set_host_notifier_fd_handler(vq, true, false);
- r = k->ioeventfd_assign(proxy, notifier, n, assign);
+ r = k->ioeventfd_assign(proxy, notifier, n, true);
if (r < 0) {
error_report("%s: unable to assign ioeventfd: %d", __func__, r);
- virtio_queue_set_host_notifier_fd_handler(vq, false, false);
- event_notifier_cleanup(notifier);
- return r;
+ goto cleanup_event_notifier;
}
+ return 0;
} else {
- k->ioeventfd_assign(proxy, notifier, n, assign);
- virtio_queue_set_host_notifier_fd_handler(vq, false, false);
- event_notifier_cleanup(notifier);
+ k->ioeventfd_assign(proxy, notifier, n, false);
}
+
+cleanup_event_notifier:
+ /* Test and clear notifier after disabling event,
+ * in case poll callback didn't have time to run.
+ */
+ virtio_queue_host_notifier_read(notifier);
+ event_notifier_cleanup(notifier);
return r;
}
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index b738163..bcbcfe0 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -2042,7 +2042,7 @@ void virtio_queue_aio_set_host_notifier_handler(VirtQueue *vq, AioContext *ctx,
}
}
-static void virtio_queue_host_notifier_read(EventNotifier *n)
+void virtio_queue_host_notifier_read(EventNotifier *n)
{
VirtQueue *vq = container_of(n, VirtQueue, host_notifier);
if (event_notifier_test_and_clear(n)) {
@@ -2050,22 +2050,6 @@ static void virtio_queue_host_notifier_read(EventNotifier *n)
}
}
-void virtio_queue_set_host_notifier_fd_handler(VirtQueue *vq, bool assign,
- bool set_handler)
-{
- if (assign && set_handler) {
- event_notifier_set_handler(&vq->host_notifier, true,
- virtio_queue_host_notifier_read);
- } else {
- event_notifier_set_handler(&vq->host_notifier, true, NULL);
- }
- if (!assign) {
- /* Test and clear notifier before after disabling event,
- * in case poll callback didn't have time to run. */
- virtio_queue_host_notifier_read(&vq->host_notifier);
- }
-}
-
EventNotifier *virtio_queue_get_host_notifier(VirtQueue *vq)
{
return &vq->host_notifier;
@@ -2148,6 +2132,7 @@ static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
int n, r, err;
for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+ VirtQueue *vq = &vdev->vq[n];
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
@@ -2156,7 +2141,8 @@ static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
err = r;
goto assign_error;
}
- virtio_queue_set_host_notifier_fd_handler(&vdev->vq[n], true, true);
+ event_notifier_set_handler(&vq->host_notifier, true,
+ virtio_queue_host_notifier_read);
}
for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
@@ -2171,10 +2157,12 @@ static int virtio_device_start_ioeventfd_impl(VirtIODevice *vdev)
assign_error:
while (--n >= 0) {
+ VirtQueue *vq = &vdev->vq[n];
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
+ event_notifier_set_handler(&vq->host_notifier, true, NULL);
r = virtio_bus_set_host_notifier(qbus, n, false);
assert(r >= 0);
}
@@ -2195,9 +2183,12 @@ static void virtio_device_stop_ioeventfd_impl(VirtIODevice *vdev)
int n, r;
for (n = 0; n < VIRTIO_QUEUE_MAX; n++) {
+ VirtQueue *vq = &vdev->vq[n];
+
if (!virtio_queue_get_num(vdev, n)) {
continue;
}
+ event_notifier_set_handler(&vq->host_notifier, true, NULL);
r = virtio_bus_set_host_notifier(qbus, n, false);
assert(r >= 0);
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 15/47] virtio: inline set_host_notifier_internal
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (13 preceding siblings ...)
2016-10-30 21:23 ` [Qemu-devel] [PULL 14/47] virtio: inline virtio_queue_set_host_notifier_fd_handler Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 16/47] cryptodev: introduce cryptodev backend interface Michael S. Tsirkin
` (34 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Paolo Bonzini, Stefan Hajnoczi, Cornelia Huck
From: Paolo Bonzini <pbonzini@redhat.com>
This is only called from virtio_bus_set_host_notifier.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/virtio-bus.c | 62 +++++++++++++++++++++-----------------------------
1 file changed, 26 insertions(+), 36 deletions(-)
diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
index 670746c..bf61f66 100644
--- a/hw/virtio/virtio-bus.c
+++ b/hw/virtio/virtio-bus.c
@@ -147,41 +147,6 @@ void virtio_bus_set_vdev_config(VirtioBusState *bus, uint8_t *config)
}
}
-static int set_host_notifier_internal(DeviceState *proxy, VirtioBusState *bus,
- int n, bool assign)
-{
- VirtIODevice *vdev = virtio_bus_get_device(bus);
- VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
- VirtQueue *vq = virtio_get_queue(vdev, n);
- EventNotifier *notifier = virtio_queue_get_host_notifier(vq);
- int r = 0;
-
- if (assign) {
- r = event_notifier_init(notifier, 1);
- if (r < 0) {
- error_report("%s: unable to init event notifier: %s (%d)",
- __func__, strerror(-r), r);
- return r;
- }
- r = k->ioeventfd_assign(proxy, notifier, n, true);
- if (r < 0) {
- error_report("%s: unable to assign ioeventfd: %d", __func__, r);
- goto cleanup_event_notifier;
- }
- return 0;
- } else {
- k->ioeventfd_assign(proxy, notifier, n, false);
- }
-
-cleanup_event_notifier:
- /* Test and clear notifier after disabling event,
- * in case poll callback didn't have time to run.
- */
- virtio_queue_host_notifier_read(notifier);
- event_notifier_cleanup(notifier);
- return r;
-}
-
int virtio_bus_start_ioeventfd(VirtioBusState *bus)
{
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
@@ -234,20 +199,45 @@ bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus)
*/
int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
{
+ VirtIODevice *vdev = virtio_bus_get_device(bus);
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(bus);
DeviceState *proxy = DEVICE(BUS(bus)->parent);
+ VirtQueue *vq = virtio_get_queue(vdev, n);
+ EventNotifier *notifier = virtio_queue_get_host_notifier(vq);
+ int r = 0;
if (!k->ioeventfd_assign) {
return -ENOSYS;
}
+
if (assign) {
assert(!bus->ioeventfd_started);
+ r = event_notifier_init(notifier, 1);
+ if (r < 0) {
+ error_report("%s: unable to init event notifier: %s (%d)",
+ __func__, strerror(-r), r);
+ return r;
+ }
+ r = k->ioeventfd_assign(proxy, notifier, n, true);
+ if (r < 0) {
+ error_report("%s: unable to assign ioeventfd: %d", __func__, r);
+ goto cleanup_event_notifier;
+ }
+ return 0;
} else {
if (!bus->ioeventfd_started) {
return 0;
}
+ k->ioeventfd_assign(proxy, notifier, n, false);
}
- return set_host_notifier_internal(proxy, bus, n, assign);
+
+cleanup_event_notifier:
+ /* Test and clear notifier after disabling event,
+ * in case poll callback didn't have time to run.
+ */
+ virtio_queue_host_notifier_read(notifier);
+ event_notifier_cleanup(notifier);
+ return r;
}
static char *virtio_bus_get_dev_path(DeviceState *dev)
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 16/47] cryptodev: introduce cryptodev backend interface
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (14 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 15/47] virtio: inline set_host_notifier_internal Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 17/47] cryptodev: add symmetric algorithm operation stuff Michael S. Tsirkin
` (33 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
cryptodev backend interface is used to realize the active work for
virtual crypto device.
This patch only add the framework, doesn't include specific operations.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/sysemu/cryptodev.h | 148 ++++++++++++++++++++++++++++++++++++++
backends/cryptodev.c | 176 +++++++++++++++++++++++++++++++++++++++++++++
backends/Makefile.objs | 2 +
3 files changed, 326 insertions(+)
create mode 100644 include/sysemu/cryptodev.h
create mode 100644 backends/cryptodev.c
diff --git a/include/sysemu/cryptodev.h b/include/sysemu/cryptodev.h
new file mode 100644
index 0000000..bfaf771
--- /dev/null
+++ b/include/sysemu/cryptodev.h
@@ -0,0 +1,148 @@
+/*
+ * QEMU Crypto Device Implementation
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#ifndef CRYPTODEV_H
+#define CRYPTODEV_H
+
+#include "qom/object.h"
+#include "qemu-common.h"
+
+/**
+ * CryptoDevBackend:
+ *
+ * The CryptoDevBackend object is an interface
+ * for different cryptodev backends, which provides crypto
+ * operation wrapper.
+ *
+ */
+
+#define TYPE_CRYPTODEV_BACKEND "cryptodev-backend"
+
+#define CRYPTODEV_BACKEND(obj) \
+ OBJECT_CHECK(CryptoDevBackend, \
+ (obj), TYPE_CRYPTODEV_BACKEND)
+#define CRYPTODEV_BACKEND_GET_CLASS(obj) \
+ OBJECT_GET_CLASS(CryptoDevBackendClass, \
+ (obj), TYPE_CRYPTODEV_BACKEND)
+#define CRYPTODEV_BACKEND_CLASS(klass) \
+ OBJECT_CLASS_CHECK(CryptoDevBackendClass, \
+ (klass), TYPE_CRYPTODEV_BACKEND)
+
+
+#define MAX_CRYPTO_QUEUE_NUM 64
+
+typedef struct CryptoDevBackendConf CryptoDevBackendConf;
+typedef struct CryptoDevBackendPeers CryptoDevBackendPeers;
+typedef struct CryptoDevBackendClient
+ CryptoDevBackendClient;
+typedef struct CryptoDevBackend CryptoDevBackend;
+
+
+typedef struct CryptoDevBackendClass {
+ ObjectClass parent_class;
+
+ void (*init)(CryptoDevBackend *backend, Error **errp);
+ void (*cleanup)(CryptoDevBackend *backend, Error **errp);
+} CryptoDevBackendClass;
+
+
+struct CryptoDevBackendClient {
+ char *model;
+ char *name;
+ char *info_str;
+ unsigned int queue_index;
+ QTAILQ_ENTRY(CryptoDevBackendClient) next;
+};
+
+struct CryptoDevBackendPeers {
+ CryptoDevBackendClient *ccs[MAX_CRYPTO_QUEUE_NUM];
+ uint32_t queues;
+};
+
+struct CryptoDevBackendConf {
+ CryptoDevBackendPeers peers;
+
+ /* Supported service mask */
+ uint32_t crypto_services;
+
+ /* Detailed algorithms mask */
+ uint32_t cipher_algo_l;
+ uint32_t cipher_algo_h;
+ uint32_t hash_algo;
+ uint32_t mac_algo_l;
+ uint32_t mac_algo_h;
+ uint32_t aead_algo;
+ /* Maximum length of cipher key */
+ uint32_t max_cipher_key_len;
+ /* Maximum length of authenticated key */
+ uint32_t max_auth_key_len;
+ /* Maximum size of each crypto request's content */
+ uint64_t max_size;
+};
+
+struct CryptoDevBackend {
+ Object parent_obj;
+
+ bool ready;
+ CryptoDevBackendConf conf;
+};
+
+/**
+ * cryptodev_backend_new_client:
+ * @model: the cryptodev backend model
+ * @name: the cryptodev backend name, can be NULL
+ *
+ * Creates a new cryptodev backend client object
+ * with the @name in the model @model.
+ *
+ * The returned object must be released with
+ * cryptodev_backend_free_client() when no
+ * longer required
+ *
+ * Returns: a new cryptodev backend client object
+ */
+CryptoDevBackendClient *
+cryptodev_backend_new_client(const char *model,
+ const char *name);
+/**
+ * cryptodev_backend_free_client:
+ * @cc: the cryptodev backend client object
+ *
+ * Release the memory associated with @cc that
+ * was previously allocated by cryptodev_backend_new_client()
+ */
+void cryptodev_backend_free_client(
+ CryptoDevBackendClient *cc);
+
+/**
+ * cryptodev_backend_cleanup:
+ * @backend: the cryptodev backend object
+ * @errp: pointer to a NULL-initialized error object
+ *
+ * Clean the resouce associated with @backend that realizaed
+ * by the specific backend's init() callback
+ */
+void cryptodev_backend_cleanup(
+ CryptoDevBackend *backend,
+ Error **errp);
+
+#endif /* CRYPTODEV_H */
diff --git a/backends/cryptodev.c b/backends/cryptodev.c
new file mode 100644
index 0000000..e8582cd
--- /dev/null
+++ b/backends/cryptodev.c
@@ -0,0 +1,176 @@
+/*
+ * QEMU Crypto Device Implementation
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include "qemu/osdep.h"
+#include "sysemu/cryptodev.h"
+#include "hw/boards.h"
+#include "qapi/error.h"
+#include "qapi/visitor.h"
+#include "qapi-types.h"
+#include "qapi-visit.h"
+#include "qemu/config-file.h"
+#include "qom/object_interfaces.h"
+
+static QTAILQ_HEAD(, CryptoDevBackendClient) crypto_clients;
+
+
+CryptoDevBackendClient *
+cryptodev_backend_new_client(const char *model,
+ const char *name)
+{
+ CryptoDevBackendClient *cc;
+
+ cc = g_malloc0(sizeof(CryptoDevBackendClient));
+ cc->model = g_strdup(model);
+ if (name) {
+ cc->name = g_strdup(name);
+ }
+
+ QTAILQ_INSERT_TAIL(&crypto_clients, cc, next);
+
+ return cc;
+}
+
+void cryptodev_backend_free_client(
+ CryptoDevBackendClient *cc)
+{
+ QTAILQ_REMOVE(&crypto_clients, cc, next);
+ g_free(cc->name);
+ g_free(cc->model);
+ g_free(cc->info_str);
+ g_free(cc);
+}
+
+void cryptodev_backend_cleanup(
+ CryptoDevBackend *backend,
+ Error **errp)
+{
+ CryptoDevBackendClass *bc =
+ CRYPTODEV_BACKEND_GET_CLASS(backend);
+
+ if (bc->cleanup) {
+ bc->cleanup(backend, errp);
+ }
+
+ backend->ready = false;
+}
+
+static void
+cryptodev_backend_get_queues(Object *obj, Visitor *v, const char *name,
+ void *opaque, Error **errp)
+{
+ CryptoDevBackend *backend = CRYPTODEV_BACKEND(obj);
+ uint32_t value = backend->conf.peers.queues;
+
+ visit_type_uint32(v, name, &value, errp);
+}
+
+static void
+cryptodev_backend_set_queues(Object *obj, Visitor *v, const char *name,
+ void *opaque, Error **errp)
+{
+ CryptoDevBackend *backend = CRYPTODEV_BACKEND(obj);
+ Error *local_err = NULL;
+ uint32_t value;
+
+ visit_type_uint32(v, name, &value, &local_err);
+ if (local_err) {
+ goto out;
+ }
+ if (!value) {
+ error_setg(&local_err, "Property '%s.%s' doesn't take value '%"
+ PRIu32 "'", object_get_typename(obj), name, value);
+ goto out;
+ }
+ backend->conf.peers.queues = value;
+out:
+ error_propagate(errp, local_err);
+}
+
+static void
+cryptodev_backend_complete(UserCreatable *uc, Error **errp)
+{
+ CryptoDevBackend *backend = CRYPTODEV_BACKEND(uc);
+ CryptoDevBackendClass *bc = CRYPTODEV_BACKEND_GET_CLASS(uc);
+ Error *local_err = NULL;
+
+ if (bc->init) {
+ bc->init(backend, &local_err);
+ if (local_err) {
+ goto out;
+ }
+ }
+ backend->ready = true;
+ return;
+
+out:
+ backend->ready = false;
+ error_propagate(errp, local_err);
+}
+
+static void cryptodev_backend_instance_init(Object *obj)
+{
+ object_property_add(obj, "queues", "int",
+ cryptodev_backend_get_queues,
+ cryptodev_backend_set_queues,
+ NULL, NULL, NULL);
+ /* Initialize devices' queues property to 1 */
+ object_property_set_int(obj, 1, "queues", NULL);
+}
+
+static void cryptodev_backend_finalize(Object *obj)
+{
+
+}
+
+static void
+cryptodev_backend_class_init(ObjectClass *oc, void *data)
+{
+ UserCreatableClass *ucc = USER_CREATABLE_CLASS(oc);
+
+ ucc->complete = cryptodev_backend_complete;
+
+ QTAILQ_INIT(&crypto_clients);
+}
+
+static const TypeInfo cryptodev_backend_info = {
+ .name = TYPE_CRYPTODEV_BACKEND,
+ .parent = TYPE_OBJECT,
+ .instance_size = sizeof(CryptoDevBackend),
+ .instance_init = cryptodev_backend_instance_init,
+ .instance_finalize = cryptodev_backend_finalize,
+ .class_size = sizeof(CryptoDevBackendClass),
+ .class_init = cryptodev_backend_class_init,
+ .interfaces = (InterfaceInfo[]) {
+ { TYPE_USER_CREATABLE },
+ { }
+ }
+};
+
+static void
+cryptodev_backend_register_types(void)
+{
+ type_register_static(&cryptodev_backend_info);
+}
+
+type_init(cryptodev_backend_register_types);
diff --git a/backends/Makefile.objs b/backends/Makefile.objs
index 31a3a89..55bd43d 100644
--- a/backends/Makefile.objs
+++ b/backends/Makefile.objs
@@ -9,3 +9,5 @@ common-obj-$(CONFIG_TPM) += tpm.o
common-obj-y += hostmem.o hostmem-ram.o
common-obj-$(CONFIG_LINUX) += hostmem-file.o
+
+common-obj-y += cryptodev.o
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 17/47] cryptodev: add symmetric algorithm operation stuff
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (15 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 16/47] cryptodev: introduce cryptodev backend interface Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 18/47] virtio-crypto: introduce virtio_crypto.h Michael S. Tsirkin
` (32 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
This patch adds session operation and crypto operation
stuff in the cryptodev backend, including function
pointers and corresponding structures.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/sysemu/cryptodev.h | 149 +++++++++++++++++++++++++++++++++++++++++++++
backends/cryptodev.c | 45 ++++++++++++++
2 files changed, 194 insertions(+)
diff --git a/include/sysemu/cryptodev.h b/include/sysemu/cryptodev.h
index bfaf771..e66bd4b 100644
--- a/include/sysemu/cryptodev.h
+++ b/include/sysemu/cryptodev.h
@@ -56,12 +56,111 @@ typedef struct CryptoDevBackendClient
CryptoDevBackendClient;
typedef struct CryptoDevBackend CryptoDevBackend;
+enum CryptoDevBackendAlgType {
+ CRYPTODEV_BACKEND_ALG_SYM,
+ CRYPTODEV_BACKEND_ALG__MAX,
+};
+
+/**
+ * CryptoDevBackendSymSessionInfo:
+ *
+ * @op_code: operation code (refer to virtio_crypto.h)
+ * @cipher_alg: algorithm type of CIPHER
+ * @key_len: byte length of cipher key
+ * @hash_alg: algorithm type of HASH/MAC
+ * @hash_result_len: byte length of HASH operation result
+ * @auth_key_len: byte length of authenticated key
+ * @add_len: byte length of additional authenticated data
+ * @op_type: operation type (refer to virtio_crypto.h)
+ * @direction: encryption or direction for CIPHER
+ * @hash_mode: HASH mode for HASH operation (refer to virtio_crypto.h)
+ * @alg_chain_order: order of algorithm chaining (CIPHER then HASH,
+ * or HASH then CIPHER)
+ * @cipher_key: point to a key of CIPHER
+ * @auth_key: point to an authenticated key of MAC
+ *
+ */
+typedef struct CryptoDevBackendSymSessionInfo {
+ /* corresponding with virtio crypto spec */
+ uint32_t op_code;
+ uint32_t cipher_alg;
+ uint32_t key_len;
+ uint32_t hash_alg;
+ uint32_t hash_result_len;
+ uint32_t auth_key_len;
+ uint32_t add_len;
+ uint8_t op_type;
+ uint8_t direction;
+ uint8_t hash_mode;
+ uint8_t alg_chain_order;
+ uint8_t *cipher_key;
+ uint8_t *auth_key;
+} CryptoDevBackendSymSessionInfo;
+
+/**
+ * CryptoDevBackendSymOpInfo:
+ *
+ * @session_id: session index which was previously
+ * created by cryptodev_backend_sym_create_session()
+ * @aad_len: byte length of additional authenticated data
+ * @iv_len: byte length of initialization vector or counter
+ * @src_len: byte length of source data
+ * @dst_len: byte length of destination data
+ * @digest_result_len: byte length of hash digest result
+ * @hash_start_src_offset: Starting point for hash processing, specified
+ * as number of bytes from start of packet in source data, only used for
+ * algorithm chain
+ * @cipher_start_src_offset: Starting point for cipher processing, specified
+ * as number of bytes from start of packet in source data, only used for
+ * algorithm chain
+ * @len_to_hash: byte length of source data on which the hash
+ * operation will be computed, only used for algorithm chain
+ * @len_to_cipher: byte length of source data on which the cipher
+ * operation will be computed, only used for algorithm chain
+ * @op_type: operation type (refer to virtio_crypto.h)
+ * @iv: point to the initialization vector or counter
+ * @src: point to the source data
+ * @dst: point to the destination data
+ * @aad_data: point to the additional authenticated data
+ * @digest_result: point to the digest result data
+ * @data[0]: point to the extensional memory by one memory allocation
+ *
+ */
+typedef struct CryptoDevBackendSymOpInfo {
+ uint64_t session_id;
+ uint32_t aad_len;
+ uint32_t iv_len;
+ uint32_t src_len;
+ uint32_t dst_len;
+ uint32_t digest_result_len;
+ uint32_t hash_start_src_offset;
+ uint32_t cipher_start_src_offset;
+ uint32_t len_to_hash;
+ uint32_t len_to_cipher;
+ uint8_t op_type;
+ uint8_t *iv;
+ uint8_t *src;
+ uint8_t *dst;
+ uint8_t *aad_data;
+ uint8_t *digest_result;
+ uint8_t data[0];
+} CryptoDevBackendSymOpInfo;
typedef struct CryptoDevBackendClass {
ObjectClass parent_class;
void (*init)(CryptoDevBackend *backend, Error **errp);
void (*cleanup)(CryptoDevBackend *backend, Error **errp);
+
+ int64_t (*create_session)(CryptoDevBackend *backend,
+ CryptoDevBackendSymSessionInfo *sess_info,
+ uint32_t queue_index, Error **errp);
+ int (*close_session)(CryptoDevBackend *backend,
+ uint64_t session_id,
+ uint32_t queue_index, Error **errp);
+ int (*do_sym_op)(CryptoDevBackend *backend,
+ CryptoDevBackendSymOpInfo *op_info,
+ uint32_t queue_index, Error **errp);
} CryptoDevBackendClass;
@@ -145,4 +244,54 @@ void cryptodev_backend_cleanup(
CryptoDevBackend *backend,
Error **errp);
+/**
+ * cryptodev_backend_sym_create_session:
+ * @backend: the cryptodev backend object
+ * @sess_info: parameters needed by session creating
+ * @queue_index: queue index of cryptodev backend client
+ * @errp: pointer to a NULL-initialized error object
+ *
+ * Create a session for symmetric algorithms
+ *
+ * Returns: session id on success, or -1 on error
+ */
+int64_t cryptodev_backend_sym_create_session(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymSessionInfo *sess_info,
+ uint32_t queue_index, Error **errp);
+
+/**
+ * cryptodev_backend_sym_close_session:
+ * @backend: the cryptodev backend object
+ * @session_id: the session id
+ * @queue_index: queue index of cryptodev backend client
+ * @errp: pointer to a NULL-initialized error object
+ *
+ * Close a session for symmetric algorithms which was previously
+ * created by cryptodev_backend_sym_create_session()
+ *
+ * Returns: 0 on success, or Negative on error
+ */
+int cryptodev_backend_sym_close_session(
+ CryptoDevBackend *backend,
+ uint64_t session_id,
+ uint32_t queue_index, Error **errp);
+
+/**
+ * cryptodev_backend_sym_operation:
+ * @backend: the cryptodev backend object
+ * @op_info: parameters needed by symmetric crypto operation
+ * @queue_index: queue index of cryptodev backend client
+ * @errp: pointer to a NULL-initialized error object
+ *
+ * Do symmetric crypto operation, such as encryption and
+ * decryption
+ *
+ * Returns: 0 on success, or Negative on error
+ */
+int cryptodev_backend_sym_operation(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymOpInfo *op_info,
+ uint32_t queue_index, Error **errp);
+
#endif /* CRYPTODEV_H */
diff --git a/backends/cryptodev.c b/backends/cryptodev.c
index e8582cd..47521cf 100644
--- a/backends/cryptodev.c
+++ b/backends/cryptodev.c
@@ -75,6 +75,51 @@ void cryptodev_backend_cleanup(
backend->ready = false;
}
+int64_t cryptodev_backend_sym_create_session(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymSessionInfo *sess_info,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendClass *bc =
+ CRYPTODEV_BACKEND_GET_CLASS(backend);
+
+ if (bc->create_session) {
+ return bc->create_session(backend, sess_info, queue_index, errp);
+ }
+
+ return -1;
+}
+
+int cryptodev_backend_sym_close_session(
+ CryptoDevBackend *backend,
+ uint64_t session_id,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendClass *bc =
+ CRYPTODEV_BACKEND_GET_CLASS(backend);
+
+ if (bc->close_session) {
+ return bc->close_session(backend, session_id, queue_index, errp);
+ }
+
+ return -1;
+}
+
+int cryptodev_backend_sym_operation(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymOpInfo *op_info,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendClass *bc =
+ CRYPTODEV_BACKEND_GET_CLASS(backend);
+
+ if (bc->do_sym_op) {
+ return bc->do_sym_op(backend, op_info, queue_index, errp);
+ }
+
+ return -1;
+}
+
static void
cryptodev_backend_get_queues(Object *obj, Visitor *v, const char *name,
void *opaque, Error **errp)
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 18/47] virtio-crypto: introduce virtio_crypto.h
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (16 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 17/47] cryptodev: add symmetric algorithm operation stuff Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 19/47] cryptodev: introduce a new cryptodev backend Michael S. Tsirkin
` (31 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Introduce the virtio_crypto.h which follows
virtio-crypto specification.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/standard-headers/linux/virtio_crypto.h | 429 +++++++++++++++++++++++++
1 file changed, 429 insertions(+)
create mode 100644 include/standard-headers/linux/virtio_crypto.h
diff --git a/include/standard-headers/linux/virtio_crypto.h b/include/standard-headers/linux/virtio_crypto.h
new file mode 100644
index 0000000..82275a8
--- /dev/null
+++ b/include/standard-headers/linux/virtio_crypto.h
@@ -0,0 +1,429 @@
+#ifndef _LINUX_VIRTIO_CRYPTO_H
+#define _LINUX_VIRTIO_CRYPTO_H
+/* This header is BSD licensed so anyone can use the definitions to implement
+ * compatible drivers/servers.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. Neither the name of IBM nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS''
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL IBM OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE. */
+
+#include "standard-headers/linux/types.h"
+#include "standard-headers/linux/virtio_config.h"
+#include "standard-headers/linux/virtio_types.h"
+
+
+#define VIRTIO_CRYPTO_SERVICE_CIPHER 0
+#define VIRTIO_CRYPTO_SERVICE_HASH 1
+#define VIRTIO_CRYPTO_SERVICE_MAC 2
+#define VIRTIO_CRYPTO_SERVICE_AEAD 3
+
+#define VIRTIO_CRYPTO_OPCODE(service, op) (((service) << 8) | (op))
+
+struct virtio_crypto_ctrl_header {
+#define VIRTIO_CRYPTO_CIPHER_CREATE_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_CIPHER, 0x02)
+#define VIRTIO_CRYPTO_CIPHER_DESTROY_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_CIPHER, 0x03)
+#define VIRTIO_CRYPTO_HASH_CREATE_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_HASH, 0x02)
+#define VIRTIO_CRYPTO_HASH_DESTROY_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_HASH, 0x03)
+#define VIRTIO_CRYPTO_MAC_CREATE_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_MAC, 0x02)
+#define VIRTIO_CRYPTO_MAC_DESTROY_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_MAC, 0x03)
+#define VIRTIO_CRYPTO_AEAD_CREATE_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_AEAD, 0x02)
+#define VIRTIO_CRYPTO_AEAD_DESTROY_SESSION \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_AEAD, 0x03)
+ __virtio32 opcode;
+ __virtio32 algo;
+ __virtio32 flag;
+ /* data virtqueue id */
+ __virtio32 queue_id;
+};
+
+struct virtio_crypto_cipher_session_para {
+#define VIRTIO_CRYPTO_NO_CIPHER 0
+#define VIRTIO_CRYPTO_CIPHER_ARC4 1
+#define VIRTIO_CRYPTO_CIPHER_AES_ECB 2
+#define VIRTIO_CRYPTO_CIPHER_AES_CBC 3
+#define VIRTIO_CRYPTO_CIPHER_AES_CTR 4
+#define VIRTIO_CRYPTO_CIPHER_DES_ECB 5
+#define VIRTIO_CRYPTO_CIPHER_DES_CBC 6
+#define VIRTIO_CRYPTO_CIPHER_3DES_ECB 7
+#define VIRTIO_CRYPTO_CIPHER_3DES_CBC 8
+#define VIRTIO_CRYPTO_CIPHER_3DES_CTR 9
+#define VIRTIO_CRYPTO_CIPHER_KASUMI_F8 10
+#define VIRTIO_CRYPTO_CIPHER_SNOW3G_UEA2 11
+#define VIRTIO_CRYPTO_CIPHER_AES_F8 12
+#define VIRTIO_CRYPTO_CIPHER_AES_XTS 13
+#define VIRTIO_CRYPTO_CIPHER_ZUC_EEA3 14
+ __virtio32 algo;
+ /* length of key */
+ __virtio32 keylen;
+
+#define VIRTIO_CRYPTO_OP_ENCRYPT 1
+#define VIRTIO_CRYPTO_OP_DECRYPT 2
+ /* encrypt or decrypt */
+ __virtio32 op;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_session_input {
+ /* Device-writable part */
+ __virtio64 session_id;
+ __virtio32 status;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_cipher_session_req {
+ struct virtio_crypto_cipher_session_para para;
+};
+
+struct virtio_crypto_hash_session_para {
+#define VIRTIO_CRYPTO_NO_HASH 0
+#define VIRTIO_CRYPTO_HASH_MD5 1
+#define VIRTIO_CRYPTO_HASH_SHA1 2
+#define VIRTIO_CRYPTO_HASH_SHA_224 3
+#define VIRTIO_CRYPTO_HASH_SHA_256 4
+#define VIRTIO_CRYPTO_HASH_SHA_384 5
+#define VIRTIO_CRYPTO_HASH_SHA_512 6
+#define VIRTIO_CRYPTO_HASH_SHA3_224 7
+#define VIRTIO_CRYPTO_HASH_SHA3_256 8
+#define VIRTIO_CRYPTO_HASH_SHA3_384 9
+#define VIRTIO_CRYPTO_HASH_SHA3_512 10
+#define VIRTIO_CRYPTO_HASH_SHA3_SHAKE128 11
+#define VIRTIO_CRYPTO_HASH_SHA3_SHAKE256 12
+ __virtio32 algo;
+ /* hash result length */
+ __virtio32 hash_result_len;
+};
+
+struct virtio_crypto_hash_create_session_req {
+ struct virtio_crypto_hash_session_para para;
+};
+
+struct virtio_crypto_mac_session_para {
+#define VIRTIO_CRYPTO_NO_MAC 0
+#define VIRTIO_CRYPTO_MAC_HMAC_MD5 1
+#define VIRTIO_CRYPTO_MAC_HMAC_SHA1 2
+#define VIRTIO_CRYPTO_MAC_HMAC_SHA_224 3
+#define VIRTIO_CRYPTO_MAC_HMAC_SHA_256 4
+#define VIRTIO_CRYPTO_MAC_HMAC_SHA_384 5
+#define VIRTIO_CRYPTO_MAC_HMAC_SHA_512 6
+#define VIRTIO_CRYPTO_MAC_CMAC_3DES 25
+#define VIRTIO_CRYPTO_MAC_CMAC_AES 26
+#define VIRTIO_CRYPTO_MAC_KASUMI_F9 27
+#define VIRTIO_CRYPTO_MAC_SNOW3G_UIA2 28
+#define VIRTIO_CRYPTO_MAC_GMAC_AES 41
+#define VIRTIO_CRYPTO_MAC_GMAC_TWOFISH 42
+#define VIRTIO_CRYPTO_MAC_CBCMAC_AES 49
+#define VIRTIO_CRYPTO_MAC_CBCMAC_KASUMI_F9 50
+#define VIRTIO_CRYPTO_MAC_XCBC_AES 53
+ __virtio32 algo;
+ /* hash result length */
+ __virtio32 hash_result_len;
+ /* length of authenticated key */
+ __virtio32 auth_key_len;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_mac_create_session_req {
+ struct virtio_crypto_mac_session_para para;
+};
+
+struct virtio_crypto_aead_session_para {
+#define VIRTIO_CRYPTO_NO_AEAD 0
+#define VIRTIO_CRYPTO_AEAD_GCM 1
+#define VIRTIO_CRYPTO_AEAD_CCM 2
+#define VIRTIO_CRYPTO_AEAD_CHACHA20_POLY1305 3
+ __virtio32 algo;
+ /* length of key */
+ __virtio32 key_len;
+ /* digest result length */
+ __virtio32 digest_result_len;
+ /* length of the additional authenticated data (AAD) in bytes */
+ __virtio32 aad_len;
+ /* encrypt or decrypt, See above VIRTIO_CRYPTO_OP_* */
+ __virtio32 op;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_aead_create_session_req {
+ struct virtio_crypto_aead_session_para para;
+};
+
+struct virtio_crypto_alg_chain_session_para {
+#define VIRTIO_CRYPTO_SYM_ALG_CHAIN_ORDER_HASH_THEN_CIPHER 1
+#define VIRTIO_CRYPTO_SYM_ALG_CHAIN_ORDER_CIPHER_THEN_HASH 2
+ __virtio32 alg_chain_order;
+/* Plain hash */
+#define VIRTIO_CRYPTO_SYM_HASH_MODE_PLAIN 1
+/* Authenticated hash (mac) */
+#define VIRTIO_CRYPTO_SYM_HASH_MODE_AUTH 2
+/* Nested hash */
+#define VIRTIO_CRYPTO_SYM_HASH_MODE_NESTED 3
+ __virtio32 hash_mode;
+ struct virtio_crypto_cipher_session_para cipher_param;
+ union {
+ struct virtio_crypto_hash_session_para hash_param;
+ struct virtio_crypto_mac_session_para mac_param;
+ } u;
+ /* length of the additional authenticated data (AAD) in bytes */
+ __virtio32 aad_len;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_alg_chain_session_req {
+ struct virtio_crypto_alg_chain_session_para para;
+};
+
+struct virtio_crypto_sym_create_session_req {
+ union {
+ struct virtio_crypto_cipher_session_req cipher;
+ struct virtio_crypto_alg_chain_session_req chain;
+ } u;
+
+ /* Device-readable part */
+
+/* No operation */
+#define VIRTIO_CRYPTO_SYM_OP_NONE 0
+/* Cipher only operation on the data */
+#define VIRTIO_CRYPTO_SYM_OP_CIPHER 1
+/* Chain any cipher with any hash or mac operation. The order
+ depends on the value of alg_chain_order param */
+#define VIRTIO_CRYPTO_SYM_OP_ALGORITHM_CHAINING 2
+ __virtio32 op_type;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_destroy_session_req {
+ /* Device-readable part */
+ __virtio64 session_id;
+};
+
+/* The request of the control viritqueue's packet */
+struct virtio_crypto_op_ctrl_req {
+ struct virtio_crypto_ctrl_header header;
+
+ union {
+ struct virtio_crypto_sym_create_session_req sym_create_session;
+ struct virtio_crypto_hash_create_session_req hash_create_session;
+ struct virtio_crypto_mac_create_session_req mac_create_session;
+ struct virtio_crypto_aead_create_session_req aead_create_session;
+ struct virtio_crypto_destroy_session_req destroy_session;
+ } u;
+};
+
+struct virtio_crypto_op_header {
+#define VIRTIO_CRYPTO_CIPHER_ENCRYPT \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_CIPHER, 0x00)
+#define VIRTIO_CRYPTO_CIPHER_DECRYPT \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_CIPHER, 0x01)
+#define VIRTIO_CRYPTO_HASH \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_HASH, 0x00)
+#define VIRTIO_CRYPTO_MAC \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_MAC, 0x00)
+#define VIRTIO_CRYPTO_AEAD_ENCRYPT \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_AEAD, 0x00)
+#define VIRTIO_CRYPTO_AEAD_DECRYPT \
+ VIRTIO_CRYPTO_OPCODE(VIRTIO_CRYPTO_SERVICE_AEAD, 0x01)
+ __virtio32 opcode;
+ /* algo should be service-specific algorithms */
+ __virtio32 algo;
+ /* session_id should be service-specific algorithms */
+ __virtio64 session_id;
+ /* control flag to control the request */
+ __virtio32 flag;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_cipher_para {
+ /*
+ * Byte Length of valid IV/Counter
+ *
+ * - For block ciphers in CBC or F8 mode, or for Kasumi in F8 mode, or for
+ * SNOW3G in UEA2 mode, this is the length of the IV (which
+ * must be the same as the block length of the cipher).
+ * - For block ciphers in CTR mode, this is the length of the counter
+ * (which must be the same as the block length of the cipher).
+ * - For AES-XTS, this is the 128bit tweak, i, from IEEE Std 1619-2007.
+ *
+ * The IV/Counter will be updated after every partial cryptographic
+ * operation.
+ */
+ __virtio32 iv_len;
+ /* length of source data */
+ __virtio32 src_data_len;
+ /* length of dst data */
+ __virtio32 dst_data_len;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_hash_para {
+ /* length of source data */
+ __virtio32 src_data_len;
+ /* hash result length */
+ __virtio32 hash_result_len;
+};
+
+struct virtio_crypto_mac_para {
+ struct virtio_crypto_hash_para hash;
+};
+
+struct virtio_crypto_aead_para {
+ /*
+ * Byte Length of valid IV data pointed to by the below iv_addr
+ * parameter.
+ *
+ * - For GCM mode, this is either 12 (for 96-bit IVs) or 16, in which
+ * case iv_addr points to J0.
+ * - For CCM mode, this is the length of the nonce, which can be in the
+ * range 7 to 13 inclusive.
+ */
+ __virtio32 iv_len;
+ /* length of additional auth data */
+ __virtio32 aad_len;
+ /* length of source data */
+ __virtio32 src_data_len;
+ /* length of dst data */
+ __virtio32 dst_data_len;
+};
+
+struct virtio_crypto_cipher_data_req {
+ /* Device-readable part */
+ struct virtio_crypto_cipher_para para;
+};
+
+struct virtio_crypto_hash_data_req {
+ /* Device-readable part */
+ struct virtio_crypto_hash_para para;
+};
+
+struct virtio_crypto_mac_data_req {
+ /* Device-readable part */
+ struct virtio_crypto_mac_para para;
+};
+
+struct virtio_crypto_alg_chain_data_para {
+ __virtio32 iv_len;
+ /* Length of source data */
+ __virtio32 src_data_len;
+ /* Length of destination data */
+ __virtio32 dst_data_len;
+ /* Starting point for cipher processing in source data */
+ __virtio32 cipher_start_src_offset;
+ /* Length of the source data that the cipher will be computed on */
+ __virtio32 len_to_cipher;
+ /* Starting point for hash processing in source data */
+ __virtio32 hash_start_src_offset;
+ /* Length of the source data that the hash will be computed on */
+ __virtio32 len_to_hash;
+ /* Length of the additional auth data */
+ __virtio32 aad_len;
+ /* Length of the hash result */
+ __virtio32 hash_result_len;
+ __virtio32 reserved;
+};
+
+struct virtio_crypto_alg_chain_data_req {
+ /* Device-readable part */
+ struct virtio_crypto_alg_chain_data_para para;
+};
+
+struct virtio_crypto_sym_data_req {
+ union {
+ struct virtio_crypto_cipher_data_req cipher;
+ struct virtio_crypto_alg_chain_data_req chain;
+ } u;
+
+ /* See above VIRTIO_CRYPTO_SYM_OP_* */
+ __virtio32 op_type;
+ __virtio32 padding;
+};
+
+struct virtio_crypto_aead_data_req {
+ /* Device-readable part */
+ struct virtio_crypto_aead_para para;
+};
+
+/* The request of the data viritqueue's packet */
+struct virtio_crypto_op_data_req {
+ struct virtio_crypto_op_header header;
+
+ union {
+ struct virtio_crypto_sym_data_req sym_req;
+ struct virtio_crypto_hash_data_req hash_req;
+ struct virtio_crypto_mac_data_req mac_req;
+ struct virtio_crypto_aead_data_req aead_req;
+ } u;
+};
+
+#define VIRTIO_CRYPTO_OK 0
+#define VIRTIO_CRYPTO_ERR 1
+#define VIRTIO_CRYPTO_BADMSG 2
+#define VIRTIO_CRYPTO_NOTSUPP 3
+#define VIRTIO_CRYPTO_INVSESS 4 /* Invaild session id */
+
+/* The accelerator hardware is ready */
+#define VIRTIO_CRYPTO_S_HW_READY (1 << 0)
+#define VIRTIO_CRYPTO_S_STARTED (1 << 1)
+
+struct virtio_crypto_config {
+ /* See VIRTIO_CRYPTO_* above */
+ __virtio32 status;
+
+ /*
+ * Maximum number of data queue legal values are between 1 and 0x8000
+ */
+ __virtio32 max_dataqueues;
+
+ /* Specifies the services mask which the devcie support,
+ see VIRTIO_CRYPTO_SERVICE_* above */
+ __virtio32 crypto_services;
+
+ /* Detailed algorithms mask */
+ __virtio32 cipher_algo_l;
+ __virtio32 cipher_algo_h;
+ __virtio32 hash_algo;
+ __virtio32 mac_algo_l;
+ __virtio32 mac_algo_h;
+ __virtio32 aead_algo;
+
+ /* Maximum length of cipher key */
+ uint32_t max_cipher_key_len;
+ /* Maximum length of authenticated key */
+ uint32_t max_auth_key_len;
+
+ __virtio32 reserve;
+
+ /* The maximum size of per request's content */
+ __virtio64 max_size;
+};
+
+struct virtio_crypto_inhdr {
+ /* See VIRTIO_CRYPTO_* above */
+ uint8_t status;
+};
+
+#endif /* _LINUX_VIRTIO_CRYPTO_H */
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 19/47] cryptodev: introduce a new cryptodev backend
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (17 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 18/47] virtio-crypto: introduce virtio_crypto.h Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 20/47] virtio-crypto: add virtio crypto device emulation Michael S. Tsirkin
` (30 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
The new cryptodev backend named cryptodev-builtin,
which realized by QEMU cipher APIs. These APIs can
be backed by either nettle or gcrypt.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
backends/cryptodev-builtin.c | 361 +++++++++++++++++++++++++++++++++++++++++++
backends/Makefile.objs | 1 +
qemu-options.hx | 18 +++
3 files changed, 380 insertions(+)
create mode 100644 backends/cryptodev-builtin.c
diff --git a/backends/cryptodev-builtin.c b/backends/cryptodev-builtin.c
new file mode 100644
index 0000000..eda954b
--- /dev/null
+++ b/backends/cryptodev-builtin.c
@@ -0,0 +1,361 @@
+/*
+ * QEMU Cryptodev backend for QEMU cipher APIs
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include "qemu/osdep.h"
+#include "sysemu/cryptodev.h"
+#include "hw/boards.h"
+#include "qapi/error.h"
+#include "standard-headers/linux/virtio_crypto.h"
+#include "crypto/cipher.h"
+
+
+/**
+ * @TYPE_CRYPTODEV_BACKEND_BUILTIN:
+ * name of backend that uses QEMU cipher API
+ */
+#define TYPE_CRYPTODEV_BACKEND_BUILTIN "cryptodev-backend-builtin"
+
+#define CRYPTODEV_BACKEND_BUILTIN(obj) \
+ OBJECT_CHECK(CryptoDevBackendBuiltin, \
+ (obj), TYPE_CRYPTODEV_BACKEND_BUILTIN)
+
+typedef struct CryptoDevBackendBuiltin
+ CryptoDevBackendBuiltin;
+
+typedef struct CryptoDevBackendBuiltinSession {
+ QCryptoCipher *cipher;
+ uint8_t direction; /* encryption or decryption */
+ uint8_t type; /* cipher? hash? aead? */
+ QTAILQ_ENTRY(CryptoDevBackendBuiltinSession) next;
+} CryptoDevBackendBuiltinSession;
+
+/* Max number of symmetric sessions */
+#define MAX_NUM_SESSIONS 256
+
+#define CRYPTODEV_BUITLIN_MAX_AUTH_KEY_LEN 512
+#define CRYPTODEV_BUITLIN_MAX_CIPHER_KEY_LEN 64
+
+struct CryptoDevBackendBuiltin {
+ CryptoDevBackend parent_obj;
+
+ CryptoDevBackendBuiltinSession *sessions[MAX_NUM_SESSIONS];
+};
+
+static void cryptodev_builtin_init(
+ CryptoDevBackend *backend, Error **errp)
+{
+ /* Only support one queue */
+ int queues = backend->conf.peers.queues;
+ CryptoDevBackendClient *cc;
+
+ if (queues != 1) {
+ error_setg(errp,
+ "Only support one queue in cryptdov-builtin backend");
+ return;
+ }
+
+ cc = cryptodev_backend_new_client(
+ "cryptodev-builtin", NULL);
+ cc->info_str = g_strdup_printf("cryptodev-builtin0");
+ cc->queue_index = 0;
+ backend->conf.peers.ccs[0] = cc;
+
+ backend->conf.crypto_services =
+ 1u << VIRTIO_CRYPTO_SERVICE_CIPHER |
+ 1u << VIRTIO_CRYPTO_SERVICE_HASH |
+ 1u << VIRTIO_CRYPTO_SERVICE_MAC;
+ backend->conf.cipher_algo_l = 1u << VIRTIO_CRYPTO_CIPHER_AES_CBC;
+ backend->conf.hash_algo = 1u << VIRTIO_CRYPTO_HASH_SHA1;
+ /*
+ * Set the Maximum length of crypto request.
+ * Why this value? Just avoid to overflow when
+ * memory allocation for each crypto request.
+ */
+ backend->conf.max_size = LONG_MAX - sizeof(CryptoDevBackendSymOpInfo);
+ backend->conf.max_cipher_key_len = CRYPTODEV_BUITLIN_MAX_CIPHER_KEY_LEN;
+ backend->conf.max_auth_key_len = CRYPTODEV_BUITLIN_MAX_AUTH_KEY_LEN;
+}
+
+static int
+cryptodev_builtin_get_unused_session_index(
+ CryptoDevBackendBuiltin *builtin)
+{
+ size_t i;
+
+ for (i = 0; i < MAX_NUM_SESSIONS; i++) {
+ if (builtin->sessions[i] == NULL) {
+ return i;
+ }
+ }
+
+ return -1;
+}
+
+static int
+cryptodev_builtin_get_aes_algo(uint32_t key_len, Error **errp)
+{
+ int algo;
+
+ if (key_len == 128 / 8) {
+ algo = QCRYPTO_CIPHER_ALG_AES_128;
+ } else if (key_len == 192 / 8) {
+ algo = QCRYPTO_CIPHER_ALG_AES_192;
+ } else if (key_len == 256 / 8) {
+ algo = QCRYPTO_CIPHER_ALG_AES_256;
+ } else {
+ error_setg(errp, "Unsupported key length :%u", key_len);
+ return -1;
+ }
+
+ return algo;
+}
+
+static int cryptodev_builtin_create_cipher_session(
+ CryptoDevBackendBuiltin *builtin,
+ CryptoDevBackendSymSessionInfo *sess_info,
+ Error **errp)
+{
+ int algo;
+ int mode;
+ QCryptoCipher *cipher;
+ int index;
+ CryptoDevBackendBuiltinSession *sess;
+
+ if (sess_info->op_type != VIRTIO_CRYPTO_SYM_OP_CIPHER) {
+ error_setg(errp, "Unsupported optype :%u", sess_info->op_type);
+ return -1;
+ }
+
+ index = cryptodev_builtin_get_unused_session_index(builtin);
+ if (index < 0) {
+ error_setg(errp, "Total number of sessions created exceeds %u",
+ MAX_NUM_SESSIONS);
+ return -1;
+ }
+
+ switch (sess_info->cipher_alg) {
+ case VIRTIO_CRYPTO_CIPHER_AES_ECB:
+ algo = cryptodev_builtin_get_aes_algo(sess_info->key_len,
+ errp);
+ if (algo < 0) {
+ return -1;
+ }
+ mode = QCRYPTO_CIPHER_MODE_ECB;
+ break;
+ case VIRTIO_CRYPTO_CIPHER_AES_CBC:
+ algo = cryptodev_builtin_get_aes_algo(sess_info->key_len,
+ errp);
+ if (algo < 0) {
+ return -1;
+ }
+ mode = QCRYPTO_CIPHER_MODE_CBC;
+ break;
+ case VIRTIO_CRYPTO_CIPHER_AES_CTR:
+ algo = cryptodev_builtin_get_aes_algo(sess_info->key_len,
+ errp);
+ if (algo < 0) {
+ return -1;
+ }
+ mode = QCRYPTO_CIPHER_MODE_CTR;
+ break;
+ case VIRTIO_CRYPTO_CIPHER_DES_ECB:
+ algo = QCRYPTO_CIPHER_ALG_DES_RFB;
+ mode = QCRYPTO_CIPHER_MODE_ECB;
+ break;
+ default:
+ error_setg(errp, "Unsupported cipher alg :%u",
+ sess_info->cipher_alg);
+ return -1;
+ }
+
+ cipher = qcrypto_cipher_new(algo, mode,
+ sess_info->cipher_key,
+ sess_info->key_len,
+ errp);
+ if (!cipher) {
+ return -1;
+ }
+
+ sess = g_new0(CryptoDevBackendBuiltinSession, 1);
+ sess->cipher = cipher;
+ sess->direction = sess_info->direction;
+ sess->type = sess_info->op_type;
+
+ builtin->sessions[index] = sess;
+
+ return index;
+}
+
+static int64_t cryptodev_builtin_sym_create_session(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymSessionInfo *sess_info,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendBuiltin *builtin =
+ CRYPTODEV_BACKEND_BUILTIN(backend);
+ int64_t session_id = -1;
+ int ret;
+
+ switch (sess_info->op_code) {
+ case VIRTIO_CRYPTO_CIPHER_CREATE_SESSION:
+ ret = cryptodev_builtin_create_cipher_session(
+ builtin, sess_info, errp);
+ if (ret < 0) {
+ return ret;
+ } else {
+ session_id = ret;
+ }
+ break;
+ case VIRTIO_CRYPTO_HASH_CREATE_SESSION:
+ case VIRTIO_CRYPTO_MAC_CREATE_SESSION:
+ default:
+ error_setg(errp, "Unsupported opcode :%" PRIu32 "",
+ sess_info->op_code);
+ return -1;
+ }
+
+ return session_id;
+}
+
+static int cryptodev_builtin_sym_close_session(
+ CryptoDevBackend *backend,
+ uint64_t session_id,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendBuiltin *builtin =
+ CRYPTODEV_BACKEND_BUILTIN(backend);
+
+ if (session_id >= MAX_NUM_SESSIONS ||
+ builtin->sessions[session_id] == NULL) {
+ error_setg(errp, "Cannot find a valid session id: %" PRIu64 "",
+ session_id);
+ return -1;
+ }
+
+ qcrypto_cipher_free(builtin->sessions[session_id]->cipher);
+ g_free(builtin->sessions[session_id]);
+ builtin->sessions[session_id] = NULL;
+ return 0;
+}
+
+static int cryptodev_builtin_sym_operation(
+ CryptoDevBackend *backend,
+ CryptoDevBackendSymOpInfo *op_info,
+ uint32_t queue_index, Error **errp)
+{
+ CryptoDevBackendBuiltin *builtin =
+ CRYPTODEV_BACKEND_BUILTIN(backend);
+ CryptoDevBackendBuiltinSession *sess;
+ int ret;
+
+ if (op_info->session_id >= MAX_NUM_SESSIONS ||
+ builtin->sessions[op_info->session_id] == NULL) {
+ error_setg(errp, "Cannot find a valid session id: %" PRIu64 "",
+ op_info->session_id);
+ return -VIRTIO_CRYPTO_INVSESS;
+ }
+
+ if (op_info->op_type == VIRTIO_CRYPTO_SYM_OP_ALGORITHM_CHAINING) {
+ error_setg(errp,
+ "Algorithm chain is unsupported for cryptdoev-builtin");
+ return -VIRTIO_CRYPTO_NOTSUPP;
+ }
+
+ sess = builtin->sessions[op_info->session_id];
+
+ ret = qcrypto_cipher_setiv(sess->cipher, op_info->iv,
+ op_info->iv_len, errp);
+ if (ret < 0) {
+ return -VIRTIO_CRYPTO_ERR;
+ }
+
+ if (sess->direction == VIRTIO_CRYPTO_OP_ENCRYPT) {
+ ret = qcrypto_cipher_encrypt(sess->cipher, op_info->src,
+ op_info->dst, op_info->src_len, errp);
+ if (ret < 0) {
+ return -VIRTIO_CRYPTO_ERR;
+ }
+ } else {
+ ret = qcrypto_cipher_decrypt(sess->cipher, op_info->src,
+ op_info->dst, op_info->src_len, errp);
+ if (ret < 0) {
+ return -VIRTIO_CRYPTO_ERR;
+ }
+ }
+ return VIRTIO_CRYPTO_OK;
+}
+
+static void cryptodev_builtin_cleanup(
+ CryptoDevBackend *backend,
+ Error **errp)
+{
+ CryptoDevBackendBuiltin *builtin =
+ CRYPTODEV_BACKEND_BUILTIN(backend);
+ size_t i;
+ int queues = backend->conf.peers.queues;
+ CryptoDevBackendClient *cc;
+
+ for (i = 0; i < MAX_NUM_SESSIONS; i++) {
+ if (builtin->sessions[i] != NULL) {
+ cryptodev_builtin_sym_close_session(
+ backend, i, 0, errp);
+ }
+ }
+
+ assert(queues == 1);
+
+ for (i = 0; i < queues; i++) {
+ cc = backend->conf.peers.ccs[i];
+ if (cc) {
+ cryptodev_backend_free_client(cc);
+ backend->conf.peers.ccs[i] = NULL;
+ }
+ }
+}
+
+static void
+cryptodev_builtin_class_init(ObjectClass *oc, void *data)
+{
+ CryptoDevBackendClass *bc = CRYPTODEV_BACKEND_CLASS(oc);
+
+ bc->init = cryptodev_builtin_init;
+ bc->cleanup = cryptodev_builtin_cleanup;
+ bc->create_session = cryptodev_builtin_sym_create_session;
+ bc->close_session = cryptodev_builtin_sym_close_session;
+ bc->do_sym_op = cryptodev_builtin_sym_operation;
+}
+
+static const TypeInfo cryptodev_builtin_info = {
+ .name = TYPE_CRYPTODEV_BACKEND_BUILTIN,
+ .parent = TYPE_CRYPTODEV_BACKEND,
+ .class_init = cryptodev_builtin_class_init,
+ .instance_size = sizeof(CryptoDevBackendBuiltin),
+};
+
+static void
+cryptodev_builtin_register_types(void)
+{
+ type_register_static(&cryptodev_builtin_info);
+}
+
+type_init(cryptodev_builtin_register_types);
diff --git a/backends/Makefile.objs b/backends/Makefile.objs
index 55bd43d..1846998 100644
--- a/backends/Makefile.objs
+++ b/backends/Makefile.objs
@@ -11,3 +11,4 @@ common-obj-y += hostmem.o hostmem-ram.o
common-obj-$(CONFIG_LINUX) += hostmem-file.o
common-obj-y += cryptodev.o
+common-obj-y += cryptodev-builtin.o
diff --git a/qemu-options.hx b/qemu-options.hx
index b1fbdb0..66f1d28 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -3948,6 +3948,24 @@ secondary:
If you want to know the detail of above command line, you can read
the colo-compare git log.
+@item -object cryptodev-backend-builtin,id=@var{id}[,queues=@var{queues}]
+
+Creates a cryptodev backend which executes crypto opreation from
+the QEMU cipher APIS. The @var{id} parameter is
+a unique ID that will be used to reference this cryptodev backend from
+the @option{virtio-crypto} device. The @var{queues} parameter is optional,
+which specify the queue number of cryptodev backend, the default of
+@var{queues} is 1.
+
+@example
+
+ # qemu-system-x86_64 \
+ [...] \
+ -object cryptodev-backend-builtin,id=cryptodev0 \
+ -device virtio-crypto-pci,id=crypto0,cryptodev=cryptodev0 \
+ [...]
+@end example
+
@item -object secret,id=@var{id},data=@var{string},format=@var{raw|base64}[,keyid=@var{secretid},iv=@var{string}]
@item -object secret,id=@var{id},file=@var{filename},format=@var{raw|base64}[,keyid=@var{secretid},iv=@var{string}]
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 20/47] virtio-crypto: add virtio crypto device emulation
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (18 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 19/47] cryptodev: introduce a new cryptodev backend Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 21/47] virtio-crypto-pci: add virtio crypto pci support Michael S. Tsirkin
` (29 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Introduce the virtio crypto realization, I'll
finish the core code in the following patches. The
thoughts came from virtio net realization.
For more information see:
http://qemu-project.org/Features/VirtioCrypto
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-crypto.h | 73 ++++++++++++++
include/standard-headers/linux/virtio_ids.h | 2 +-
hw/virtio/virtio-crypto.c | 151 ++++++++++++++++++++++++++++
hw/virtio/Makefile.objs | 1 +
4 files changed, 226 insertions(+), 1 deletion(-)
create mode 100644 include/hw/virtio/virtio-crypto.h
create mode 100644 hw/virtio/virtio-crypto.c
diff --git a/include/hw/virtio/virtio-crypto.h b/include/hw/virtio/virtio-crypto.h
new file mode 100644
index 0000000..4652c21
--- /dev/null
+++ b/include/hw/virtio/virtio-crypto.h
@@ -0,0 +1,73 @@
+/*
+ * Virtio crypto Support
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or
+ * (at your option) any later version. See the COPYING file in the
+ * top-level directory.
+ */
+
+#ifndef _QEMU_VIRTIO_CRYPTO_H
+#define _QEMU_VIRTIO_CRYPTO_H
+
+#include "standard-headers/linux/virtio_crypto.h"
+#include "hw/virtio/virtio.h"
+#include "sysemu/iothread.h"
+#include "sysemu/cryptodev.h"
+
+
+#define DEBUG_VIRTIO_CRYPTO 0
+
+#define DPRINTF(fmt, ...) \
+do { \
+ if (DEBUG_VIRTIO_CRYPTO) { \
+ fprintf(stderr, "virtio_crypto: " fmt, ##__VA_ARGS__); \
+ } \
+} while (0)
+
+
+#define TYPE_VIRTIO_CRYPTO "virtio-crypto-device"
+#define VIRTIO_CRYPTO(obj) \
+ OBJECT_CHECK(VirtIOCrypto, (obj), TYPE_VIRTIO_CRYPTO)
+#define VIRTIO_CRYPTO_GET_PARENT_CLASS(obj) \
+ OBJECT_GET_PARENT_CLASS(obj, TYPE_VIRTIO_CRYPTO)
+
+
+typedef struct VirtIOCryptoConf {
+ CryptoDevBackend *cryptodev;
+} VirtIOCryptoConf;
+
+struct VirtIOCrypto;
+
+typedef struct VirtIOCryptoReq {
+ VirtQueueElement elem;
+ /* flags of operation, such as type of algorithm */
+ uint32_t flags;
+ VirtQueue *vq;
+ struct VirtIOCrypto *vcrypto;
+ union {
+ CryptoDevBackendSymOpInfo *sym_op_info;
+ } u;
+} VirtIOCryptoReq;
+
+typedef struct VirtIOCrypto {
+ VirtIODevice parent_obj;
+
+ VirtQueue *ctrl_vq;
+
+ VirtIOCryptoConf conf;
+ CryptoDevBackend *cryptodev;
+
+ uint32_t max_queues;
+ uint32_t status;
+
+ int multiqueue;
+ uint32_t curr_queues;
+ size_t config_size;
+} VirtIOCrypto;
+
+#endif /* _QEMU_VIRTIO_CRYPTO_H */
diff --git a/include/standard-headers/linux/virtio_ids.h b/include/standard-headers/linux/virtio_ids.h
index 3228d58..fe74e42 100644
--- a/include/standard-headers/linux/virtio_ids.h
+++ b/include/standard-headers/linux/virtio_ids.h
@@ -42,5 +42,5 @@
#define VIRTIO_ID_GPU 16 /* virtio GPU */
#define VIRTIO_ID_INPUT 18 /* virtio input */
#define VIRTIO_ID_VSOCK 19 /* virtio vsock transport */
-
+#define VIRTIO_ID_CRYPTO 20 /* virtio crypto */
#endif /* _LINUX_VIRTIO_IDS_H */
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
new file mode 100644
index 0000000..109a504
--- /dev/null
+++ b/hw/virtio/virtio-crypto.c
@@ -0,0 +1,151 @@
+/*
+ * Virtio crypto Support
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or
+ * (at your option) any later version. See the COPYING file in the
+ * top-level directory.
+ */
+#include "qemu/osdep.h"
+#include "qemu/iov.h"
+#include "hw/qdev.h"
+#include "qapi/error.h"
+#include "qemu/error-report.h"
+
+#include "hw/virtio/virtio.h"
+#include "hw/virtio/virtio-crypto.h"
+#include "hw/virtio/virtio-access.h"
+#include "standard-headers/linux/virtio_ids.h"
+
+#define VIRTIO_CRYPTO_VM_VERSION 1
+
+static uint64_t virtio_crypto_get_features(VirtIODevice *vdev,
+ uint64_t features,
+ Error **errp)
+{
+ return features;
+}
+
+static void virtio_crypto_reset(VirtIODevice *vdev)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ /* multiqueue is disabled by default */
+ vcrypto->curr_queues = 1;
+ if (!vcrypto->cryptodev->ready) {
+ vcrypto->status &= ~VIRTIO_CRYPTO_S_HW_READY;
+ } else {
+ vcrypto->status |= VIRTIO_CRYPTO_S_HW_READY;
+ }
+}
+
+static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
+{
+ VirtIODevice *vdev = VIRTIO_DEVICE(dev);
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(dev);
+ int i;
+
+ vcrypto->cryptodev = vcrypto->conf.cryptodev;
+ if (vcrypto->cryptodev == NULL) {
+ error_setg(errp, "'cryptodev' parameter expects a valid object");
+ return;
+ }
+
+ vcrypto->max_queues = MAX(vcrypto->cryptodev->conf.peers.queues, 1);
+ if (vcrypto->max_queues + 1 > VIRTIO_QUEUE_MAX) {
+ error_setg(errp, "Invalid number of queues (= %" PRIu16 "), "
+ "must be a postive integer less than %d.",
+ vcrypto->max_queues, VIRTIO_QUEUE_MAX);
+ return;
+ }
+
+ virtio_init(vdev, "virtio-crypto", VIRTIO_ID_CRYPTO, vcrypto->config_size);
+ vcrypto->curr_queues = 1;
+
+ for (i = 0; i < vcrypto->max_queues; i++) {
+ virtio_add_queue(vdev, 1024, NULL);
+ }
+
+ vcrypto->ctrl_vq = virtio_add_queue(vdev, 64, NULL);
+ if (!vcrypto->cryptodev->ready) {
+ vcrypto->status &= ~VIRTIO_CRYPTO_S_HW_READY;
+ } else {
+ vcrypto->status |= VIRTIO_CRYPTO_S_HW_READY;
+ }
+}
+
+static void virtio_crypto_device_unrealize(DeviceState *dev, Error **errp)
+{
+ VirtIODevice *vdev = VIRTIO_DEVICE(dev);
+
+ virtio_cleanup(vdev);
+}
+
+static const VMStateDescription vmstate_virtio_crypto = {
+ .name = "virtio-crypto",
+ .minimum_version_id = VIRTIO_CRYPTO_VM_VERSION,
+ .version_id = VIRTIO_CRYPTO_VM_VERSION,
+ .fields = (VMStateField[]) {
+ VMSTATE_VIRTIO_DEVICE,
+ VMSTATE_END_OF_LIST()
+ },
+};
+
+static Property virtio_crypto_properties[] = {
+ DEFINE_PROP_END_OF_LIST(),
+};
+
+static void virtio_crypto_get_config(VirtIODevice *vdev, uint8_t *config)
+{
+
+}
+
+static void virtio_crypto_class_init(ObjectClass *klass, void *data)
+{
+ DeviceClass *dc = DEVICE_CLASS(klass);
+ VirtioDeviceClass *vdc = VIRTIO_DEVICE_CLASS(klass);
+
+ dc->props = virtio_crypto_properties;
+ dc->vmsd = &vmstate_virtio_crypto;
+ set_bit(DEVICE_CATEGORY_MISC, dc->categories);
+ vdc->realize = virtio_crypto_device_realize;
+ vdc->unrealize = virtio_crypto_device_unrealize;
+ vdc->get_config = virtio_crypto_get_config;
+ vdc->get_features = virtio_crypto_get_features;
+ vdc->reset = virtio_crypto_reset;
+}
+
+static void virtio_crypto_instance_init(Object *obj)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(obj);
+
+ /*
+ * The default config_size is sizeof(struct virtio_crypto_config).
+ * Can be overriden with virtio_crypto_set_config_size.
+ */
+ vcrypto->config_size = sizeof(struct virtio_crypto_config);
+
+ object_property_add_link(obj, "cryptodev",
+ TYPE_CRYPTODEV_BACKEND,
+ (Object **)&vcrypto->conf.cryptodev,
+ qdev_prop_allow_set_link_before_realize,
+ OBJ_PROP_LINK_UNREF_ON_RELEASE, NULL);
+}
+
+static const TypeInfo virtio_crypto_info = {
+ .name = TYPE_VIRTIO_CRYPTO,
+ .parent = TYPE_VIRTIO_DEVICE,
+ .instance_size = sizeof(VirtIOCrypto),
+ .instance_init = virtio_crypto_instance_init,
+ .class_init = virtio_crypto_class_init,
+};
+
+static void virtio_register_types(void)
+{
+ type_register_static(&virtio_crypto_info);
+}
+
+type_init(virtio_register_types)
diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
index e716308..968f392 100644
--- a/hw/virtio/Makefile.objs
+++ b/hw/virtio/Makefile.objs
@@ -7,3 +7,4 @@ obj-y += virtio.o virtio-balloon.o
obj-$(CONFIG_LINUX) += vhost.o vhost-backend.o vhost-user.o
obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
+obj-y += virtio-crypto.o
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 21/47] virtio-crypto-pci: add virtio crypto pci support
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (19 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 20/47] virtio-crypto: add virtio crypto device emulation Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 22/47] virtio-crypto: set capacity of algorithms supported Michael S. Tsirkin
` (28 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
This patch adds virtio-crypto-pci, which is the pci proxy for the virtio
crypto device.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/virtio-pci.h | 15 +++++++++
hw/virtio/virtio-crypto-pci.c | 77 +++++++++++++++++++++++++++++++++++++++++++
hw/virtio/Makefile.objs | 1 +
3 files changed, 93 insertions(+)
create mode 100644 hw/virtio/virtio-crypto-pci.c
diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h
index 4d206c8..b2a996f 100644
--- a/hw/virtio/virtio-pci.h
+++ b/hw/virtio/virtio-pci.h
@@ -25,6 +25,8 @@
#include "hw/virtio/virtio-bus.h"
#include "hw/virtio/virtio-input.h"
#include "hw/virtio/virtio-gpu.h"
+#include "hw/virtio/virtio-crypto.h"
+
#ifdef CONFIG_VIRTFS
#include "hw/9pfs/virtio-9p.h"
#endif
@@ -48,6 +50,7 @@ typedef struct VirtIOInputHIDPCI VirtIOInputHIDPCI;
typedef struct VirtIOInputHostPCI VirtIOInputHostPCI;
typedef struct VirtIOGPUPCI VirtIOGPUPCI;
typedef struct VHostVSockPCI VHostVSockPCI;
+typedef struct VirtIOCryptoPCI VirtIOCryptoPCI;
/* virtio-pci-bus */
@@ -350,6 +353,18 @@ struct VHostVSockPCI {
};
#endif
+/*
+ * virtio-crypto-pci: This extends VirtioPCIProxy.
+ */
+#define TYPE_VIRTIO_CRYPTO_PCI "virtio-crypto-pci"
+#define VIRTIO_CRYPTO_PCI(obj) \
+ OBJECT_CHECK(VirtIOCryptoPCI, (obj), TYPE_VIRTIO_CRYPTO_PCI)
+
+struct VirtIOCryptoPCI {
+ VirtIOPCIProxy parent_obj;
+ VirtIOCrypto vdev;
+};
+
/* Virtio ABI version, if we increment this, we break the guest driver. */
#define VIRTIO_PCI_ABI_VERSION 0
diff --git a/hw/virtio/virtio-crypto-pci.c b/hw/virtio/virtio-crypto-pci.c
new file mode 100644
index 0000000..21d9984
--- /dev/null
+++ b/hw/virtio/virtio-crypto-pci.c
@@ -0,0 +1,77 @@
+/*
+ * Virtio crypto device
+ *
+ * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
+ *
+ * Authors:
+ * Gonglei <arei.gonglei@huawei.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or
+ * (at your option) any later version. See the COPYING file in the
+ * top-level directory.
+ *
+ */
+#include "qemu/osdep.h"
+#include "hw/pci/pci.h"
+#include "hw/virtio/virtio.h"
+#include "hw/virtio/virtio-bus.h"
+#include "hw/virtio/virtio-pci.h"
+#include "hw/virtio/virtio-crypto.h"
+#include "qapi/error.h"
+
+static Property virtio_crypto_pci_properties[] = {
+ DEFINE_PROP_BIT("ioeventfd", VirtIOPCIProxy, flags,
+ VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT, true),
+ DEFINE_PROP_UINT32("vectors", VirtIOPCIProxy, nvectors, 2),
+ DEFINE_PROP_END_OF_LIST(),
+};
+
+static void virtio_crypto_pci_realize(VirtIOPCIProxy *vpci_dev, Error **errp)
+{
+ VirtIOCryptoPCI *vcrypto = VIRTIO_CRYPTO_PCI(vpci_dev);
+ DeviceState *vdev = DEVICE(&vcrypto->vdev);
+
+ qdev_set_parent_bus(vdev, BUS(&vpci_dev->bus));
+ virtio_pci_force_virtio_1(vpci_dev);
+ object_property_set_bool(OBJECT(vdev), true, "realized", errp);
+ object_property_set_link(OBJECT(vcrypto),
+ OBJECT(vcrypto->vdev.conf.cryptodev), "cryptodev",
+ NULL);
+}
+
+static void virtio_crypto_pci_class_init(ObjectClass *klass, void *data)
+{
+ DeviceClass *dc = DEVICE_CLASS(klass);
+ VirtioPCIClass *k = VIRTIO_PCI_CLASS(klass);
+ PCIDeviceClass *pcidev_k = PCI_DEVICE_CLASS(klass);
+
+ k->realize = virtio_crypto_pci_realize;
+ set_bit(DEVICE_CATEGORY_MISC, dc->categories);
+ dc->props = virtio_crypto_pci_properties;
+
+ pcidev_k->class_id = PCI_CLASS_OTHERS;
+}
+
+static void virtio_crypto_initfn(Object *obj)
+{
+ VirtIOCryptoPCI *dev = VIRTIO_CRYPTO_PCI(obj);
+
+ virtio_instance_init_common(obj, &dev->vdev, sizeof(dev->vdev),
+ TYPE_VIRTIO_CRYPTO);
+ object_property_add_alias(obj, "cryptodev", OBJECT(&dev->vdev),
+ "cryptodev", &error_abort);
+}
+
+static const TypeInfo virtio_crypto_pci_info = {
+ .name = TYPE_VIRTIO_CRYPTO_PCI,
+ .parent = TYPE_VIRTIO_PCI,
+ .instance_size = sizeof(VirtIOCryptoPCI),
+ .instance_init = virtio_crypto_initfn,
+ .class_init = virtio_crypto_pci_class_init,
+};
+
+static void virtio_crypto_pci_register_types(void)
+{
+ type_register_static(&virtio_crypto_pci_info);
+}
+type_init(virtio_crypto_pci_register_types)
diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
index 968f392..95c4c30 100644
--- a/hw/virtio/Makefile.objs
+++ b/hw/virtio/Makefile.objs
@@ -8,3 +8,4 @@ obj-$(CONFIG_LINUX) += vhost.o vhost-backend.o vhost-user.o
obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o
obj-y += virtio-crypto.o
+obj-$(CONFIG_VIRTIO_PCI) += virtio-crypto-pci.o
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 22/47] virtio-crypto: set capacity of algorithms supported
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (20 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 21/47] virtio-crypto-pci: add virtio crypto pci support Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 23/47] virtio-crypto: add control queue handler Michael S. Tsirkin
` (27 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Expose the capacity of algorithms supported by
virtio crypto device to the frontend driver using
pci configuration space.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-crypto.h | 18 ++++++++++++++++
hw/virtio/virtio-crypto.c | 43 +++++++++++++++++++++++++++++++++++++++
2 files changed, 61 insertions(+)
diff --git a/include/hw/virtio/virtio-crypto.h b/include/hw/virtio/virtio-crypto.h
index 4652c21..783ea23 100644
--- a/include/hw/virtio/virtio-crypto.h
+++ b/include/hw/virtio/virtio-crypto.h
@@ -39,6 +39,24 @@ do { \
typedef struct VirtIOCryptoConf {
CryptoDevBackend *cryptodev;
+
+ /* Supported service mask */
+ uint32_t crypto_services;
+
+ /* Detailed algorithms mask */
+ uint32_t cipher_algo_l;
+ uint32_t cipher_algo_h;
+ uint32_t hash_algo;
+ uint32_t mac_algo_l;
+ uint32_t mac_algo_h;
+ uint32_t aead_algo;
+
+ /* Maximum length of cipher key */
+ uint32_t max_cipher_key_len;
+ /* Maximum length of authenticated key */
+ uint32_t max_auth_key_len;
+ /* Maximum size of each crypto request's content */
+ uint64_t max_size;
} VirtIOCryptoConf;
struct VirtIOCrypto;
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
index 109a504..4ded704 100644
--- a/hw/virtio/virtio-crypto.c
+++ b/hw/virtio/virtio-crypto.c
@@ -42,6 +42,27 @@ static void virtio_crypto_reset(VirtIODevice *vdev)
}
}
+static void virtio_crypto_init_config(VirtIODevice *vdev)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+
+ vcrypto->conf.crypto_services =
+ vcrypto->conf.cryptodev->conf.crypto_services;
+ vcrypto->conf.cipher_algo_l =
+ vcrypto->conf.cryptodev->conf.cipher_algo_l;
+ vcrypto->conf.cipher_algo_h =
+ vcrypto->conf.cryptodev->conf.cipher_algo_h;
+ vcrypto->conf.hash_algo = vcrypto->conf.cryptodev->conf.hash_algo;
+ vcrypto->conf.mac_algo_l = vcrypto->conf.cryptodev->conf.mac_algo_l;
+ vcrypto->conf.mac_algo_h = vcrypto->conf.cryptodev->conf.mac_algo_h;
+ vcrypto->conf.aead_algo = vcrypto->conf.cryptodev->conf.aead_algo;
+ vcrypto->conf.max_cipher_key_len =
+ vcrypto->conf.cryptodev->conf.max_cipher_key_len;
+ vcrypto->conf.max_auth_key_len =
+ vcrypto->conf.cryptodev->conf.max_auth_key_len;
+ vcrypto->conf.max_size = vcrypto->conf.cryptodev->conf.max_size;
+}
+
static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
{
VirtIODevice *vdev = VIRTIO_DEVICE(dev);
@@ -75,6 +96,8 @@ static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
} else {
vcrypto->status |= VIRTIO_CRYPTO_S_HW_READY;
}
+
+ virtio_crypto_init_config(vdev);
}
static void virtio_crypto_device_unrealize(DeviceState *dev, Error **errp)
@@ -100,7 +123,27 @@ static Property virtio_crypto_properties[] = {
static void virtio_crypto_get_config(VirtIODevice *vdev, uint8_t *config)
{
+ VirtIOCrypto *c = VIRTIO_CRYPTO(vdev);
+ struct virtio_crypto_config crypto_cfg;
+ /*
+ * Virtio-crypto device conforms to VIRTIO 1.0 which is always LE,
+ * so we can use LE accessors directly.
+ */
+ stl_le_p(&crypto_cfg.status, c->status);
+ stl_le_p(&crypto_cfg.max_dataqueues, c->max_queues);
+ stl_le_p(&crypto_cfg.crypto_services, c->conf.crypto_services);
+ stl_le_p(&crypto_cfg.cipher_algo_l, c->conf.cipher_algo_l);
+ stl_le_p(&crypto_cfg.cipher_algo_h, c->conf.cipher_algo_h);
+ stl_le_p(&crypto_cfg.hash_algo, c->conf.hash_algo);
+ stl_le_p(&crypto_cfg.mac_algo_l, c->conf.mac_algo_l);
+ stl_le_p(&crypto_cfg.mac_algo_h, c->conf.mac_algo_h);
+ stl_le_p(&crypto_cfg.aead_algo, c->conf.aead_algo);
+ stl_le_p(&crypto_cfg.max_cipher_key_len, c->conf.max_cipher_key_len);
+ stl_le_p(&crypto_cfg.max_auth_key_len, c->conf.max_auth_key_len);
+ stq_le_p(&crypto_cfg.max_size, c->conf.max_size);
+
+ memcpy(config, &crypto_cfg, c->config_size);
}
static void virtio_crypto_class_init(ObjectClass *klass, void *data)
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 23/47] virtio-crypto: add control queue handler
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (21 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 22/47] virtio-crypto: set capacity of algorithms supported Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 24/47] virtio-crypto: add data queue processing handler Michael S. Tsirkin
` (26 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Realize the symmetric algorithm control queue handler,
including plain cipher and chainning algorithms.
Currently the control queue is used to create and
close session for symmetric algorithm.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/virtio-crypto.c | 299 +++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 298 insertions(+), 1 deletion(-)
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
index 4ded704..195a266 100644
--- a/hw/virtio/virtio-crypto.c
+++ b/hw/virtio/virtio-crypto.c
@@ -23,6 +23,303 @@
#define VIRTIO_CRYPTO_VM_VERSION 1
+/*
+ * Transfer virtqueue index to crypto queue index.
+ * The control virtqueue is after the data virtqueues
+ * so the input value doesn't need to be adjusted
+ */
+static inline int virtio_crypto_vq2q(int queue_index)
+{
+ return queue_index;
+}
+
+static int
+virtio_crypto_cipher_session_helper(VirtIODevice *vdev,
+ CryptoDevBackendSymSessionInfo *info,
+ struct virtio_crypto_cipher_session_para *cipher_para,
+ struct iovec **iov, unsigned int *out_num)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ unsigned int num = *out_num;
+
+ info->cipher_alg = ldl_le_p(&cipher_para->algo);
+ info->key_len = ldl_le_p(&cipher_para->keylen);
+ info->direction = ldl_le_p(&cipher_para->op);
+ DPRINTF("cipher_alg=%" PRIu32 ", info->direction=%" PRIu32 "\n",
+ info->cipher_alg, info->direction);
+
+ if (info->key_len > vcrypto->conf.max_cipher_key_len) {
+ error_report("virtio-crypto length of cipher key is too big: %u",
+ info->key_len);
+ return -VIRTIO_CRYPTO_ERR;
+ }
+ /* Get cipher key */
+ if (info->key_len > 0) {
+ size_t s;
+ DPRINTF("keylen=%" PRIu32 "\n", info->key_len);
+
+ info->cipher_key = g_malloc(info->key_len);
+ s = iov_to_buf(*iov, num, 0, info->cipher_key, info->key_len);
+ if (unlikely(s != info->key_len)) {
+ virtio_error(vdev, "virtio-crypto cipher key incorrect");
+ return -EFAULT;
+ }
+ iov_discard_front(iov, &num, info->key_len);
+ *out_num = num;
+ }
+
+ return 0;
+}
+
+static int64_t
+virtio_crypto_create_sym_session(VirtIOCrypto *vcrypto,
+ struct virtio_crypto_sym_create_session_req *sess_req,
+ uint32_t queue_id,
+ uint32_t opcode,
+ struct iovec *iov, unsigned int out_num)
+{
+ VirtIODevice *vdev = VIRTIO_DEVICE(vcrypto);
+ CryptoDevBackendSymSessionInfo info;
+ int64_t session_id;
+ int queue_index;
+ uint32_t op_type;
+ Error *local_err = NULL;
+ int ret;
+
+ memset(&info, 0, sizeof(info));
+ op_type = ldl_le_p(&sess_req->op_type);
+ info.op_type = op_type;
+ info.op_code = opcode;
+
+ if (op_type == VIRTIO_CRYPTO_SYM_OP_CIPHER) {
+ ret = virtio_crypto_cipher_session_helper(vdev, &info,
+ &sess_req->u.cipher.para,
+ &iov, &out_num);
+ if (ret < 0) {
+ goto err;
+ }
+ } else if (op_type == VIRTIO_CRYPTO_SYM_OP_ALGORITHM_CHAINING) {
+ size_t s;
+ /* cipher part */
+ ret = virtio_crypto_cipher_session_helper(vdev, &info,
+ &sess_req->u.chain.para.cipher_param,
+ &iov, &out_num);
+ if (ret < 0) {
+ goto err;
+ }
+ /* hash part */
+ info.alg_chain_order = ldl_le_p(
+ &sess_req->u.chain.para.alg_chain_order);
+ info.add_len = ldl_le_p(&sess_req->u.chain.para.aad_len);
+ info.hash_mode = ldl_le_p(&sess_req->u.chain.para.hash_mode);
+ if (info.hash_mode == VIRTIO_CRYPTO_SYM_HASH_MODE_AUTH) {
+ info.hash_alg = ldl_le_p(&sess_req->u.chain.para.u.mac_param.algo);
+ info.auth_key_len = ldl_le_p(
+ &sess_req->u.chain.para.u.mac_param.auth_key_len);
+ info.hash_result_len = ldl_le_p(
+ &sess_req->u.chain.para.u.mac_param.hash_result_len);
+ if (info.auth_key_len > vcrypto->conf.max_auth_key_len) {
+ error_report("virtio-crypto length of auth key is too big: %u",
+ info.auth_key_len);
+ ret = -VIRTIO_CRYPTO_ERR;
+ goto err;
+ }
+ /* get auth key */
+ if (info.auth_key_len > 0) {
+ DPRINTF("auth_keylen=%" PRIu32 "\n", info.auth_key_len);
+ info.auth_key = g_malloc(info.auth_key_len);
+ s = iov_to_buf(iov, out_num, 0, info.auth_key,
+ info.auth_key_len);
+ if (unlikely(s != info.auth_key_len)) {
+ virtio_error(vdev,
+ "virtio-crypto authenticated key incorrect");
+ ret = -EFAULT;
+ goto err;
+ }
+ iov_discard_front(&iov, &out_num, info.auth_key_len);
+ }
+ } else if (info.hash_mode == VIRTIO_CRYPTO_SYM_HASH_MODE_PLAIN) {
+ info.hash_alg = ldl_le_p(
+ &sess_req->u.chain.para.u.hash_param.algo);
+ info.hash_result_len = ldl_le_p(
+ &sess_req->u.chain.para.u.hash_param.hash_result_len);
+ } else {
+ /* VIRTIO_CRYPTO_SYM_HASH_MODE_NESTED */
+ error_report("unsupported hash mode");
+ ret = -VIRTIO_CRYPTO_NOTSUPP;
+ goto err;
+ }
+ } else {
+ /* VIRTIO_CRYPTO_SYM_OP_NONE */
+ error_report("unsupported cipher op_type: VIRTIO_CRYPTO_SYM_OP_NONE");
+ ret = -VIRTIO_CRYPTO_NOTSUPP;
+ goto err;
+ }
+
+ queue_index = virtio_crypto_vq2q(queue_id);
+ session_id = cryptodev_backend_sym_create_session(
+ vcrypto->cryptodev,
+ &info, queue_index, &local_err);
+ if (session_id >= 0) {
+ DPRINTF("create session_id=%" PRIu64 " successfully\n",
+ session_id);
+
+ ret = session_id;
+ } else {
+ if (local_err) {
+ error_report_err(local_err);
+ }
+ ret = -VIRTIO_CRYPTO_ERR;
+ }
+
+err:
+ g_free(info.cipher_key);
+ g_free(info.auth_key);
+ return ret;
+}
+
+static uint8_t
+virtio_crypto_handle_close_session(VirtIOCrypto *vcrypto,
+ struct virtio_crypto_destroy_session_req *close_sess_req,
+ uint32_t queue_id)
+{
+ int ret;
+ uint64_t session_id;
+ uint32_t status;
+ Error *local_err = NULL;
+
+ session_id = ldq_le_p(&close_sess_req->session_id);
+ DPRINTF("close session, id=%" PRIu64 "\n", session_id);
+
+ ret = cryptodev_backend_sym_close_session(
+ vcrypto->cryptodev, session_id, queue_id, &local_err);
+ if (ret == 0) {
+ status = VIRTIO_CRYPTO_OK;
+ } else {
+ if (local_err) {
+ error_report_err(local_err);
+ } else {
+ error_report("destroy session failed");
+ }
+ status = VIRTIO_CRYPTO_ERR;
+ }
+
+ return status;
+}
+
+static void virtio_crypto_handle_ctrl(VirtIODevice *vdev, VirtQueue *vq)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ struct virtio_crypto_op_ctrl_req ctrl;
+ VirtQueueElement *elem;
+ struct iovec *in_iov;
+ struct iovec *out_iov;
+ unsigned in_num;
+ unsigned out_num;
+ uint32_t queue_id;
+ uint32_t opcode;
+ struct virtio_crypto_session_input input;
+ int64_t session_id;
+ uint8_t status;
+ size_t s;
+
+ for (;;) {
+ elem = virtqueue_pop(vq, sizeof(VirtQueueElement));
+ if (!elem) {
+ break;
+ }
+ if (elem->out_num < 1 || elem->in_num < 1) {
+ virtio_error(vdev, "virtio-crypto ctrl missing headers");
+ virtqueue_detach_element(vq, elem, 0);
+ g_free(elem);
+ break;
+ }
+
+ out_num = elem->out_num;
+ out_iov = elem->out_sg;
+ in_num = elem->in_num;
+ in_iov = elem->in_sg;
+ if (unlikely(iov_to_buf(out_iov, out_num, 0, &ctrl, sizeof(ctrl))
+ != sizeof(ctrl))) {
+ virtio_error(vdev, "virtio-crypto request ctrl_hdr too short");
+ virtqueue_detach_element(vq, elem, 0);
+ g_free(elem);
+ break;
+ }
+ iov_discard_front(&out_iov, &out_num, sizeof(ctrl));
+
+ opcode = ldl_le_p(&ctrl.header.opcode);
+ queue_id = ldl_le_p(&ctrl.header.queue_id);
+
+ switch (opcode) {
+ case VIRTIO_CRYPTO_CIPHER_CREATE_SESSION:
+ memset(&input, 0, sizeof(input));
+ session_id = virtio_crypto_create_sym_session(vcrypto,
+ &ctrl.u.sym_create_session,
+ queue_id, opcode,
+ out_iov, out_num);
+ /* Serious errors, need to reset virtio crypto device */
+ if (session_id == -EFAULT) {
+ virtqueue_detach_element(vq, elem, 0);
+ break;
+ } else if (session_id == -VIRTIO_CRYPTO_NOTSUPP) {
+ stl_le_p(&input.status, VIRTIO_CRYPTO_NOTSUPP);
+ } else if (session_id == -VIRTIO_CRYPTO_ERR) {
+ stl_le_p(&input.status, VIRTIO_CRYPTO_ERR);
+ } else {
+ /* Set the session id */
+ stq_le_p(&input.session_id, session_id);
+ stl_le_p(&input.status, VIRTIO_CRYPTO_OK);
+ }
+
+ s = iov_from_buf(in_iov, in_num, 0, &input, sizeof(input));
+ if (unlikely(s != sizeof(input))) {
+ virtio_error(vdev, "virtio-crypto input incorrect");
+ virtqueue_detach_element(vq, elem, 0);
+ break;
+ }
+ virtqueue_push(vq, elem, sizeof(input));
+ virtio_notify(vdev, vq);
+ break;
+ case VIRTIO_CRYPTO_CIPHER_DESTROY_SESSION:
+ case VIRTIO_CRYPTO_HASH_DESTROY_SESSION:
+ case VIRTIO_CRYPTO_MAC_DESTROY_SESSION:
+ case VIRTIO_CRYPTO_AEAD_DESTROY_SESSION:
+ status = virtio_crypto_handle_close_session(vcrypto,
+ &ctrl.u.destroy_session, queue_id);
+ /* The status only occupy one byte, we can directly use it */
+ s = iov_from_buf(in_iov, in_num, 0, &status, sizeof(status));
+ if (unlikely(s != sizeof(status))) {
+ virtio_error(vdev, "virtio-crypto status incorrect");
+ virtqueue_detach_element(vq, elem, 0);
+ break;
+ }
+ virtqueue_push(vq, elem, sizeof(status));
+ virtio_notify(vdev, vq);
+ break;
+ case VIRTIO_CRYPTO_HASH_CREATE_SESSION:
+ case VIRTIO_CRYPTO_MAC_CREATE_SESSION:
+ case VIRTIO_CRYPTO_AEAD_CREATE_SESSION:
+ default:
+ error_report("virtio-crypto unsupported ctrl opcode: %d", opcode);
+ memset(&input, 0, sizeof(input));
+ stl_le_p(&input.status, VIRTIO_CRYPTO_NOTSUPP);
+ s = iov_from_buf(in_iov, in_num, 0, &input, sizeof(input));
+ if (unlikely(s != sizeof(input))) {
+ virtio_error(vdev, "virtio-crypto input incorrect");
+ virtqueue_detach_element(vq, elem, 0);
+ break;
+ }
+ virtqueue_push(vq, elem, sizeof(input));
+ virtio_notify(vdev, vq);
+
+ break;
+ } /* end switch case */
+
+ g_free(elem);
+ } /* end for loop */
+}
+
static uint64_t virtio_crypto_get_features(VirtIODevice *vdev,
uint64_t features,
Error **errp)
@@ -90,7 +387,7 @@ static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
virtio_add_queue(vdev, 1024, NULL);
}
- vcrypto->ctrl_vq = virtio_add_queue(vdev, 64, NULL);
+ vcrypto->ctrl_vq = virtio_add_queue(vdev, 64, virtio_crypto_handle_ctrl);
if (!vcrypto->cryptodev->ready) {
vcrypto->status &= ~VIRTIO_CRYPTO_S_HW_READY;
} else {
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 24/47] virtio-crypto: add data queue processing handler
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (22 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 23/47] virtio-crypto: add control queue handler Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 25/47] cryptodev: introduce an unified wrapper for crypto operation Michael S. Tsirkin
` (25 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Introduces VirtIOCryptoReq structure to store
crypto request so that we can easily support
asynchronous crypto operation in the future.
At present, we only support cipher and algorithm
chaining.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-crypto.h | 4 +
hw/virtio/virtio-crypto.c | 358 +++++++++++++++++++++++++++++++++++++-
2 files changed, 361 insertions(+), 1 deletion(-)
diff --git a/include/hw/virtio/virtio-crypto.h b/include/hw/virtio/virtio-crypto.h
index 783ea23..db5c941 100644
--- a/include/hw/virtio/virtio-crypto.h
+++ b/include/hw/virtio/virtio-crypto.h
@@ -65,6 +65,10 @@ typedef struct VirtIOCryptoReq {
VirtQueueElement elem;
/* flags of operation, such as type of algorithm */
uint32_t flags;
+ struct virtio_crypto_inhdr *in;
+ struct iovec *in_iov; /* Head address of dest iovec */
+ unsigned int in_num; /* Number of dest iovec */
+ size_t in_len;
VirtQueue *vq;
struct VirtIOCrypto *vcrypto;
union {
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
index 195a266..6c8770a 100644
--- a/hw/virtio/virtio-crypto.c
+++ b/hw/virtio/virtio-crypto.c
@@ -320,6 +320,362 @@ static void virtio_crypto_handle_ctrl(VirtIODevice *vdev, VirtQueue *vq)
} /* end for loop */
}
+static void virtio_crypto_init_request(VirtIOCrypto *vcrypto, VirtQueue *vq,
+ VirtIOCryptoReq *req)
+{
+ req->vcrypto = vcrypto;
+ req->vq = vq;
+ req->in = NULL;
+ req->in_iov = NULL;
+ req->in_num = 0;
+ req->in_len = 0;
+ req->flags = CRYPTODEV_BACKEND_ALG__MAX;
+ req->u.sym_op_info = NULL;
+}
+
+static void virtio_crypto_free_request(VirtIOCryptoReq *req)
+{
+ if (req) {
+ if (req->flags == CRYPTODEV_BACKEND_ALG_SYM) {
+ g_free(req->u.sym_op_info);
+ }
+ g_free(req);
+ }
+}
+
+static void
+virtio_crypto_sym_input_data_helper(VirtIODevice *vdev,
+ VirtIOCryptoReq *req,
+ uint32_t status,
+ CryptoDevBackendSymOpInfo *sym_op_info)
+{
+ size_t s, len;
+
+ if (status != VIRTIO_CRYPTO_OK) {
+ return;
+ }
+
+ len = sym_op_info->dst_len;
+ /* Save the cipher result */
+ s = iov_from_buf(req->in_iov, req->in_num, 0, sym_op_info->dst, len);
+ if (s != len) {
+ virtio_error(vdev, "virtio-crypto dest data incorrect");
+ return;
+ }
+
+ iov_discard_front(&req->in_iov, &req->in_num, len);
+
+ if (sym_op_info->op_type ==
+ VIRTIO_CRYPTO_SYM_OP_ALGORITHM_CHAINING) {
+ /* Save the digest result */
+ s = iov_from_buf(req->in_iov, req->in_num, 0,
+ sym_op_info->digest_result,
+ sym_op_info->digest_result_len);
+ if (s != sym_op_info->digest_result_len) {
+ virtio_error(vdev, "virtio-crypto digest result incorrect");
+ }
+ }
+}
+
+static void virtio_crypto_req_complete(VirtIOCryptoReq *req, uint8_t status)
+{
+ VirtIOCrypto *vcrypto = req->vcrypto;
+ VirtIODevice *vdev = VIRTIO_DEVICE(vcrypto);
+
+ if (req->flags == CRYPTODEV_BACKEND_ALG_SYM) {
+ virtio_crypto_sym_input_data_helper(vdev, req, status,
+ req->u.sym_op_info);
+ }
+ stb_p(&req->in->status, status);
+ virtqueue_push(req->vq, &req->elem, req->in_len);
+ virtio_notify(vdev, req->vq);
+}
+
+static VirtIOCryptoReq *
+virtio_crypto_get_request(VirtIOCrypto *s, VirtQueue *vq)
+{
+ VirtIOCryptoReq *req = virtqueue_pop(vq, sizeof(VirtIOCryptoReq));
+
+ if (req) {
+ virtio_crypto_init_request(s, vq, req);
+ }
+ return req;
+}
+
+static CryptoDevBackendSymOpInfo *
+virtio_crypto_sym_op_helper(VirtIODevice *vdev,
+ struct virtio_crypto_cipher_para *cipher_para,
+ struct virtio_crypto_alg_chain_data_para *alg_chain_para,
+ struct iovec *iov, unsigned int out_num)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ CryptoDevBackendSymOpInfo *op_info;
+ uint32_t src_len = 0, dst_len = 0;
+ uint32_t iv_len = 0;
+ uint32_t aad_len = 0, hash_result_len = 0;
+ uint32_t hash_start_src_offset = 0, len_to_hash = 0;
+ uint32_t cipher_start_src_offset = 0, len_to_cipher = 0;
+
+ size_t max_len, curr_size = 0;
+ size_t s;
+
+ /* Plain cipher */
+ if (cipher_para) {
+ iv_len = ldl_le_p(&cipher_para->iv_len);
+ src_len = ldl_le_p(&cipher_para->src_data_len);
+ dst_len = ldl_le_p(&cipher_para->dst_data_len);
+ } else if (alg_chain_para) { /* Algorithm chain */
+ iv_len = ldl_le_p(&alg_chain_para->iv_len);
+ src_len = ldl_le_p(&alg_chain_para->src_data_len);
+ dst_len = ldl_le_p(&alg_chain_para->dst_data_len);
+
+ aad_len = ldl_le_p(&alg_chain_para->aad_len);
+ hash_result_len = ldl_le_p(&alg_chain_para->hash_result_len);
+ hash_start_src_offset = ldl_le_p(
+ &alg_chain_para->hash_start_src_offset);
+ cipher_start_src_offset = ldl_le_p(
+ &alg_chain_para->cipher_start_src_offset);
+ len_to_cipher = ldl_le_p(&alg_chain_para->len_to_cipher);
+ len_to_hash = ldl_le_p(&alg_chain_para->len_to_hash);
+ } else {
+ return NULL;
+ }
+
+ max_len = iv_len + aad_len + src_len + dst_len + hash_result_len;
+ if (unlikely(max_len > vcrypto->conf.max_size)) {
+ virtio_error(vdev, "virtio-crypto too big length");
+ return NULL;
+ }
+
+ op_info = g_malloc0(sizeof(CryptoDevBackendSymOpInfo) + max_len);
+ op_info->iv_len = iv_len;
+ op_info->src_len = src_len;
+ op_info->dst_len = dst_len;
+ op_info->aad_len = aad_len;
+ op_info->digest_result_len = hash_result_len;
+ op_info->hash_start_src_offset = hash_start_src_offset;
+ op_info->len_to_hash = len_to_hash;
+ op_info->cipher_start_src_offset = cipher_start_src_offset;
+ op_info->len_to_cipher = len_to_cipher;
+ /* Handle the initilization vector */
+ if (op_info->iv_len > 0) {
+ DPRINTF("iv_len=%" PRIu32 "\n", op_info->iv_len);
+ op_info->iv = op_info->data + curr_size;
+
+ s = iov_to_buf(iov, out_num, 0, op_info->iv, op_info->iv_len);
+ if (unlikely(s != op_info->iv_len)) {
+ virtio_error(vdev, "virtio-crypto iv incorrect");
+ goto err;
+ }
+ iov_discard_front(&iov, &out_num, op_info->iv_len);
+ curr_size += op_info->iv_len;
+ }
+
+ /* Handle additional authentication data if exists */
+ if (op_info->aad_len > 0) {
+ DPRINTF("aad_len=%" PRIu32 "\n", op_info->aad_len);
+ op_info->aad_data = op_info->data + curr_size;
+
+ s = iov_to_buf(iov, out_num, 0, op_info->aad_data, op_info->aad_len);
+ if (unlikely(s != op_info->aad_len)) {
+ virtio_error(vdev, "virtio-crypto additional auth data incorrect");
+ goto err;
+ }
+ iov_discard_front(&iov, &out_num, op_info->aad_len);
+
+ curr_size += op_info->aad_len;
+ }
+
+ /* Handle the source data */
+ if (op_info->src_len > 0) {
+ DPRINTF("src_len=%" PRIu32 "\n", op_info->src_len);
+ op_info->src = op_info->data + curr_size;
+
+ s = iov_to_buf(iov, out_num, 0, op_info->src, op_info->src_len);
+ if (unlikely(s != op_info->src_len)) {
+ virtio_error(vdev, "virtio-crypto source data incorrect");
+ goto err;
+ }
+ iov_discard_front(&iov, &out_num, op_info->src_len);
+
+ curr_size += op_info->src_len;
+ }
+
+ /* Handle the destination data */
+ op_info->dst = op_info->data + curr_size;
+ curr_size += op_info->dst_len;
+
+ DPRINTF("dst_len=%" PRIu32 "\n", op_info->dst_len);
+
+ /* Handle the hash digest result */
+ if (hash_result_len > 0) {
+ DPRINTF("hash_result_len=%" PRIu32 "\n", hash_result_len);
+ op_info->digest_result = op_info->data + curr_size;
+ }
+
+ return op_info;
+
+err:
+ g_free(op_info);
+ return NULL;
+}
+
+static int
+virtio_crypto_handle_sym_req(VirtIOCrypto *vcrypto,
+ struct virtio_crypto_sym_data_req *req,
+ CryptoDevBackendSymOpInfo **sym_op_info,
+ struct iovec *iov, unsigned int out_num)
+{
+ VirtIODevice *vdev = VIRTIO_DEVICE(vcrypto);
+ uint32_t op_type;
+ CryptoDevBackendSymOpInfo *op_info;
+
+ op_type = ldl_le_p(&req->op_type);
+
+ if (op_type == VIRTIO_CRYPTO_SYM_OP_CIPHER) {
+ op_info = virtio_crypto_sym_op_helper(vdev, &req->u.cipher.para,
+ NULL, iov, out_num);
+ if (!op_info) {
+ return -EFAULT;
+ }
+ op_info->op_type = op_type;
+ } else if (op_type == VIRTIO_CRYPTO_SYM_OP_ALGORITHM_CHAINING) {
+ op_info = virtio_crypto_sym_op_helper(vdev, NULL,
+ &req->u.chain.para,
+ iov, out_num);
+ if (!op_info) {
+ return -EFAULT;
+ }
+ op_info->op_type = op_type;
+ } else {
+ /* VIRTIO_CRYPTO_SYM_OP_NONE */
+ error_report("virtio-crypto unsupported cipher type");
+ return -VIRTIO_CRYPTO_NOTSUPP;
+ }
+
+ *sym_op_info = op_info;
+
+ return 0;
+}
+
+static int
+virtio_crypto_handle_request(VirtIOCryptoReq *request)
+{
+ VirtIOCrypto *vcrypto = request->vcrypto;
+ VirtIODevice *vdev = VIRTIO_DEVICE(vcrypto);
+ VirtQueueElement *elem = &request->elem;
+ int queue_index = virtio_crypto_vq2q(virtio_get_queue_index(request->vq));
+ struct virtio_crypto_op_data_req req;
+ int ret;
+ struct iovec *in_iov;
+ struct iovec *out_iov;
+ unsigned in_num;
+ unsigned out_num;
+ uint32_t opcode;
+ uint8_t status = VIRTIO_CRYPTO_ERR;
+ uint64_t session_id;
+ CryptoDevBackendSymOpInfo *sym_op_info = NULL;
+ Error *local_err = NULL;
+
+ if (elem->out_num < 1 || elem->in_num < 1) {
+ virtio_error(vdev, "virtio-crypto dataq missing headers");
+ return -1;
+ }
+
+ out_num = elem->out_num;
+ out_iov = elem->out_sg;
+ in_num = elem->in_num;
+ in_iov = elem->in_sg;
+ if (unlikely(iov_to_buf(out_iov, out_num, 0, &req, sizeof(req))
+ != sizeof(req))) {
+ virtio_error(vdev, "virtio-crypto request outhdr too short");
+ return -1;
+ }
+ iov_discard_front(&out_iov, &out_num, sizeof(req));
+
+ if (in_iov[in_num - 1].iov_len <
+ sizeof(struct virtio_crypto_inhdr)) {
+ virtio_error(vdev, "virtio-crypto request inhdr too short");
+ return -1;
+ }
+ /* We always touch the last byte, so just see how big in_iov is. */
+ request->in_len = iov_size(in_iov, in_num);
+ request->in = (void *)in_iov[in_num - 1].iov_base
+ + in_iov[in_num - 1].iov_len
+ - sizeof(struct virtio_crypto_inhdr);
+ iov_discard_back(in_iov, &in_num, sizeof(struct virtio_crypto_inhdr));
+
+ /*
+ * The length of operation result, including dest_data
+ * and digest_result if exists.
+ */
+ request->in_num = in_num;
+ request->in_iov = in_iov;
+
+ opcode = ldl_le_p(&req.header.opcode);
+ session_id = ldq_le_p(&req.header.session_id);
+
+ switch (opcode) {
+ case VIRTIO_CRYPTO_CIPHER_ENCRYPT:
+ case VIRTIO_CRYPTO_CIPHER_DECRYPT:
+ ret = virtio_crypto_handle_sym_req(vcrypto,
+ &req.u.sym_req,
+ &sym_op_info,
+ out_iov, out_num);
+ /* Serious errors, need to reset virtio crypto device */
+ if (ret == -EFAULT) {
+ return -1;
+ } else if (ret == -VIRTIO_CRYPTO_NOTSUPP) {
+ virtio_crypto_req_complete(request, VIRTIO_CRYPTO_NOTSUPP);
+ virtio_crypto_free_request(request);
+ } else {
+ sym_op_info->session_id = session_id;
+
+ /* Set request's parameter */
+ request->flags = CRYPTODEV_BACKEND_ALG_SYM;
+ request->u.sym_op_info = sym_op_info;
+ ret = cryptodev_backend_sym_operation(vcrypto->cryptodev,
+ sym_op_info, queue_index, &local_err);
+ if (ret < 0) {
+ status = VIRTIO_CRYPTO_ERR;
+ if (local_err) {
+ error_report_err(local_err);
+ }
+ } else { /* ret >= 0 */
+ status = VIRTIO_CRYPTO_OK;
+ }
+ virtio_crypto_req_complete(request, status);
+ virtio_crypto_free_request(request);
+ }
+ break;
+ case VIRTIO_CRYPTO_HASH:
+ case VIRTIO_CRYPTO_MAC:
+ case VIRTIO_CRYPTO_AEAD_ENCRYPT:
+ case VIRTIO_CRYPTO_AEAD_DECRYPT:
+ default:
+ error_report("virtio-crypto unsupported dataq opcode: %u",
+ opcode);
+ virtio_crypto_req_complete(request, VIRTIO_CRYPTO_NOTSUPP);
+ virtio_crypto_free_request(request);
+ }
+
+ return 0;
+}
+
+static void virtio_crypto_handle_dataq(VirtIODevice *vdev, VirtQueue *vq)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ VirtIOCryptoReq *req;
+
+ while ((req = virtio_crypto_get_request(vcrypto, vq))) {
+ if (virtio_crypto_handle_request(req) < 0) {
+ virtqueue_detach_element(req->vq, &req->elem, 0);
+ virtio_crypto_free_request(req);
+ break;
+ }
+ }
+}
+
static uint64_t virtio_crypto_get_features(VirtIODevice *vdev,
uint64_t features,
Error **errp)
@@ -384,7 +740,7 @@ static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
vcrypto->curr_queues = 1;
for (i = 0; i < vcrypto->max_queues; i++) {
- virtio_add_queue(vdev, 1024, NULL);
+ virtio_add_queue(vdev, 1024, virtio_crypto_handle_dataq);
}
vcrypto->ctrl_vq = virtio_add_queue(vdev, 64, virtio_crypto_handle_ctrl);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 25/47] cryptodev: introduce an unified wrapper for crypto operation
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (23 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 24/47] virtio-crypto: add data queue processing handler Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 26/47] virtio-crypto: using bh to handle dataq's requests Michael S. Tsirkin
` (24 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
We use an opaque point to the VirtIOCryptoReq which
can support different packets based on different
algorithms.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/sysemu/cryptodev.h | 13 +++++++------
backends/cryptodev.c | 28 ++++++++++++++++++++++++++--
hw/virtio/virtio-crypto.c | 10 +++++-----
3 files changed, 38 insertions(+), 13 deletions(-)
diff --git a/include/sysemu/cryptodev.h b/include/sysemu/cryptodev.h
index e66bd4b..84526c0 100644
--- a/include/sysemu/cryptodev.h
+++ b/include/sysemu/cryptodev.h
@@ -278,20 +278,21 @@ int cryptodev_backend_sym_close_session(
uint32_t queue_index, Error **errp);
/**
- * cryptodev_backend_sym_operation:
+ * cryptodev_backend_crypto_operation:
* @backend: the cryptodev backend object
- * @op_info: parameters needed by symmetric crypto operation
+ * @opaque: pointer to a VirtIOCryptoReq object
* @queue_index: queue index of cryptodev backend client
* @errp: pointer to a NULL-initialized error object
*
- * Do symmetric crypto operation, such as encryption and
+ * Do crypto operation, such as encryption and
* decryption
*
- * Returns: 0 on success, or Negative on error
+ * Returns: VIRTIO_CRYPTO_OK on success,
+ * or -VIRTIO_CRYPTO_* on error
*/
-int cryptodev_backend_sym_operation(
+int cryptodev_backend_crypto_operation(
CryptoDevBackend *backend,
- CryptoDevBackendSymOpInfo *op_info,
+ void *opaque,
uint32_t queue_index, Error **errp);
#endif /* CRYPTODEV_H */
diff --git a/backends/cryptodev.c b/backends/cryptodev.c
index 47521cf..4a49f97 100644
--- a/backends/cryptodev.c
+++ b/backends/cryptodev.c
@@ -30,6 +30,8 @@
#include "qapi-visit.h"
#include "qemu/config-file.h"
#include "qom/object_interfaces.h"
+#include "hw/virtio/virtio-crypto.h"
+
static QTAILQ_HEAD(, CryptoDevBackendClient) crypto_clients;
@@ -105,7 +107,7 @@ int cryptodev_backend_sym_close_session(
return -1;
}
-int cryptodev_backend_sym_operation(
+static int cryptodev_backend_sym_operation(
CryptoDevBackend *backend,
CryptoDevBackendSymOpInfo *op_info,
uint32_t queue_index, Error **errp)
@@ -117,7 +119,29 @@ int cryptodev_backend_sym_operation(
return bc->do_sym_op(backend, op_info, queue_index, errp);
}
- return -1;
+ return -VIRTIO_CRYPTO_ERR;
+}
+
+int cryptodev_backend_crypto_operation(
+ CryptoDevBackend *backend,
+ void *opaque,
+ uint32_t queue_index, Error **errp)
+{
+ VirtIOCryptoReq *req = opaque;
+
+ if (req->flags == CRYPTODEV_BACKEND_ALG_SYM) {
+ CryptoDevBackendSymOpInfo *op_info;
+ op_info = req->u.sym_op_info;
+
+ return cryptodev_backend_sym_operation(backend,
+ op_info, queue_index, errp);
+ } else {
+ error_setg(errp, "Unsupported cryptodev alg type: %" PRIu32 "",
+ req->flags);
+ return -VIRTIO_CRYPTO_NOTSUPP;
+ }
+
+ return -VIRTIO_CRYPTO_ERR;
}
static void
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
index 6c8770a..6faacb7 100644
--- a/hw/virtio/virtio-crypto.c
+++ b/hw/virtio/virtio-crypto.c
@@ -634,15 +634,15 @@ virtio_crypto_handle_request(VirtIOCryptoReq *request)
/* Set request's parameter */
request->flags = CRYPTODEV_BACKEND_ALG_SYM;
request->u.sym_op_info = sym_op_info;
- ret = cryptodev_backend_sym_operation(vcrypto->cryptodev,
- sym_op_info, queue_index, &local_err);
+ ret = cryptodev_backend_crypto_operation(vcrypto->cryptodev,
+ request, queue_index, &local_err);
if (ret < 0) {
- status = VIRTIO_CRYPTO_ERR;
+ status = -ret;
if (local_err) {
error_report_err(local_err);
}
- } else { /* ret >= 0 */
- status = VIRTIO_CRYPTO_OK;
+ } else { /* ret == VIRTIO_CRYPTO_OK */
+ status = ret;
}
virtio_crypto_req_complete(request, status);
virtio_crypto_free_request(request);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 26/47] virtio-crypto: using bh to handle dataq's requests
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (24 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 25/47] cryptodev: introduce an unified wrapper for crypto operation Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 27/47] virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer Michael S. Tsirkin
` (23 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Gonglei
From: Gonglei <arei.gonglei@huawei.com>
Make crypto operations are executed asynchronously,
so that other QEMU threads and monitor couldn't
be blocked at the virtqueue handling context.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/virtio-crypto.h | 8 +++++-
hw/virtio/virtio-crypto.c | 55 +++++++++++++++++++++++++++++++++++++--
2 files changed, 60 insertions(+), 3 deletions(-)
diff --git a/include/hw/virtio/virtio-crypto.h b/include/hw/virtio/virtio-crypto.h
index db5c941..a00a0bf 100644
--- a/include/hw/virtio/virtio-crypto.h
+++ b/include/hw/virtio/virtio-crypto.h
@@ -76,11 +76,17 @@ typedef struct VirtIOCryptoReq {
} u;
} VirtIOCryptoReq;
+typedef struct VirtIOCryptoQueue {
+ VirtQueue *dataq;
+ QEMUBH *dataq_bh;
+ struct VirtIOCrypto *vcrypto;
+} VirtIOCryptoQueue;
+
typedef struct VirtIOCrypto {
VirtIODevice parent_obj;
VirtQueue *ctrl_vq;
-
+ VirtIOCryptoQueue *vqs;
VirtIOCryptoConf conf;
CryptoDevBackend *cryptodev;
diff --git a/hw/virtio/virtio-crypto.c b/hw/virtio/virtio-crypto.c
index 6faacb7..dcb1e06 100644
--- a/hw/virtio/virtio-crypto.c
+++ b/hw/virtio/virtio-crypto.c
@@ -676,6 +676,41 @@ static void virtio_crypto_handle_dataq(VirtIODevice *vdev, VirtQueue *vq)
}
}
+static void virtio_crypto_dataq_bh(void *opaque)
+{
+ VirtIOCryptoQueue *q = opaque;
+ VirtIOCrypto *vcrypto = q->vcrypto;
+ VirtIODevice *vdev = VIRTIO_DEVICE(vcrypto);
+
+ /* This happens when device was stopped but BH wasn't. */
+ if (!vdev->vm_running) {
+ return;
+ }
+
+ /* Just in case the driver is not ready on more */
+ if (unlikely(!(vdev->status & VIRTIO_CONFIG_S_DRIVER_OK))) {
+ return;
+ }
+
+ virtio_crypto_handle_dataq(vdev, q->dataq);
+ virtio_queue_set_notification(q->dataq, 1);
+}
+
+static void
+virtio_crypto_handle_dataq_bh(VirtIODevice *vdev, VirtQueue *vq)
+{
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(vdev);
+ VirtIOCryptoQueue *q =
+ &vcrypto->vqs[virtio_crypto_vq2q(virtio_get_queue_index(vq))];
+
+ /* This happens when device was stopped but VCPU wasn't. */
+ if (!vdev->vm_running) {
+ return;
+ }
+ virtio_queue_set_notification(vq, 0);
+ qemu_bh_schedule(q->dataq_bh);
+}
+
static uint64_t virtio_crypto_get_features(VirtIODevice *vdev,
uint64_t features,
Error **errp)
@@ -738,9 +773,13 @@ static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
virtio_init(vdev, "virtio-crypto", VIRTIO_ID_CRYPTO, vcrypto->config_size);
vcrypto->curr_queues = 1;
-
+ vcrypto->vqs = g_malloc0(sizeof(VirtIOCryptoQueue) * vcrypto->max_queues);
for (i = 0; i < vcrypto->max_queues; i++) {
- virtio_add_queue(vdev, 1024, virtio_crypto_handle_dataq);
+ vcrypto->vqs[i].dataq =
+ virtio_add_queue(vdev, 1024, virtio_crypto_handle_dataq_bh);
+ vcrypto->vqs[i].dataq_bh =
+ qemu_bh_new(virtio_crypto_dataq_bh, &vcrypto->vqs[i]);
+ vcrypto->vqs[i].vcrypto = vcrypto;
}
vcrypto->ctrl_vq = virtio_add_queue(vdev, 64, virtio_crypto_handle_ctrl);
@@ -756,6 +795,18 @@ static void virtio_crypto_device_realize(DeviceState *dev, Error **errp)
static void virtio_crypto_device_unrealize(DeviceState *dev, Error **errp)
{
VirtIODevice *vdev = VIRTIO_DEVICE(dev);
+ VirtIOCrypto *vcrypto = VIRTIO_CRYPTO(dev);
+ VirtIOCryptoQueue *q;
+ int i, max_queues;
+
+ max_queues = vcrypto->multiqueue ? vcrypto->max_queues : 1;
+ for (i = 0; i < max_queues; i++) {
+ virtio_del_queue(vdev, i);
+ q = &vcrypto->vqs[i];
+ qemu_bh_delete(q->dataq_bh);
+ }
+
+ g_free(vcrypto->vqs);
virtio_cleanup(vdev);
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 27/47] virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (25 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 26/47] virtio-crypto: using bh to handle dataq's requests Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 28/47] acpi nvdimm: fix wrong buffer size returned by DSM method Michael S. Tsirkin
` (22 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Gonglei, Thomas Huth, Michael Tokarev, Markus Armbruster
From: Gonglei <arei.gonglei@huawei.com>
This patch includes two parts: Cryptodev Backends
and virtio-crypto stuff. I can maintain cryptodev backends
which introduced by myself. For virtio-crypto stuff, I can
share the work with Michael (The whole virtio supporter).
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
MAINTAINERS | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 82d4d00..a396261 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1005,6 +1005,13 @@ F: include/sysemu/rng*.h
F: backends/rng*.c
F: tests/virtio-rng-test.c
+virtio-crypto
+M: Gonglei <arei.gonglei@huawei.com>
+S: Supported
+F: hw/virtio/virtio-crypto.c
+F: hw/virtio/virtio-crypto-pci.c
+F: include/hw/virtio/virtio-crypto.h
+
nvme
M: Keith Busch <keith.busch@intel.com>
L: qemu-block@nongnu.org
@@ -1250,6 +1257,12 @@ S: Maintained
F: backends/hostmem*.c
F: include/sysemu/hostmem.h
+Cryptodev Backends
+M: Gonglei <arei.gonglei@huawei.com>
+S: Maintained
+F: include/sysemu/cryptodev*.h
+F: backends/cryptodev*.c
+
QAPI
M: Markus Armbruster <armbru@redhat.com>
M: Michael Roth <mdroth@linux.vnet.ibm.com>
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 28/47] acpi nvdimm: fix wrong buffer size returned by DSM method
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (26 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 27/47] virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 29/47] acpi nvdimm: fix OperationRegion definition Michael S. Tsirkin
` (21 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Currently, 'RLEN' is the totally buffer size written by QEMU and it is
ACPI internally used only. The buffer size returned to guest should
not include 'RLEN' itself
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index e486128..24a2b3b 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -862,7 +862,8 @@ static void nvdimm_build_common_dsm(Aml *dev)
aml_append(method, aml_store(dsm_mem, aml_name("NTFI")));
result_size = aml_local(1);
- aml_append(method, aml_store(aml_name("RLEN"), result_size));
+ /* RLEN is not included in the payload returned to guest. */
+ aml_append(method, aml_subtract(aml_name("RLEN"), aml_int(4), result_size));
aml_append(method, aml_store(aml_shiftleft(result_size, aml_int(3)),
result_size));
aml_append(method, aml_create_field(aml_name("ODAT"), aml_int(0),
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 29/47] acpi nvdimm: fix OperationRegion definition
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (27 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 28/47] acpi nvdimm: fix wrong buffer size returned by DSM method Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base Michael S. Tsirkin
` (20 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Based on ACPI spec:
RegionOffset := TermArg => Integer
However, Named object is not a TermArg.
This patch moves OperationRegion to NCAL() and uses localX as
its RegionOffset
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 122 ++++++++++++++++++++++++++++---------------------------
1 file changed, 62 insertions(+), 60 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index 24a2b3b..bbb2cfd 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -781,14 +781,73 @@ static void nvdimm_build_common_dsm(Aml *dev)
{
Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem, *result_size;
Aml *elsectx, *unsupport, *unpatched, *expected_uuid, *uuid_invalid;
- Aml *pckg, *pckg_index, *pckg_buf;
+ Aml *pckg, *pckg_index, *pckg_buf, *field;
uint8_t byte_list[1];
method = aml_method(NVDIMM_COMMON_DSM, 5, AML_SERIALIZED);
uuid = aml_arg(0);
function = aml_arg(2);
handle = aml_arg(4);
- dsm_mem = aml_name(NVDIMM_ACPI_MEM_ADDR);
+ dsm_mem = aml_local(6);
+
+ aml_append(method, aml_store(aml_name(NVDIMM_ACPI_MEM_ADDR), dsm_mem));
+
+ /* map DSM memory and IO into ACPI namespace. */
+ aml_append(method, aml_operation_region("NPIO", AML_SYSTEM_IO,
+ aml_int(NVDIMM_ACPI_IO_BASE), NVDIMM_ACPI_IO_LEN));
+ aml_append(method, aml_operation_region("NRAM", AML_SYSTEM_MEMORY,
+ dsm_mem, sizeof(NvdimmDsmIn)));
+
+ /*
+ * DSM notifier:
+ * NTFI: write the address of DSM memory and notify QEMU to emulate
+ * the access.
+ *
+ * It is the IO port so that accessing them will cause VM-exit, the
+ * control will be transferred to QEMU.
+ */
+ field = aml_field("NPIO", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
+ aml_append(field, aml_named_field("NTFI",
+ sizeof(uint32_t) * BITS_PER_BYTE));
+ aml_append(method, field);
+
+ /*
+ * DSM input:
+ * HDLE: store device's handle, it's zero if the _DSM call happens
+ * on NVDIMM Root Device.
+ * REVS: store the Arg1 of _DSM call.
+ * FUNC: store the Arg2 of _DSM call.
+ * ARG3: store the Arg3 of _DSM call.
+ *
+ * They are RAM mapping on host so that these accesses never cause
+ * VM-EXIT.
+ */
+ field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
+ aml_append(field, aml_named_field("HDLE",
+ sizeof(typeof_field(NvdimmDsmIn, handle)) * BITS_PER_BYTE));
+ aml_append(field, aml_named_field("REVS",
+ sizeof(typeof_field(NvdimmDsmIn, revision)) * BITS_PER_BYTE));
+ aml_append(field, aml_named_field("FUNC",
+ sizeof(typeof_field(NvdimmDsmIn, function)) * BITS_PER_BYTE));
+ aml_append(field, aml_named_field("ARG3",
+ (sizeof(NvdimmDsmIn) - offsetof(NvdimmDsmIn, arg3)) * BITS_PER_BYTE));
+ aml_append(method, field);
+
+ /*
+ * DSM output:
+ * RLEN: the size of the buffer filled by QEMU.
+ * ODAT: the buffer QEMU uses to store the result.
+ *
+ * Since the page is reused by both input and out, the input data
+ * will be lost after storing new result into ODAT so we should fetch
+ * all the input data before writing the result.
+ */
+ field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
+ aml_append(field, aml_named_field("RLEN",
+ sizeof(typeof_field(NvdimmDsmOut, len)) * BITS_PER_BYTE));
+ aml_append(field, aml_named_field("ODAT",
+ (sizeof(NvdimmDsmOut) - offsetof(NvdimmDsmOut, data)) * BITS_PER_BYTE));
+ aml_append(method, field);
/*
* do not support any method if DSM memory address has not been
@@ -915,7 +974,7 @@ static void nvdimm_build_ssdt(GSList *device_list, GArray *table_offsets,
GArray *table_data, BIOSLinker *linker,
GArray *dsm_dma_arrea)
{
- Aml *ssdt, *sb_scope, *dev, *field;
+ Aml *ssdt, *sb_scope, *dev;
int mem_addr_offset, nvdimm_ssdt;
acpi_add_table(table_offsets, table_data);
@@ -940,63 +999,6 @@ static void nvdimm_build_ssdt(GSList *device_list, GArray *table_offsets,
*/
aml_append(dev, aml_name_decl("_HID", aml_string("ACPI0012")));
- /* map DSM memory and IO into ACPI namespace. */
- aml_append(dev, aml_operation_region("NPIO", AML_SYSTEM_IO,
- aml_int(NVDIMM_ACPI_IO_BASE), NVDIMM_ACPI_IO_LEN));
- aml_append(dev, aml_operation_region("NRAM", AML_SYSTEM_MEMORY,
- aml_name(NVDIMM_ACPI_MEM_ADDR), sizeof(NvdimmDsmIn)));
-
- /*
- * DSM notifier:
- * NTFI: write the address of DSM memory and notify QEMU to emulate
- * the access.
- *
- * It is the IO port so that accessing them will cause VM-exit, the
- * control will be transferred to QEMU.
- */
- field = aml_field("NPIO", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("NTFI",
- sizeof(uint32_t) * BITS_PER_BYTE));
- aml_append(dev, field);
-
- /*
- * DSM input:
- * HDLE: store device's handle, it's zero if the _DSM call happens
- * on NVDIMM Root Device.
- * REVS: store the Arg1 of _DSM call.
- * FUNC: store the Arg2 of _DSM call.
- * ARG3: store the Arg3 of _DSM call.
- *
- * They are RAM mapping on host so that these accesses never cause
- * VM-EXIT.
- */
- field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("HDLE",
- sizeof(typeof_field(NvdimmDsmIn, handle)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("REVS",
- sizeof(typeof_field(NvdimmDsmIn, revision)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("FUNC",
- sizeof(typeof_field(NvdimmDsmIn, function)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("ARG3",
- (sizeof(NvdimmDsmIn) - offsetof(NvdimmDsmIn, arg3)) * BITS_PER_BYTE));
- aml_append(dev, field);
-
- /*
- * DSM output:
- * RLEN: the size of the buffer filled by QEMU.
- * ODAT: the buffer QEMU uses to store the result.
- *
- * Since the page is reused by both input and out, the input data
- * will be lost after storing new result into ODAT so we should fetch
- * all the input data before writing the result.
- */
- field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("RLEN",
- sizeof(typeof_field(NvdimmDsmOut, len)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("ODAT",
- (sizeof(NvdimmDsmOut) - offsetof(NvdimmDsmOut, data)) * BITS_PER_BYTE));
- aml_append(dev, field);
-
nvdimm_build_common_dsm(dev);
/* 0 is reserved for root device. */
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (28 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 29/47] acpi nvdimm: fix OperationRegion definition Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-31 9:20 ` Igor Mammedov
2016-10-30 21:24 ` [Qemu-devel] [PULL 31/47] acpi nvdimm: fix ARG3 conflict Michael S. Tsirkin
` (19 subsequent siblings)
49 siblings, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
According to ACPI 6.0 spec, "Memory Device Physical Address
Region Base" in memdev is defined as "This field provides the
Device Physical Address base of the region". This field should
be zero in our case
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index bbb2cfd..c2f5caa 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -289,8 +289,6 @@ static void
nvdimm_build_structure_memdev(GArray *structures, DeviceState *dev)
{
NvdimmNfitMemDev *nfit_memdev;
- uint64_t addr = object_property_get_int(OBJECT(dev), PC_DIMM_ADDR_PROP,
- NULL);
uint64_t size = object_property_get_int(OBJECT(dev), PC_DIMM_SIZE_PROP,
NULL);
int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
@@ -314,7 +312,8 @@ nvdimm_build_structure_memdev(GArray *structures, DeviceState *dev)
/* The memory region on the device. */
nfit_memdev->region_len = cpu_to_le64(size);
- nfit_memdev->region_dpa = cpu_to_le64(addr);
+ /* The device address starts from 0. */
+ nfit_memdev->region_dpa = cpu_to_le64(0);
/* Only one interleave for PMEM. */
nfit_memdev->interleave_ways = cpu_to_le16(1);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 31/47] acpi nvdimm: fix ARG3 conflict
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (29 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 32/47] acpi nvdimm: fix Arg6 usage Michael S. Tsirkin
` (18 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
As ARG3 is a reserved name, we rename it to FARG
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index c2f5caa..0b8c275 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -816,7 +816,8 @@ static void nvdimm_build_common_dsm(Aml *dev)
* on NVDIMM Root Device.
* REVS: store the Arg1 of _DSM call.
* FUNC: store the Arg2 of _DSM call.
- * ARG3: store the Arg3 of _DSM call.
+ * FARG: store the Arg3 of _DSM call which is a Package containing
+ * function-specific arguments.
*
* They are RAM mapping on host so that these accesses never cause
* VM-EXIT.
@@ -828,7 +829,7 @@ static void nvdimm_build_common_dsm(Aml *dev)
sizeof(typeof_field(NvdimmDsmIn, revision)) * BITS_PER_BYTE));
aml_append(field, aml_named_field("FUNC",
sizeof(typeof_field(NvdimmDsmIn, function)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("ARG3",
+ aml_append(field, aml_named_field("FARG",
(sizeof(NvdimmDsmIn) - offsetof(NvdimmDsmIn, arg3)) * BITS_PER_BYTE));
aml_append(method, field);
@@ -910,7 +911,7 @@ static void nvdimm_build_common_dsm(Aml *dev)
pckg_buf = aml_local(3);
aml_append(ifctx, aml_store(aml_index(pckg, aml_int(0)), pckg_index));
aml_append(ifctx, aml_store(aml_derefof(pckg_index), pckg_buf));
- aml_append(ifctx, aml_store(pckg_buf, aml_name("ARG3")));
+ aml_append(ifctx, aml_store(pckg_buf, aml_name("FARG")));
aml_append(method, ifctx);
/*
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 32/47] acpi nvdimm: fix Arg6 usage
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (30 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 31/47] acpi nvdimm: fix ARG3 conflict Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 33/47] nvdimm acpi: compile nvdimm acpi code arch-independently Michael S. Tsirkin
` (17 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
As the function only has 5 args, we use local7 instead of it
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index 0b8c275..cc958a4 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -780,7 +780,7 @@ static void nvdimm_build_common_dsm(Aml *dev)
{
Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem, *result_size;
Aml *elsectx, *unsupport, *unpatched, *expected_uuid, *uuid_invalid;
- Aml *pckg, *pckg_index, *pckg_buf, *field;
+ Aml *pckg, *pckg_index, *pckg_buf, *field, *dsm_out_buf;
uint8_t byte_list[1];
method = aml_method(NVDIMM_COMMON_DSM, 5, AML_SERIALIZED);
@@ -788,6 +788,7 @@ static void nvdimm_build_common_dsm(Aml *dev)
function = aml_arg(2);
handle = aml_arg(4);
dsm_mem = aml_local(6);
+ dsm_out_buf = aml_local(7);
aml_append(method, aml_store(aml_name(NVDIMM_ACPI_MEM_ADDR), dsm_mem));
@@ -928,8 +929,8 @@ static void nvdimm_build_common_dsm(Aml *dev)
aml_append(method, aml_create_field(aml_name("ODAT"), aml_int(0),
result_size, "OBUF"));
aml_append(method, aml_concatenate(aml_buffer(0, NULL), aml_name("OBUF"),
- aml_arg(6)));
- aml_append(method, aml_return(aml_arg(6)));
+ dsm_out_buf));
+ aml_append(method, aml_return(dsm_out_buf));
aml_append(dev, method);
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 33/47] nvdimm acpi: compile nvdimm acpi code arch-independently
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (31 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 32/47] acpi nvdimm: fix Arg6 usage Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 34/47] acpi nvdimm: rename result_size to dsm_out_buf_siz Michael S. Tsirkin
` (16 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
As the arch dependent info, TARGET_PAGE_SIZE, has been dropped
from nvdimm acpi code, it can be compiled arch-independently
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/Makefile.objs | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/acpi/Makefile.objs b/hw/acpi/Makefile.objs
index 4b7da66..489e63b 100644
--- a/hw/acpi/Makefile.objs
+++ b/hw/acpi/Makefile.objs
@@ -3,7 +3,7 @@ common-obj-$(CONFIG_ACPI_X86_ICH) += ich9.o tco.o
common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu_hotplug.o
common-obj-$(CONFIG_ACPI_MEMORY_HOTPLUG) += memory_hotplug.o memory_hotplug_acpi_table.o
common-obj-$(CONFIG_ACPI_CPU_HOTPLUG) += cpu.o
-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o
+common-obj-$(CONFIG_ACPI_NVDIMM) += nvdimm.o
common-obj-$(CONFIG_ACPI) += acpi_interface.o
common-obj-$(CONFIG_ACPI) += bios-linker-loader.o
common-obj-$(CONFIG_ACPI) += aml-build.o
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 34/47] acpi nvdimm: rename result_size to dsm_out_buf_siz
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (32 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 33/47] nvdimm acpi: compile nvdimm acpi code arch-independently Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 35/47] nvdimm acpi: use common macros instead of magic names Michael S. Tsirkin
` (15 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Rename it as dsm_out_buf_siz is more descriptive
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index cc958a4..12126d1 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -778,9 +778,9 @@ void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
static void nvdimm_build_common_dsm(Aml *dev)
{
- Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem, *result_size;
+ Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem;
Aml *elsectx, *unsupport, *unpatched, *expected_uuid, *uuid_invalid;
- Aml *pckg, *pckg_index, *pckg_buf, *field, *dsm_out_buf;
+ Aml *pckg, *pckg_index, *pckg_buf, *field, *dsm_out_buf, *dsm_out_buf_size;
uint8_t byte_list[1];
method = aml_method(NVDIMM_COMMON_DSM, 5, AML_SERIALIZED);
@@ -921,13 +921,14 @@ static void nvdimm_build_common_dsm(Aml *dev)
*/
aml_append(method, aml_store(dsm_mem, aml_name("NTFI")));
- result_size = aml_local(1);
+ dsm_out_buf_size = aml_local(1);
/* RLEN is not included in the payload returned to guest. */
- aml_append(method, aml_subtract(aml_name("RLEN"), aml_int(4), result_size));
- aml_append(method, aml_store(aml_shiftleft(result_size, aml_int(3)),
- result_size));
+ aml_append(method, aml_subtract(aml_name("RLEN"), aml_int(4),
+ dsm_out_buf_size));
+ aml_append(method, aml_store(aml_shiftleft(dsm_out_buf_size, aml_int(3)),
+ dsm_out_buf_size));
aml_append(method, aml_create_field(aml_name("ODAT"), aml_int(0),
- result_size, "OBUF"));
+ dsm_out_buf_size, "OBUF"));
aml_append(method, aml_concatenate(aml_buffer(0, NULL), aml_name("OBUF"),
dsm_out_buf));
aml_append(method, aml_return(dsm_out_buf));
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 35/47] nvdimm acpi: use common macros instead of magic names
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (33 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 34/47] acpi nvdimm: rename result_size to dsm_out_buf_siz Michael S. Tsirkin
@ 2016-10-30 21:24 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 36/47] nvdimm acpi: prebuild nvdimm devices for available slots Michael S. Tsirkin
` (14 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:24 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
There are some names repeatedly used in acpi code, define them
as macros to refine the code
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 83 +++++++++++++++++++++++++++++++++-----------------------
1 file changed, 49 insertions(+), 34 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index 12126d1..bb896c9 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -773,8 +773,20 @@ void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
state->dsm_mem->len);
}
-#define NVDIMM_COMMON_DSM "NCAL"
-#define NVDIMM_ACPI_MEM_ADDR "MEMA"
+#define NVDIMM_COMMON_DSM "NCAL"
+#define NVDIMM_ACPI_MEM_ADDR "MEMA"
+
+#define NVDIMM_DSM_MEMORY "NRAM"
+#define NVDIMM_DSM_IOPORT "NPIO"
+
+#define NVDIMM_DSM_NOTIFY "NTFI"
+#define NVDIMM_DSM_HANDLE "HDLE"
+#define NVDIMM_DSM_REVISION "REVS"
+#define NVDIMM_DSM_FUNCTION "FUNC"
+#define NVDIMM_DSM_ARG3 "FARG"
+
+#define NVDIMM_DSM_OUT_BUF_SIZE "RLEN"
+#define NVDIMM_DSM_OUT_BUF "ODAT"
static void nvdimm_build_common_dsm(Aml *dev)
{
@@ -793,60 +805,63 @@ static void nvdimm_build_common_dsm(Aml *dev)
aml_append(method, aml_store(aml_name(NVDIMM_ACPI_MEM_ADDR), dsm_mem));
/* map DSM memory and IO into ACPI namespace. */
- aml_append(method, aml_operation_region("NPIO", AML_SYSTEM_IO,
+ aml_append(method, aml_operation_region(NVDIMM_DSM_IOPORT, AML_SYSTEM_IO,
aml_int(NVDIMM_ACPI_IO_BASE), NVDIMM_ACPI_IO_LEN));
- aml_append(method, aml_operation_region("NRAM", AML_SYSTEM_MEMORY,
- dsm_mem, sizeof(NvdimmDsmIn)));
+ aml_append(method, aml_operation_region(NVDIMM_DSM_MEMORY,
+ AML_SYSTEM_MEMORY, dsm_mem, sizeof(NvdimmDsmIn)));
/*
* DSM notifier:
- * NTFI: write the address of DSM memory and notify QEMU to emulate
- * the access.
+ * NVDIMM_DSM_NOTIFY: write the address of DSM memory and notify QEMU to
+ * emulate the access.
*
* It is the IO port so that accessing them will cause VM-exit, the
* control will be transferred to QEMU.
*/
- field = aml_field("NPIO", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("NTFI",
+ field = aml_field(NVDIMM_DSM_IOPORT, AML_DWORD_ACC, AML_NOLOCK,
+ AML_PRESERVE);
+ aml_append(field, aml_named_field(NVDIMM_DSM_NOTIFY,
sizeof(uint32_t) * BITS_PER_BYTE));
aml_append(method, field);
/*
* DSM input:
- * HDLE: store device's handle, it's zero if the _DSM call happens
- * on NVDIMM Root Device.
- * REVS: store the Arg1 of _DSM call.
- * FUNC: store the Arg2 of _DSM call.
- * FARG: store the Arg3 of _DSM call which is a Package containing
- * function-specific arguments.
+ * NVDIMM_DSM_HANDLE: store device's handle, it's zero if the _DSM call
+ * happens on NVDIMM Root Device.
+ * NVDIMM_DSM_REVISION: store the Arg1 of _DSM call.
+ * NVDIMM_DSM_FUNCTION: store the Arg2 of _DSM call.
+ * NVDIMM_DSM_ARG3: store the Arg3 of _DSM call which is a Package
+ * containing function-specific arguments.
*
* They are RAM mapping on host so that these accesses never cause
* VM-EXIT.
*/
- field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("HDLE",
+ field = aml_field(NVDIMM_DSM_MEMORY, AML_DWORD_ACC, AML_NOLOCK,
+ AML_PRESERVE);
+ aml_append(field, aml_named_field(NVDIMM_DSM_HANDLE,
sizeof(typeof_field(NvdimmDsmIn, handle)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("REVS",
+ aml_append(field, aml_named_field(NVDIMM_DSM_REVISION,
sizeof(typeof_field(NvdimmDsmIn, revision)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("FUNC",
+ aml_append(field, aml_named_field(NVDIMM_DSM_FUNCTION,
sizeof(typeof_field(NvdimmDsmIn, function)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("FARG",
+ aml_append(field, aml_named_field(NVDIMM_DSM_ARG3,
(sizeof(NvdimmDsmIn) - offsetof(NvdimmDsmIn, arg3)) * BITS_PER_BYTE));
aml_append(method, field);
/*
* DSM output:
- * RLEN: the size of the buffer filled by QEMU.
- * ODAT: the buffer QEMU uses to store the result.
+ * NVDIMM_DSM_OUT_BUF_SIZE: the size of the buffer filled by QEMU.
+ * NVDIMM_DSM_OUT_BUF: the buffer QEMU uses to store the result.
*
* Since the page is reused by both input and out, the input data
* will be lost after storing new result into ODAT so we should fetch
* all the input data before writing the result.
*/
- field = aml_field("NRAM", AML_DWORD_ACC, AML_NOLOCK, AML_PRESERVE);
- aml_append(field, aml_named_field("RLEN",
+ field = aml_field(NVDIMM_DSM_MEMORY, AML_DWORD_ACC, AML_NOLOCK,
+ AML_PRESERVE);
+ aml_append(field, aml_named_field(NVDIMM_DSM_OUT_BUF_SIZE,
sizeof(typeof_field(NvdimmDsmOut, len)) * BITS_PER_BYTE));
- aml_append(field, aml_named_field("ODAT",
+ aml_append(field, aml_named_field(NVDIMM_DSM_OUT_BUF,
(sizeof(NvdimmDsmOut) - offsetof(NvdimmDsmOut, data)) * BITS_PER_BYTE));
aml_append(method, field);
@@ -892,9 +907,9 @@ static void nvdimm_build_common_dsm(Aml *dev)
* it reserves 0 for root device and is the handle for NVDIMM devices.
* See the comments in nvdimm_slot_to_handle().
*/
- aml_append(method, aml_store(handle, aml_name("HDLE")));
- aml_append(method, aml_store(aml_arg(1), aml_name("REVS")));
- aml_append(method, aml_store(aml_arg(2), aml_name("FUNC")));
+ aml_append(method, aml_store(handle, aml_name(NVDIMM_DSM_HANDLE)));
+ aml_append(method, aml_store(aml_arg(1), aml_name(NVDIMM_DSM_REVISION)));
+ aml_append(method, aml_store(aml_arg(2), aml_name(NVDIMM_DSM_FUNCTION)));
/*
* The fourth parameter (Arg3) of _DSM is a package which contains
@@ -912,23 +927,23 @@ static void nvdimm_build_common_dsm(Aml *dev)
pckg_buf = aml_local(3);
aml_append(ifctx, aml_store(aml_index(pckg, aml_int(0)), pckg_index));
aml_append(ifctx, aml_store(aml_derefof(pckg_index), pckg_buf));
- aml_append(ifctx, aml_store(pckg_buf, aml_name("FARG")));
+ aml_append(ifctx, aml_store(pckg_buf, aml_name(NVDIMM_DSM_ARG3)));
aml_append(method, ifctx);
/*
* tell QEMU about the real address of DSM memory, then QEMU
* gets the control and fills the result in DSM memory.
*/
- aml_append(method, aml_store(dsm_mem, aml_name("NTFI")));
+ aml_append(method, aml_store(dsm_mem, aml_name(NVDIMM_DSM_NOTIFY)));
dsm_out_buf_size = aml_local(1);
/* RLEN is not included in the payload returned to guest. */
- aml_append(method, aml_subtract(aml_name("RLEN"), aml_int(4),
- dsm_out_buf_size));
+ aml_append(method, aml_subtract(aml_name(NVDIMM_DSM_OUT_BUF_SIZE),
+ aml_int(4), dsm_out_buf_size));
aml_append(method, aml_store(aml_shiftleft(dsm_out_buf_size, aml_int(3)),
dsm_out_buf_size));
- aml_append(method, aml_create_field(aml_name("ODAT"), aml_int(0),
- dsm_out_buf_size, "OBUF"));
+ aml_append(method, aml_create_field(aml_name(NVDIMM_DSM_OUT_BUF),
+ aml_int(0), dsm_out_buf_size, "OBUF"));
aml_append(method, aml_concatenate(aml_buffer(0, NULL), aml_name("OBUF"),
dsm_out_buf));
aml_append(method, aml_return(dsm_out_buf));
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 36/47] nvdimm acpi: prebuild nvdimm devices for available slots
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (34 preceding siblings ...)
2016-10-30 21:24 ` [Qemu-devel] [PULL 35/47] nvdimm acpi: use common macros instead of magic names Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer Michael S. Tsirkin
` (13 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Xiao Guangrong, Stefan Hajnoczi, Igor Mammedov,
Paolo Bonzini, Richard Henderson, Eduardo Habkost
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
For each NVDIMM present or intended to be supported by platform,
platform firmware also exposes an ACPI Namespace Device under
the root device
So it builds nvdimm devices for all slots to support vNVDIMM hotplug
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/mem/nvdimm.h | 3 ++-
hw/acpi/nvdimm.c | 41 ++++++++++++++++++++++++-----------------
hw/i386/acpi-build.c | 2 +-
3 files changed, 27 insertions(+), 19 deletions(-)
diff --git a/include/hw/mem/nvdimm.h b/include/hw/mem/nvdimm.h
index 1cfe9e0..63a2b20 100644
--- a/include/hw/mem/nvdimm.h
+++ b/include/hw/mem/nvdimm.h
@@ -112,5 +112,6 @@ typedef struct AcpiNVDIMMState AcpiNVDIMMState;
void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
FWCfgState *fw_cfg, Object *owner);
void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
- BIOSLinker *linker, GArray *dsm_dma_arrea);
+ BIOSLinker *linker, GArray *dsm_dma_arrea,
+ uint32_t ram_slots);
#endif
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index bb896c9..b8a2e62 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -961,12 +961,11 @@ static void nvdimm_build_device_dsm(Aml *dev, uint32_t handle)
aml_append(dev, method);
}
-static void nvdimm_build_nvdimm_devices(GSList *device_list, Aml *root_dev)
+static void nvdimm_build_nvdimm_devices(Aml *root_dev, uint32_t ram_slots)
{
- for (; device_list; device_list = device_list->next) {
- DeviceState *dev = device_list->data;
- int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
- NULL);
+ uint32_t slot;
+
+ for (slot = 0; slot < ram_slots; slot++) {
uint32_t handle = nvdimm_slot_to_handle(slot);
Aml *nvdimm_dev;
@@ -987,9 +986,9 @@ static void nvdimm_build_nvdimm_devices(GSList *device_list, Aml *root_dev)
}
}
-static void nvdimm_build_ssdt(GSList *device_list, GArray *table_offsets,
- GArray *table_data, BIOSLinker *linker,
- GArray *dsm_dma_arrea)
+static void nvdimm_build_ssdt(GArray *table_offsets, GArray *table_data,
+ BIOSLinker *linker, GArray *dsm_dma_arrea,
+ uint32_t ram_slots)
{
Aml *ssdt, *sb_scope, *dev;
int mem_addr_offset, nvdimm_ssdt;
@@ -1021,7 +1020,7 @@ static void nvdimm_build_ssdt(GSList *device_list, GArray *table_offsets,
/* 0 is reserved for root device. */
nvdimm_build_device_dsm(dev, 0);
- nvdimm_build_nvdimm_devices(device_list, dev);
+ nvdimm_build_nvdimm_devices(dev, ram_slots);
aml_append(sb_scope, dev);
aml_append(ssdt, sb_scope);
@@ -1046,17 +1045,25 @@ static void nvdimm_build_ssdt(GSList *device_list, GArray *table_offsets,
}
void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
- BIOSLinker *linker, GArray *dsm_dma_arrea)
+ BIOSLinker *linker, GArray *dsm_dma_arrea,
+ uint32_t ram_slots)
{
GSList *device_list;
- /* no NVDIMM device is plugged. */
device_list = nvdimm_get_plugged_device_list();
- if (!device_list) {
- return;
+
+ /* NVDIMM device is plugged. */
+ if (device_list) {
+ nvdimm_build_nfit(device_list, table_offsets, table_data, linker);
+ g_slist_free(device_list);
+ }
+
+ /*
+ * NVDIMM device is allowed to be plugged only if there is available
+ * slot.
+ */
+ if (ram_slots) {
+ nvdimm_build_ssdt(table_offsets, table_data, linker, dsm_dma_arrea,
+ ram_slots);
}
- nvdimm_build_nfit(device_list, table_offsets, table_data, linker);
- nvdimm_build_ssdt(device_list, table_offsets, table_data, linker,
- dsm_dma_arrea);
- g_slist_free(device_list);
}
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 93be96f..cec4b4e 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -2811,7 +2811,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
}
if (pcms->acpi_nvdimm_state.is_enabled) {
nvdimm_build_acpi(table_offsets, tables_blob, tables->linker,
- pcms->acpi_nvdimm_state.dsm_mem);
+ pcms->acpi_nvdimm_state.dsm_mem, machine->ram_slots);
}
/* Add tables supplied by user (if any) */
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (35 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 36/47] nvdimm acpi: prebuild nvdimm devices for available slots Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-31 9:45 ` Igor Mammedov
2016-10-30 21:25 ` [Qemu-devel] [PULL 38/47] nvdimm acpi: introduce _FIT Michael S. Tsirkin
` (12 subsequent siblings)
49 siblings, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov, Paolo Bonzini,
Richard Henderson, Eduardo Habkost
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
The buffer is used to save the FIT info for all the presented nvdimm
devices which is updated after the nvdimm device is plugged or
unplugged. In the later patch, it will be used to construct NVDIMM
ACPI _FIT method which reflects the presented nvdimm devices after
nvdimm hotplug
As FIT buffer can not completely mapped into guest address space,
OSPM will exit to QEMU multiple times, however, there is the race
condition - FIT may be changed during these multiple exits, so that
some rules are introduced:
1) the user should hold the @lock to access the buffer and
2) mark @dirty whenever the buffer is updated.
@dirty is cleared for the first time OSPM gets fit buffer, if
dirty is detected in the later access, OSPM will restart the
access
As fit should be updated after nvdimm device is successfully realized
so that a new hotplug callback, post_hotplug, is introduced
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/hotplug.h | 10 +++++++++
include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
hw/core/hotplug.c | 11 +++++++++
hw/core/qdev.c | 20 +++++++++++++----
hw/i386/acpi-build.c | 2 +-
hw/i386/pc.c | 19 ++++++++++++++++
7 files changed, 124 insertions(+), 23 deletions(-)
diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
index c0db869..10ca5b6 100644
--- a/include/hw/hotplug.h
+++ b/include/hw/hotplug.h
@@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
* @parent: Opaque parent interface.
* @pre_plug: pre plug callback called at start of device.realize(true)
* @plug: plug callback called at end of device.realize(true).
+ * @post_pug: post plug callback called after device is successfully plugged.
* @unplug_request: unplug request callback.
* Used as a means to initiate device unplug for devices that
* require asynchronous unplug handling.
@@ -61,6 +62,7 @@ typedef struct HotplugHandlerClass {
/* <public> */
hotplug_fn pre_plug;
hotplug_fn plug;
+ hotplug_fn post_plug;
hotplug_fn unplug_request;
hotplug_fn unplug;
} HotplugHandlerClass;
@@ -83,6 +85,14 @@ void hotplug_handler_pre_plug(HotplugHandler *plug_handler,
DeviceState *plugged_dev,
Error **errp);
+/**
+ * hotplug_handler_post_plug:
+ *
+ * Call #HotplugHandlerClass.post_plug callback of @plug_handler.
+ */
+void hotplug_handler_post_plug(HotplugHandler *plug_handler,
+ DeviceState *plugged_dev,
+ Error **errp);
/**
* hotplug_handler_unplug_request:
diff --git a/include/hw/mem/nvdimm.h b/include/hw/mem/nvdimm.h
index 63a2b20..33cd421 100644
--- a/include/hw/mem/nvdimm.h
+++ b/include/hw/mem/nvdimm.h
@@ -98,12 +98,35 @@ typedef struct NVDIMMClass NVDIMMClass;
#define NVDIMM_ACPI_IO_BASE 0x0a18
#define NVDIMM_ACPI_IO_LEN 4
+/*
+ * The buffer, @fit, saves the FIT info for all the presented NVDIMM
+ * devices which is updated after the NVDIMM device is plugged or
+ * unplugged.
+ *
+ * Rules to use the buffer:
+ * 1) the user should hold the @lock to access the buffer.
+ * 2) mark @dirty whenever the buffer is updated.
+ *
+ * These rules preserve NVDIMM ACPI _FIT method to read incomplete
+ * or obsolete fit info if fit update happens during multiple RFIT
+ * calls.
+ */
+struct NvdimmFitBuffer {
+ QemuMutex lock;
+ GArray *fit;
+ bool dirty;
+};
+typedef struct NvdimmFitBuffer NvdimmFitBuffer;
+
struct AcpiNVDIMMState {
/* detect if NVDIMM support is enabled. */
bool is_enabled;
/* the data of the fw_cfg file NVDIMM_DSM_MEM_FILE. */
GArray *dsm_mem;
+
+ NvdimmFitBuffer fit_buf;
+
/* the IO region used by OSPM to transfer control to QEMU. */
MemoryRegion io_mr;
};
@@ -112,6 +135,7 @@ typedef struct AcpiNVDIMMState AcpiNVDIMMState;
void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
FWCfgState *fw_cfg, Object *owner);
void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
- BIOSLinker *linker, GArray *dsm_dma_arrea,
+ BIOSLinker *linker, AcpiNVDIMMState *state,
uint32_t ram_slots);
+void nvdimm_acpi_hotplug(AcpiNVDIMMState *state);
#endif
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index b8a2e62..5f728a6 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -348,8 +348,9 @@ static void nvdimm_build_structure_dcr(GArray *structures, DeviceState *dev)
(DSM) in DSM Spec Rev1.*/);
}
-static GArray *nvdimm_build_device_structure(GSList *device_list)
+static GArray *nvdimm_build_device_structure(void)
{
+ GSList *device_list = nvdimm_get_plugged_device_list();
GArray *structures = g_array_new(false, true /* clear */, 1);
for (; device_list; device_list = device_list->next) {
@@ -367,28 +368,58 @@ static GArray *nvdimm_build_device_structure(GSList *device_list)
/* build NVDIMM Control Region Structure. */
nvdimm_build_structure_dcr(structures, dev);
}
+ g_slist_free(device_list);
return structures;
}
-static void nvdimm_build_nfit(GSList *device_list, GArray *table_offsets,
+static void nvdimm_init_fit_buffer(NvdimmFitBuffer *fit_buf)
+{
+ qemu_mutex_init(&fit_buf->lock);
+ fit_buf->fit = g_array_new(false, true /* clear */, 1);
+}
+
+static void nvdimm_build_fit_buffer(NvdimmFitBuffer *fit_buf)
+{
+ qemu_mutex_lock(&fit_buf->lock);
+ g_array_free(fit_buf->fit, true);
+ fit_buf->fit = nvdimm_build_device_structure();
+ fit_buf->dirty = true;
+ qemu_mutex_unlock(&fit_buf->lock);
+}
+
+void nvdimm_acpi_hotplug(AcpiNVDIMMState *state)
+{
+ nvdimm_build_fit_buffer(&state->fit_buf);
+}
+
+static void nvdimm_build_nfit(AcpiNVDIMMState *state, GArray *table_offsets,
GArray *table_data, BIOSLinker *linker)
{
- GArray *structures = nvdimm_build_device_structure(device_list);
+ NvdimmFitBuffer *fit_buf = &state->fit_buf;
unsigned int header;
+ qemu_mutex_lock(&fit_buf->lock);
+
+ /* NVDIMM device is not plugged? */
+ if (!fit_buf->fit->len) {
+ goto exit;
+ }
+
acpi_add_table(table_offsets, table_data);
/* NFIT header. */
header = table_data->len;
acpi_data_push(table_data, sizeof(NvdimmNfitHeader));
/* NVDIMM device structures. */
- g_array_append_vals(table_data, structures->data, structures->len);
+ g_array_append_vals(table_data, fit_buf->fit->data, fit_buf->fit->len);
build_header(linker, table_data,
(void *)(table_data->data + header), "NFIT",
- sizeof(NvdimmNfitHeader) + structures->len, 1, NULL, NULL);
- g_array_free(structures, true);
+ sizeof(NvdimmNfitHeader) + fit_buf->fit->len, 1, NULL, NULL);
+
+exit:
+ qemu_mutex_unlock(&fit_buf->lock);
}
struct NvdimmDsmIn {
@@ -771,6 +802,8 @@ void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
acpi_data_push(state->dsm_mem, sizeof(NvdimmDsmIn));
fw_cfg_add_file(fw_cfg, NVDIMM_DSM_MEM_FILE, state->dsm_mem->data,
state->dsm_mem->len);
+
+ nvdimm_init_fit_buffer(&state->fit_buf);
}
#define NVDIMM_COMMON_DSM "NCAL"
@@ -1045,25 +1078,17 @@ static void nvdimm_build_ssdt(GArray *table_offsets, GArray *table_data,
}
void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
- BIOSLinker *linker, GArray *dsm_dma_arrea,
+ BIOSLinker *linker, AcpiNVDIMMState *state,
uint32_t ram_slots)
{
- GSList *device_list;
-
- device_list = nvdimm_get_plugged_device_list();
-
- /* NVDIMM device is plugged. */
- if (device_list) {
- nvdimm_build_nfit(device_list, table_offsets, table_data, linker);
- g_slist_free(device_list);
- }
+ nvdimm_build_nfit(state, table_offsets, table_data, linker);
/*
* NVDIMM device is allowed to be plugged only if there is available
* slot.
*/
if (ram_slots) {
- nvdimm_build_ssdt(table_offsets, table_data, linker, dsm_dma_arrea,
+ nvdimm_build_ssdt(table_offsets, table_data, linker, state->dsm_mem,
ram_slots);
}
}
diff --git a/hw/core/hotplug.c b/hw/core/hotplug.c
index 17ac986..ab34c19 100644
--- a/hw/core/hotplug.c
+++ b/hw/core/hotplug.c
@@ -35,6 +35,17 @@ void hotplug_handler_plug(HotplugHandler *plug_handler,
}
}
+void hotplug_handler_post_plug(HotplugHandler *plug_handler,
+ DeviceState *plugged_dev,
+ Error **errp)
+{
+ HotplugHandlerClass *hdc = HOTPLUG_HANDLER_GET_CLASS(plug_handler);
+
+ if (hdc->post_plug) {
+ hdc->post_plug(plug_handler, plugged_dev, errp);
+ }
+}
+
void hotplug_handler_unplug_request(HotplugHandler *plug_handler,
DeviceState *plugged_dev,
Error **errp)
diff --git a/hw/core/qdev.c b/hw/core/qdev.c
index 5783442..d835e62 100644
--- a/hw/core/qdev.c
+++ b/hw/core/qdev.c
@@ -945,10 +945,21 @@ static void device_set_realized(Object *obj, bool value, Error **errp)
goto child_realize_fail;
}
}
+
if (dev->hotplugged) {
device_reset(dev);
}
dev->pending_deleted_event = false;
+ dev->realized = value;
+
+ if (hotplug_ctrl) {
+ hotplug_handler_post_plug(hotplug_ctrl, dev, &local_err);
+ }
+
+ if (local_err != NULL) {
+ dev->realized = value;
+ goto post_realize_fail;
+ }
} else if (!value && dev->realized) {
Error **local_errp = NULL;
QLIST_FOREACH(bus, &dev->child_bus, sibling) {
@@ -965,13 +976,14 @@ static void device_set_realized(Object *obj, bool value, Error **errp)
}
dev->pending_deleted_event = true;
DEVICE_LISTENER_CALL(unrealize, Reverse, dev);
- }
- if (local_err != NULL) {
- goto fail;
+ if (local_err != NULL) {
+ goto fail;
+ }
+
+ dev->realized = value;
}
- dev->realized = value;
return;
child_realize_fail:
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index cec4b4e..03a5386 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -2811,7 +2811,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
}
if (pcms->acpi_nvdimm_state.is_enabled) {
nvdimm_build_acpi(table_offsets, tables_blob, tables->linker,
- pcms->acpi_nvdimm_state.dsm_mem, machine->ram_slots);
+ &pcms->acpi_nvdimm_state, machine->ram_slots);
}
/* Add tables supplied by user (if any) */
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index f56ea0f..b395717 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1721,6 +1721,16 @@ out:
error_propagate(errp, local_err);
}
+static void pc_dimm_post_plug(HotplugHandler *hotplug_dev,
+ DeviceState *dev, Error **errp)
+{
+ PCMachineState *pcms = PC_MACHINE(hotplug_dev);
+
+ if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
+ nvdimm_acpi_hotplug(&pcms->acpi_nvdimm_state);
+ }
+}
+
static void pc_dimm_unplug_request(HotplugHandler *hotplug_dev,
DeviceState *dev, Error **errp)
{
@@ -1986,6 +1996,14 @@ static void pc_machine_device_plug_cb(HotplugHandler *hotplug_dev,
}
}
+static void pc_machine_device_post_plug_cb(HotplugHandler *hotplug_dev,
+ DeviceState *dev, Error **errp)
+{
+ if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
+ pc_dimm_post_plug(hotplug_dev, dev, errp);
+ }
+}
+
static void pc_machine_device_unplug_request_cb(HotplugHandler *hotplug_dev,
DeviceState *dev, Error **errp)
{
@@ -2290,6 +2308,7 @@ static void pc_machine_class_init(ObjectClass *oc, void *data)
mc->reset = pc_machine_reset;
hc->pre_plug = pc_machine_device_pre_plug_cb;
hc->plug = pc_machine_device_plug_cb;
+ hc->post_plug = pc_machine_device_post_plug_cb;
hc->unplug_request = pc_machine_device_unplug_request_cb;
hc->unplug = pc_machine_device_unplug_cb;
nc->nmi_monitor_handler = x86_nmi;
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 38/47] nvdimm acpi: introduce _FIT
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (36 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug Michael S. Tsirkin
` (11 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
_FIT is required for hotplug support, guest will inquire the updated
device info from it if a hotplug event is received
As FIT buffer is not completely mapped into guest address space, so a
new function, Read FIT whose UUID is UUID
648B9CF2-CDA1-4312-8AD9-49C4AF32BD62, handle 0x10000, function index
is 0x1, is reserved by QEMU to read the piece of FIT buffer. The buffer
is concatenated before _FIT return
Refer to docs/specs/acpi-nvdimm.txt for detailed design
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 204 ++++++++++++++++++++++++++++++++++++++++++++-
docs/specs/acpi_nvdimm.txt | 58 ++++++++++++-
2 files changed, 257 insertions(+), 5 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index 5f728a6..fc1a012 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -496,6 +496,22 @@ typedef struct NvdimmFuncSetLabelDataIn NvdimmFuncSetLabelDataIn;
QEMU_BUILD_BUG_ON(sizeof(NvdimmFuncSetLabelDataIn) +
offsetof(NvdimmDsmIn, arg3) > 4096);
+struct NvdimmFuncReadFITIn {
+ uint32_t offset; /* the offset of FIT buffer. */
+} QEMU_PACKED;
+typedef struct NvdimmFuncReadFITIn NvdimmFuncReadFITIn;
+QEMU_BUILD_BUG_ON(sizeof(NvdimmFuncReadFITIn) +
+ offsetof(NvdimmDsmIn, arg3) > 4096);
+
+struct NvdimmFuncReadFITOut {
+ /* the size of buffer filled by QEMU. */
+ uint32_t len;
+ uint32_t func_ret_status; /* return status code. */
+ uint8_t fit[0]; /* the FIT data. */
+} QEMU_PACKED;
+typedef struct NvdimmFuncReadFITOut NvdimmFuncReadFITOut;
+QEMU_BUILD_BUG_ON(sizeof(NvdimmFuncReadFITOut) > 4096);
+
static void
nvdimm_dsm_function0(uint32_t supported_func, hwaddr dsm_mem_addr)
{
@@ -516,6 +532,74 @@ nvdimm_dsm_no_payload(uint32_t func_ret_status, hwaddr dsm_mem_addr)
cpu_physical_memory_write(dsm_mem_addr, &out, sizeof(out));
}
+#define NVDIMM_QEMU_RSVD_HANDLE_ROOT 0x10000
+
+/* Read FIT data, defined in docs/specs/acpi_nvdimm.txt. */
+static void nvdimm_dsm_func_read_fit(AcpiNVDIMMState *state, NvdimmDsmIn *in,
+ hwaddr dsm_mem_addr)
+{
+ NvdimmFitBuffer *fit_buf = &state->fit_buf;
+ NvdimmFuncReadFITIn *read_fit;
+ NvdimmFuncReadFITOut *read_fit_out;
+ GArray *fit;
+ uint32_t read_len = 0, func_ret_status;
+ int size;
+
+ read_fit = (NvdimmFuncReadFITIn *)in->arg3;
+ le32_to_cpus(&read_fit->offset);
+
+ qemu_mutex_lock(&fit_buf->lock);
+ fit = fit_buf->fit;
+
+ nvdimm_debug("Read FIT: offset %#x FIT size %#x Dirty %s.\n",
+ read_fit->offset, fit->len, fit_buf->dirty ? "Yes" : "No");
+
+ if (read_fit->offset > fit->len) {
+ func_ret_status = 3 /* Invalid Input Parameters */;
+ goto exit;
+ }
+
+ /* It is the first time to read FIT. */
+ if (!read_fit->offset) {
+ fit_buf->dirty = false;
+ } else if (fit_buf->dirty) { /* FIT has been changed during RFIT. */
+ func_ret_status = 0x100 /* fit changed */;
+ goto exit;
+ }
+
+ func_ret_status = 0 /* Success */;
+ read_len = MIN(fit->len - read_fit->offset,
+ 4096 - sizeof(NvdimmFuncReadFITOut));
+
+exit:
+ size = sizeof(NvdimmFuncReadFITOut) + read_len;
+ read_fit_out = g_malloc(size);
+
+ read_fit_out->len = cpu_to_le32(size);
+ read_fit_out->func_ret_status = cpu_to_le32(func_ret_status);
+ memcpy(read_fit_out->fit, fit->data + read_fit->offset, read_len);
+
+ cpu_physical_memory_write(dsm_mem_addr, read_fit_out, size);
+
+ g_free(read_fit_out);
+ qemu_mutex_unlock(&fit_buf->lock);
+}
+
+static void nvdimm_dsm_reserved_root(AcpiNVDIMMState *state, NvdimmDsmIn *in,
+ hwaddr dsm_mem_addr)
+{
+ switch (in->function) {
+ case 0x0:
+ nvdimm_dsm_function0(0x1 | 1 << 1 /* Read FIT */, dsm_mem_addr);
+ return;
+ case 0x1 /*Read FIT */:
+ nvdimm_dsm_func_read_fit(state, in, dsm_mem_addr);
+ return;
+ }
+
+ nvdimm_dsm_no_payload(1 /* Not Supported */, dsm_mem_addr);
+}
+
static void nvdimm_dsm_root(NvdimmDsmIn *in, hwaddr dsm_mem_addr)
{
/*
@@ -742,6 +826,7 @@ nvdimm_dsm_read(void *opaque, hwaddr addr, unsigned size)
static void
nvdimm_dsm_write(void *opaque, hwaddr addr, uint64_t val, unsigned size)
{
+ AcpiNVDIMMState *state = opaque;
NvdimmDsmIn *in;
hwaddr dsm_mem_addr = val;
@@ -769,6 +854,11 @@ nvdimm_dsm_write(void *opaque, hwaddr addr, uint64_t val, unsigned size)
goto exit;
}
+ if (in->handle == NVDIMM_QEMU_RSVD_HANDLE_ROOT) {
+ nvdimm_dsm_reserved_root(state, in, dsm_mem_addr);
+ goto exit;
+ }
+
/* Handle 0 is reserved for NVDIMM Root Device. */
if (!in->handle) {
nvdimm_dsm_root(in, dsm_mem_addr);
@@ -821,9 +911,13 @@ void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
#define NVDIMM_DSM_OUT_BUF_SIZE "RLEN"
#define NVDIMM_DSM_OUT_BUF "ODAT"
+#define NVDIMM_DSM_RFIT_STATUS "RSTA"
+
+#define NVDIMM_QEMU_RSVD_UUID "648B9CF2-CDA1-4312-8AD9-49C4AF32BD62"
+
static void nvdimm_build_common_dsm(Aml *dev)
{
- Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem;
+ Aml *method, *ifctx, *function, *handle, *uuid, *dsm_mem, *elsectx2;
Aml *elsectx, *unsupport, *unpatched, *expected_uuid, *uuid_invalid;
Aml *pckg, *pckg_index, *pckg_buf, *field, *dsm_out_buf, *dsm_out_buf_size;
uint8_t byte_list[1];
@@ -912,9 +1006,15 @@ static void nvdimm_build_common_dsm(Aml *dev)
/* UUID for NVDIMM Root Device */, expected_uuid));
aml_append(method, ifctx);
elsectx = aml_else();
- aml_append(elsectx, aml_store(
+ ifctx = aml_if(aml_equal(handle, aml_int(NVDIMM_QEMU_RSVD_HANDLE_ROOT)));
+ aml_append(ifctx, aml_store(aml_touuid(NVDIMM_QEMU_RSVD_UUID
+ /* UUID for QEMU internal use */), expected_uuid));
+ aml_append(elsectx, ifctx);
+ elsectx2 = aml_else();
+ aml_append(elsectx2, aml_store(
aml_touuid("4309AC30-0D11-11E4-9191-0800200C9A66")
/* UUID for NVDIMM Devices */, expected_uuid));
+ aml_append(elsectx, elsectx2);
aml_append(method, elsectx);
uuid_invalid = aml_lnot(aml_equal(uuid, expected_uuid));
@@ -994,6 +1094,105 @@ static void nvdimm_build_device_dsm(Aml *dev, uint32_t handle)
aml_append(dev, method);
}
+static void nvdimm_build_fit(Aml *dev)
+{
+ Aml *method, *pkg, *buf, *buf_size, *offset, *call_result;
+ Aml *whilectx, *ifcond, *ifctx, *elsectx, *fit;
+
+ buf = aml_local(0);
+ buf_size = aml_local(1);
+ fit = aml_local(2);
+
+ aml_append(dev, aml_create_dword_field(aml_buffer(4, NULL),
+ aml_int(0), NVDIMM_DSM_RFIT_STATUS));
+
+ /* build helper function, RFIT. */
+ method = aml_method("RFIT", 1, AML_SERIALIZED);
+ aml_append(method, aml_create_dword_field(aml_buffer(4, NULL),
+ aml_int(0), "OFST"));
+
+ /* prepare input package. */
+ pkg = aml_package(1);
+ aml_append(method, aml_store(aml_arg(0), aml_name("OFST")));
+ aml_append(pkg, aml_name("OFST"));
+
+ /* call Read_FIT function. */
+ call_result = aml_call5(NVDIMM_COMMON_DSM,
+ aml_touuid(NVDIMM_QEMU_RSVD_UUID),
+ aml_int(1) /* Revision 1 */,
+ aml_int(0x1) /* Read FIT */,
+ pkg, aml_int(NVDIMM_QEMU_RSVD_HANDLE_ROOT));
+ aml_append(method, aml_store(call_result, buf));
+
+ /* handle _DSM result. */
+ aml_append(method, aml_create_dword_field(buf,
+ aml_int(0) /* offset at byte 0 */, "STAU"));
+
+ aml_append(method, aml_store(aml_name("STAU"),
+ aml_name(NVDIMM_DSM_RFIT_STATUS)));
+
+ /* if something is wrong during _DSM. */
+ ifcond = aml_equal(aml_int(0 /* Success */), aml_name("STAU"));
+ ifctx = aml_if(aml_lnot(ifcond));
+ aml_append(ifctx, aml_return(aml_buffer(0, NULL)));
+ aml_append(method, ifctx);
+
+ aml_append(method, aml_store(aml_sizeof(buf), buf_size));
+ aml_append(method, aml_subtract(buf_size,
+ aml_int(4) /* the size of "STAU" */,
+ buf_size));
+
+ /* if we read the end of fit. */
+ ifctx = aml_if(aml_equal(buf_size, aml_int(0)));
+ aml_append(ifctx, aml_return(aml_buffer(0, NULL)));
+ aml_append(method, ifctx);
+
+ aml_append(method, aml_store(aml_shiftleft(buf_size, aml_int(3)),
+ buf_size));
+ aml_append(method, aml_create_field(buf,
+ aml_int(4 * BITS_PER_BYTE), /* offset at byte 4.*/
+ buf_size, "BUFF"));
+ aml_append(method, aml_return(aml_name("BUFF")));
+ aml_append(dev, method);
+
+ /* build _FIT. */
+ method = aml_method("_FIT", 0, AML_SERIALIZED);
+ offset = aml_local(3);
+
+ aml_append(method, aml_store(aml_buffer(0, NULL), fit));
+ aml_append(method, aml_store(aml_int(0), offset));
+
+ whilectx = aml_while(aml_int(1));
+ aml_append(whilectx, aml_store(aml_call1("RFIT", offset), buf));
+ aml_append(whilectx, aml_store(aml_sizeof(buf), buf_size));
+
+ /*
+ * if fit buffer was changed during RFIT, read from the beginning
+ * again.
+ */
+ ifctx = aml_if(aml_equal(aml_name(NVDIMM_DSM_RFIT_STATUS),
+ aml_int(0x100 /* fit changed */)));
+ aml_append(ifctx, aml_store(aml_buffer(0, NULL), fit));
+ aml_append(ifctx, aml_store(aml_int(0), offset));
+ aml_append(whilectx, ifctx);
+
+ elsectx = aml_else();
+
+ /* finish fit read if no data is read out. */
+ ifctx = aml_if(aml_equal(buf_size, aml_int(0)));
+ aml_append(ifctx, aml_return(fit));
+ aml_append(elsectx, ifctx);
+
+ /* update the offset. */
+ aml_append(elsectx, aml_add(offset, buf_size, offset));
+ /* append the data we read out to the fit buffer. */
+ aml_append(elsectx, aml_concatenate(fit, buf, fit));
+ aml_append(whilectx, elsectx);
+ aml_append(method, whilectx);
+
+ aml_append(dev, method);
+}
+
static void nvdimm_build_nvdimm_devices(Aml *root_dev, uint32_t ram_slots)
{
uint32_t slot;
@@ -1052,6 +1251,7 @@ static void nvdimm_build_ssdt(GArray *table_offsets, GArray *table_data,
/* 0 is reserved for root device. */
nvdimm_build_device_dsm(dev, 0);
+ nvdimm_build_fit(dev);
nvdimm_build_nvdimm_devices(dev, ram_slots);
diff --git a/docs/specs/acpi_nvdimm.txt b/docs/specs/acpi_nvdimm.txt
index 0fdd251..4aa5e3d 100644
--- a/docs/specs/acpi_nvdimm.txt
+++ b/docs/specs/acpi_nvdimm.txt
@@ -127,6 +127,58 @@ _DSM process diagram:
| result from the page | | |
+--------------------------+ +--------------+
- _FIT implementation
- -------------------
- TODO (will fill it when nvdimm hotplug is introduced)
+Device Handle Reservation
+-------------------------
+As we mentioned above, byte 0 ~ byte 3 in the DSM memory save NVDIMM device
+handle. The handle is completely QEMU internal thing, the values in range
+[0, 0xFFFF] indicate nvdimm device (O means nvdimm root device named NVDR),
+other values are reserved by other purpose.
+
+Current reserved handle:
+0x10000 is reserved for QEMU internal DSM function called on the root
+device.
+
+QEMU internal use only _DSM function
+------------------------------------
+UUID, 648B9CF2-CDA1-4312-8AD9-49C4AF32BD62, is reserved for QEMU internal
+DSM function.
+
+There is the function introduced by QEMU and only used by QEMU internal.
+
+1) Read FIT
+ As we only reserved one page for NVDIMM ACPI it is impossible to map the
+ whole FIT data to guest's address space. This function is used by _FIT
+ method to read a piece of FIT data from QEMU.
+
+ Input parameters:
+ Arg0 – UUID {set to 648B9CF2-CDA1-4312-8AD9-49C4AF32BD62}
+ Arg1 – Revision ID (set to 1)
+ Arg2 - Function Index, 0x1
+ Arg3 - A package containing a buffer whose layout is as follows:
+
+ +----------+-------------+-------------+-----------------------------------+
+ | Filed | Byte Length | Byte Offset | Description |
+ +----------+-------------+-------------+-----------------------------------+
+ | offset | 4 | 0 | the offset of FIT buffer |
+ +----------+-------------+-------------+-----------------------------------+
+
+ Output:
+ +----------+-------------+-------------+-----------------------------------+
+ | Filed | Byte Length | Byte Offset | Description |
+ +----------+-------------+-------------+-----------------------------------+
+ | | | | return status codes |
+ | | | | 0x100 indicates fit has been |
+ | status | 4 | 0 | updated |
+ | | | | other follows Chapter 3 in DSM |
+ | | | | Spec Rev1 |
+ +----------+-------------+-------------+-----------------------------------+
+ | fit data | Varies | 4 | FIT data |
+ | | | | |
+ +----------+-------------+-------------+-----------------------------------+
+
+ The FIT offset is maintained by the caller itself, current offset plugs
+ the length returned by the function is the next offset we should read.
+ When all the FIT data has been read out, zero length is returned.
+
+ If it returns 0x100, OSPM should restart to read FIT (read from offset 0
+ again).
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (37 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 38/47] nvdimm acpi: introduce _FIT Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-11-02 11:19 ` Igor Mammedov
2016-10-30 21:25 ` [Qemu-devel] [PULL 40/47] ipmi: Remove hotplug from IPMI BMCs Michael S. Tsirkin
` (10 subsequent siblings)
49 siblings, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Xiao Guangrong, Igor Mammedov, Paolo Bonzini,
Richard Henderson, Eduardo Habkost
From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
_GPE.E04 is dedicated for nvdimm device hotplug
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/acpi/acpi_dev_interface.h | 1 +
hw/acpi/memory_hotplug.c | 31 +++++++++++++++++++++++--------
hw/i386/acpi-build.c | 7 +++++++
hw/i386/pc.c | 12 ++++++++++++
hw/mem/nvdimm.c | 4 ----
docs/specs/acpi_mem_hotplug.txt | 3 +++
6 files changed, 46 insertions(+), 12 deletions(-)
diff --git a/include/hw/acpi/acpi_dev_interface.h b/include/hw/acpi/acpi_dev_interface.h
index da4ef7f..901a4ae 100644
--- a/include/hw/acpi/acpi_dev_interface.h
+++ b/include/hw/acpi/acpi_dev_interface.h
@@ -10,6 +10,7 @@ typedef enum {
ACPI_PCI_HOTPLUG_STATUS = 2,
ACPI_CPU_HOTPLUG_STATUS = 4,
ACPI_MEMORY_HOTPLUG_STATUS = 8,
+ ACPI_NVDIMM_HOTPLUG_STATUS = 16,
} AcpiEventStatusBits;
#define TYPE_ACPI_DEVICE_IF "acpi-device-interface"
diff --git a/hw/acpi/memory_hotplug.c b/hw/acpi/memory_hotplug.c
index ec4e64b..70f6451 100644
--- a/hw/acpi/memory_hotplug.c
+++ b/hw/acpi/memory_hotplug.c
@@ -2,6 +2,7 @@
#include "hw/acpi/memory_hotplug.h"
#include "hw/acpi/pc-hotplug.h"
#include "hw/mem/pc-dimm.h"
+#include "hw/mem/nvdimm.h"
#include "hw/boards.h"
#include "hw/qdev-core.h"
#include "trace.h"
@@ -232,11 +233,8 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
DeviceState *dev, Error **errp)
{
MemStatus *mdev;
- DeviceClass *dc = DEVICE_GET_CLASS(dev);
-
- if (!dc->hotpluggable) {
- return;
- }
+ AcpiEventStatusBits event;
+ bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
mdev = acpi_memory_slot_status(mem_st, dev, errp);
if (!mdev) {
@@ -244,10 +242,23 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
}
mdev->dimm = dev;
- mdev->is_enabled = true;
+
+ /*
+ * do not set is_enabled and is_inserting if the slot is plugged with
+ * a nvdimm device to stop OSPM inquires memory region from the slot.
+ */
+ if (is_nvdimm) {
+ event = ACPI_NVDIMM_HOTPLUG_STATUS;
+ } else {
+ mdev->is_enabled = true;
+ event = ACPI_MEMORY_HOTPLUG_STATUS;
+ }
+
if (dev->hotplugged) {
- mdev->is_inserting = true;
- acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
+ if (!is_nvdimm) {
+ mdev->is_inserting = true;
+ }
+ acpi_send_event(DEVICE(hotplug_dev), event);
}
}
@@ -262,6 +273,8 @@ void acpi_memory_unplug_request_cb(HotplugHandler *hotplug_dev,
return;
}
+ /* nvdimm device hot unplug is not supported yet. */
+ assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
mdev->is_removing = true;
acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
}
@@ -276,6 +289,8 @@ void acpi_memory_unplug_cb(MemHotplugState *mem_st,
return;
}
+ /* nvdimm device hot unplug is not supported yet. */
+ assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
mdev->is_enabled = false;
mdev->dimm = NULL;
}
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 03a5386..7aaa07a 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -2069,6 +2069,13 @@ build_dsdt(GArray *table_data, BIOSLinker *linker,
method = aml_method("_E03", 0, AML_NOTSERIALIZED);
aml_append(method, aml_call0(MEMORY_HOTPLUG_HANDLER_PATH));
aml_append(scope, method);
+
+ if (pcms->acpi_nvdimm_state.is_enabled) {
+ method = aml_method("_E04", 0, AML_NOTSERIALIZED);
+ aml_append(method, aml_notify(aml_name("\\_SB.NVDR"),
+ aml_int(0x80)));
+ aml_append(scope, method);
+ }
}
aml_append(dsdt, scope);
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index b395717..c011552 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1744,6 +1744,12 @@ static void pc_dimm_unplug_request(HotplugHandler *hotplug_dev,
goto out;
}
+ if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
+ error_setg(&local_err,
+ "nvdimm device hot unplug is not supported yet.");
+ goto out;
+ }
+
hhc = HOTPLUG_HANDLER_GET_CLASS(pcms->acpi_dev);
hhc->unplug_request(HOTPLUG_HANDLER(pcms->acpi_dev), dev, &local_err);
@@ -1761,6 +1767,12 @@ static void pc_dimm_unplug(HotplugHandler *hotplug_dev,
HotplugHandlerClass *hhc;
Error *local_err = NULL;
+ if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
+ error_setg(&local_err,
+ "nvdimm device hot unplug is not supported yet.");
+ goto out;
+ }
+
hhc = HOTPLUG_HANDLER_GET_CLASS(pcms->acpi_dev);
hhc->unplug(HOTPLUG_HANDLER(pcms->acpi_dev), dev, &local_err);
diff --git a/hw/mem/nvdimm.c b/hw/mem/nvdimm.c
index 7895805..db896b0 100644
--- a/hw/mem/nvdimm.c
+++ b/hw/mem/nvdimm.c
@@ -148,13 +148,9 @@ static MemoryRegion *nvdimm_get_vmstate_memory_region(PCDIMMDevice *dimm)
static void nvdimm_class_init(ObjectClass *oc, void *data)
{
- DeviceClass *dc = DEVICE_CLASS(oc);
PCDIMMDeviceClass *ddc = PC_DIMM_CLASS(oc);
NVDIMMClass *nvc = NVDIMM_CLASS(oc);
- /* nvdimm hotplug has not been supported yet. */
- dc->hotpluggable = false;
-
ddc->realize = nvdimm_realize;
ddc->get_memory_region = nvdimm_get_memory_region;
ddc->get_vmstate_memory_region = nvdimm_get_vmstate_memory_region;
diff --git a/docs/specs/acpi_mem_hotplug.txt b/docs/specs/acpi_mem_hotplug.txt
index 3df3620..cb26dd2 100644
--- a/docs/specs/acpi_mem_hotplug.txt
+++ b/docs/specs/acpi_mem_hotplug.txt
@@ -4,6 +4,9 @@ QEMU<->ACPI BIOS memory hotplug interface
ACPI BIOS GPE.3 handler is dedicated for notifying OS about memory hot-add
and hot-remove events.
+ACPI BIOS GPE.4 handler is dedicated for notifying OS about nvdimm device
+hot-add and hot-remove events.
+
Memory hot-plug interface (IO port 0xa00-0xa17, 1-4 byte access):
---------------------------------------------------------------
0xa00:
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 40/47] ipmi: Remove hotplug from IPMI BMCs
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (38 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 41/47] ipmi_bmc_sim: Remove an unnecessary mutex Michael S. Tsirkin
` (9 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Corey Minyard, Marc-André Lureau,
Paolo Bonzini, Cédric Le Goater, Marcel Apfelbaum,
Greg Kurz
From: Corey Minyard <cminyard@mvista.com>
No hotplug support, make sure it doesn't happen.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/ipmi_bmc_extern.c | 1 +
hw/ipmi/ipmi_bmc_sim.c | 1 +
2 files changed, 2 insertions(+)
diff --git a/hw/ipmi/ipmi_bmc_extern.c b/hw/ipmi/ipmi_bmc_extern.c
index 4b310e5..d30b286 100644
--- a/hw/ipmi/ipmi_bmc_extern.c
+++ b/hw/ipmi/ipmi_bmc_extern.c
@@ -512,6 +512,7 @@ static void ipmi_bmc_extern_class_init(ObjectClass *oc, void *data)
bk->handle_command = ipmi_bmc_extern_handle_command;
bk->handle_reset = ipmi_bmc_extern_handle_reset;
+ dc->hotpluggable = false;
dc->realize = ipmi_bmc_extern_realize;
dc->props = ipmi_bmc_extern_properties;
}
diff --git a/hw/ipmi/ipmi_bmc_sim.c b/hw/ipmi/ipmi_bmc_sim.c
index 17c7c0e..a0282cb 100644
--- a/hw/ipmi/ipmi_bmc_sim.c
+++ b/hw/ipmi/ipmi_bmc_sim.c
@@ -1791,6 +1791,7 @@ static void ipmi_sim_class_init(ObjectClass *oc, void *data)
DeviceClass *dc = DEVICE_CLASS(oc);
IPMIBmcClass *bk = IPMI_BMC_CLASS(oc);
+ dc->hotpluggable = false;
dc->realize = ipmi_sim_realize;
bk->handle_command = ipmi_sim_handle_command;
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 41/47] ipmi_bmc_sim: Remove an unnecessary mutex
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (39 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 40/47] ipmi: Remove hotplug from IPMI BMCs Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 42/47] ipmi: chassis poweroff should use qemu_system_shutdown_request() Michael S. Tsirkin
` (8 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Corey Minyard, Marc-André Lureau,
Cédric Le Goater, Marcel Apfelbaum, Greg Kurz
From: Corey Minyard <cminyard@mvista.com>
Get rid of the unnecessary mutex, it was a vestige
of something else that was not done. That way we don't
have to free it.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/ipmi_bmc_sim.c | 6 ------
1 file changed, 6 deletions(-)
diff --git a/hw/ipmi/ipmi_bmc_sim.c b/hw/ipmi/ipmi_bmc_sim.c
index a0282cb..c7883d6 100644
--- a/hw/ipmi/ipmi_bmc_sim.c
+++ b/hw/ipmi/ipmi_bmc_sim.c
@@ -217,7 +217,6 @@ struct IPMIBmcSim {
/* Odd netfns are for responses, so we only need the even ones. */
const IPMINetfn *netfns[MAX_NETFNS / 2];
- QemuMutex lock;
/* We allow one event in the buffer */
uint8_t evtbuf[16];
@@ -940,7 +939,6 @@ static void get_msg(IPMIBmcSim *ibs,
{
IPMIRcvBufEntry *msg;
- qemu_mutex_lock(&ibs->lock);
if (QTAILQ_EMPTY(&ibs->rcvbufs)) {
rsp_buffer_set_error(rsp, 0x80); /* Queue empty */
goto out;
@@ -960,7 +958,6 @@ static void get_msg(IPMIBmcSim *ibs,
}
out:
- qemu_mutex_unlock(&ibs->lock);
return;
}
@@ -1055,11 +1052,9 @@ static void send_msg(IPMIBmcSim *ibs,
end_msg:
msg->buf[msg->len] = ipmb_checksum(msg->buf, msg->len, 0);
msg->len++;
- qemu_mutex_lock(&ibs->lock);
QTAILQ_INSERT_TAIL(&ibs->rcvbufs, msg, entry);
ibs->msg_flags |= IPMI_BMC_MSG_FLAG_RCV_MSG_QUEUE;
k->set_atn(s, 1, attn_irq_enabled(ibs));
- qemu_mutex_unlock(&ibs->lock);
}
static void do_watchdog_reset(IPMIBmcSim *ibs)
@@ -1753,7 +1748,6 @@ static void ipmi_sim_realize(DeviceState *dev, Error **errp)
unsigned int i;
IPMIBmcSim *ibs = IPMI_BMC_SIMULATOR(b);
- qemu_mutex_init(&ibs->lock);
QTAILQ_INIT(&ibs->rcvbufs);
ibs->bmc_global_enables = (1 << IPMI_BMC_EVENT_LOG_BIT);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 42/47] ipmi: chassis poweroff should use qemu_system_shutdown_request()
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (40 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 41/47] ipmi_bmc_sim: Remove an unnecessary mutex Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 43/47] ipmi: Implement shutdown via ACPI overtemp Michael S. Tsirkin
` (7 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Cédric Le Goater, Corey Minyard, Paolo Bonzini
From: Cédric Le Goater <clg@kaod.org>
When issuing a chassis 'powerdown' control command, the routine
qemu_system_shutdown_request() should be used to exit the guest.
qemu_system_powerdown_request() will initiate a soft shutdown which is
not what is required by the IPMI (28.3 Chassis Control Command):
0h = power down. Force system into soft off (S4/S45) state. This
is for 'emergency' management power down actions. The command does
not initiate a clean shut-down of the operating system prior to
powering down the system
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/ipmi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/ipmi/ipmi.c b/hw/ipmi/ipmi.c
index f09f217..f91c7b7 100644
--- a/hw/ipmi/ipmi.c
+++ b/hw/ipmi/ipmi.c
@@ -51,7 +51,7 @@ static int ipmi_do_hw_op(IPMIInterface *s, enum ipmi_op op, int checkonly)
if (checkonly) {
return 0;
}
- qemu_system_powerdown_request();
+ qemu_system_shutdown_request();
return 0;
case IPMI_SEND_NMI:
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 43/47] ipmi: Implement shutdown via ACPI overtemp
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (41 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 42/47] ipmi: chassis poweroff should use qemu_system_shutdown_request() Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 44/47] ipmi: fix build config variable name for ipmi_bmc_extern.o Michael S. Tsirkin
` (6 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Corey Minyard, Cédric Le Goater, Paolo Bonzini
From: Corey Minyard <cminyard@mvista.com>
This is allowed by the IPMI specification for graceful shutdown,
so implement it.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/ipmi.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/hw/ipmi/ipmi.c b/hw/ipmi/ipmi.c
index f91c7b7..5cf1caa 100644
--- a/hw/ipmi/ipmi.c
+++ b/hw/ipmi/ipmi.c
@@ -61,9 +61,15 @@ static int ipmi_do_hw_op(IPMIInterface *s, enum ipmi_op op, int checkonly)
qmp_inject_nmi(NULL);
return 0;
+ case IPMI_SHUTDOWN_VIA_ACPI_OVERTEMP:
+ if (checkonly) {
+ return 0;
+ }
+ qemu_system_powerdown_request();
+ return 0;
+
case IPMI_POWERCYCLE_CHASSIS:
case IPMI_PULSE_DIAG_IRQ:
- case IPMI_SHUTDOWN_VIA_ACPI_OVERTEMP:
case IPMI_POWERON_CHASSIS:
default:
return IPMI_CC_COMMAND_NOT_SUPPORTED;
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 44/47] ipmi: fix build config variable name for ipmi_bmc_extern.o
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (42 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 43/47] ipmi: Implement shutdown via ACPI overtemp Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 45/47] ipmi: Add graceful shutdown handling to the external BMC Michael S. Tsirkin
` (5 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Daniel P. Berrange, Corey Minyard
From: "Daniel P. Berrange" <berrange@redhat.com>
The original commit:
commit 67aa56fc03bea44ccf384ea400515a8a58844a50
Author: Corey Minyard <cminyard@mvista.com>
Date: Thu Dec 17 12:50:06 2015 -0600
ipmi: Add an external connection simulation interface
defined a new variable CONFIG_IPMI_EXTERN, but then went
on to mistakely use the pre-existing CONFIG_IPMI_LOCAL
variable.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/Makefile.objs | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/ipmi/Makefile.objs b/hw/ipmi/Makefile.objs
index a90318d..1b422bb 100644
--- a/hw/ipmi/Makefile.objs
+++ b/hw/ipmi/Makefile.objs
@@ -1,5 +1,5 @@
common-obj-$(CONFIG_IPMI) += ipmi.o
common-obj-$(CONFIG_IPMI_LOCAL) += ipmi_bmc_sim.o
-common-obj-$(CONFIG_IPMI_LOCAL) += ipmi_bmc_extern.o
+common-obj-$(CONFIG_IPMI_EXTERN) += ipmi_bmc_extern.o
common-obj-$(CONFIG_ISA_IPMI_KCS) += isa_ipmi_kcs.o
common-obj-$(CONFIG_ISA_IPMI_BT) += isa_ipmi_bt.o
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 45/47] ipmi: Add graceful shutdown handling to the external BMC
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (43 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 44/47] ipmi: fix build config variable name for ipmi_bmc_extern.o Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 46/47] acpi/ipmi: Initialize the fwinfo before fetching it Michael S. Tsirkin
` (4 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Corey Minyard, Paolo Bonzini,
Marc-André Lureau, Daniel P. Berrange, Eric Blake,
Michael Tokarev
From: Corey Minyard <cminyard@mvista.com>
I misunderstood the workings of the power settings, the power off
is a force off operation and there needs to be a separate graceful
shutdown operation. So replace the force off operation with a
graceful shutdown.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/ipmi/ipmi_bmc_extern.c | 11 ++++++++---
tests/ipmi-bt-test.c | 2 +-
2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/hw/ipmi/ipmi_bmc_extern.c b/hw/ipmi/ipmi_bmc_extern.c
index d30b286..e8e3d25 100644
--- a/hw/ipmi/ipmi_bmc_extern.c
+++ b/hw/ipmi/ipmi_bmc_extern.c
@@ -54,7 +54,8 @@
#define VM_CAPABILITIES_IRQ 0x04
#define VM_CAPABILITIES_NMI 0x08
#define VM_CAPABILITIES_ATTN 0x10
-#define VM_CMD_FORCEOFF 0x09
+#define VM_CAPABILITIES_GRACEFUL_SHUTDOWN 0x20
+#define VM_CMD_GRACEFUL_SHUTDOWN 0x09
#define TYPE_IPMI_BMC_EXTERN "ipmi-bmc-extern"
#define IPMI_BMC_EXTERN(obj) OBJECT_CHECK(IPMIBmcExtern, (obj), \
@@ -276,8 +277,8 @@ static void handle_hw_op(IPMIBmcExtern *ibe, unsigned char hw_op)
k->do_hw_op(s, IPMI_SEND_NMI, 0);
break;
- case VM_CMD_FORCEOFF:
- qemu_system_shutdown_request();
+ case VM_CMD_GRACEFUL_SHUTDOWN:
+ k->do_hw_op(s, IPMI_SHUTDOWN_VIA_ACPI_OVERTEMP, 0);
break;
}
}
@@ -401,6 +402,10 @@ static void chr_event(void *opaque, int event)
if (k->do_hw_op(ibe->parent.intf, IPMI_POWEROFF_CHASSIS, 1) == 0) {
v |= VM_CAPABILITIES_POWER;
}
+ if (k->do_hw_op(ibe->parent.intf, IPMI_SHUTDOWN_VIA_ACPI_OVERTEMP, 1)
+ == 0) {
+ v |= VM_CAPABILITIES_GRACEFUL_SHUTDOWN;
+ }
if (k->do_hw_op(ibe->parent.intf, IPMI_RESET_CHASSIS, 1) == 0) {
v |= VM_CAPABILITIES_RESET;
}
diff --git a/tests/ipmi-bt-test.c b/tests/ipmi-bt-test.c
index be9005e..65d05b3 100644
--- a/tests/ipmi-bt-test.c
+++ b/tests/ipmi-bt-test.c
@@ -309,7 +309,7 @@ static void test_connect(void)
uint8_t msg[100];
unsigned int msglen;
static uint8_t exp1[] = { 0xff, 0x01, 0xa1 }; /* A protocol version */
- static uint8_t exp2[] = { 0x08, 0x1f, 0xa1 }; /* A capabilities cmd */
+ static uint8_t exp2[] = { 0x08, 0x3f, 0xa1 }; /* A capabilities cmd */
FD_ZERO(&readfds);
FD_SET(emu_lfd, &readfds);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 46/47] acpi/ipmi: Initialize the fwinfo before fetching it
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (44 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 45/47] ipmi: Add graceful shutdown handling to the external BMC Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 47/47] acpi: fix assert failure caused by commit 35c5a52d Michael S. Tsirkin
` (3 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Corey Minyard, qemu-stable, Igor Mammedov
From: Corey Minyard <cminyard@mvista.com>
The initialization was missed before, resulting in some
bad data in the smbus case.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/ipmi.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/hw/acpi/ipmi.c b/hw/acpi/ipmi.c
index 7e74ce4..651e2e9 100644
--- a/hw/acpi/ipmi.c
+++ b/hw/acpi/ipmi.c
@@ -99,6 +99,7 @@ void build_acpi_ipmi_devices(Aml *scope, BusState *bus)
ii = IPMI_INTERFACE(obj);
iic = IPMI_INTERFACE_GET_CLASS(obj);
+ memset(&info, 0, sizeof(info));
iic->get_fwinfo(ii, &info);
aml_append(scope, aml_ipmi_device(&info));
}
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* [Qemu-devel] [PULL 47/47] acpi: fix assert failure caused by commit 35c5a52d
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (45 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 46/47] acpi/ipmi: Initialize the fwinfo before fetching it Michael S. Tsirkin
@ 2016-10-30 21:25 ` Michael S. Tsirkin
2016-10-31 9:50 ` [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Igor Mammedov
` (2 subsequent siblings)
49 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-30 21:25 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Haozhong Zhang, Dan Williams, Xiao Guangrong,
Igor Mammedov
From: Haozhong Zhang <haozhong.zhang@intel.com>
Commit 35c5a52d "acpi: do not use TARGET_PAGE_SIZE" changed struct
NvdimmDsmIn from a variable-size structure to a fixed-size structure of
4096 bytes. It forgot to adjust an assert in
nvdimm_dsm_set_label_data(..., NvdimmDsmIn *in, ...):
assert(sizeof(*in) + sizeof(*set_label_data) + set_label_data->length <=
4096);
which could crash QEMU when guest writes NVDIMM labels.
Fix it by replacing sizeof(*in) by offsetof(NvdimmDsmIn, arg3).
Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
Reported-by: Dan Williams <dan.j.williams@intel.com>
Tested-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/acpi/nvdimm.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index fc1a012..602ec54 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -757,8 +757,8 @@ static void nvdimm_dsm_set_label_data(NVDIMMDevice *nvdimm, NvdimmDsmIn *in,
return;
}
- assert(sizeof(*in) + sizeof(*set_label_data) + set_label_data->length <=
- 4096);
+ assert(offsetof(NvdimmDsmIn, arg3) +
+ sizeof(*set_label_data) + set_label_data->length <= 4096);
nvc->write_label_data(nvdimm, set_label_data->in_buf,
set_label_data->length, set_label_data->offset);
--
MST
^ permalink raw reply related [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-30 21:24 ` [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base Michael S. Tsirkin
@ 2016-10-31 9:20 ` Igor Mammedov
2016-10-31 9:23 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 9:20 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell, Xiao Guangrong
On Sun, 30 Oct 2016 23:24:46 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>
> According to ACPI 6.0 spec, "Memory Device Physical Address
> Region Base" in memdev is defined as "This field provides the
> Device Physical Address base of the region". This field should
> be zero in our case
I'm not sure that it should be a zero,
care to point source which tells that it should be zero?
> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> hw/acpi/nvdimm.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
> index bbb2cfd..c2f5caa 100644
> --- a/hw/acpi/nvdimm.c
> +++ b/hw/acpi/nvdimm.c
> @@ -289,8 +289,6 @@ static void
> nvdimm_build_structure_memdev(GArray *structures, DeviceState *dev)
> {
> NvdimmNfitMemDev *nfit_memdev;
> - uint64_t addr = object_property_get_int(OBJECT(dev), PC_DIMM_ADDR_PROP,
> - NULL);
> uint64_t size = object_property_get_int(OBJECT(dev), PC_DIMM_SIZE_PROP,
> NULL);
> int slot = object_property_get_int(OBJECT(dev), PC_DIMM_SLOT_PROP,
> @@ -314,7 +312,8 @@ nvdimm_build_structure_memdev(GArray *structures, DeviceState *dev)
>
> /* The memory region on the device. */
> nfit_memdev->region_len = cpu_to_le64(size);
> - nfit_memdev->region_dpa = cpu_to_le64(addr);
> + /* The device address starts from 0. */
> + nfit_memdev->region_dpa = cpu_to_le64(0);
>
> /* Only one interleave for PMEM. */
> nfit_memdev->interleave_ways = cpu_to_le16(1);
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-31 9:20 ` Igor Mammedov
@ 2016-10-31 9:23 ` Xiao Guangrong
2016-10-31 10:56 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-10-31 9:23 UTC (permalink / raw)
To: Igor Mammedov, Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell
On 10/31/2016 05:20 PM, Igor Mammedov wrote:
> On Sun, 30 Oct 2016 23:24:46 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>
>> According to ACPI 6.0 spec, "Memory Device Physical Address
>> Region Base" in memdev is defined as "This field provides the
>> Device Physical Address base of the region". This field should
>> be zero in our case
> I'm not sure that it should be a zero,
> care to point source which tells that it should be zero?
The spec says that this is the Device Physical Address, so that
it is the device internal address, it should be zero as we do not
reserve any thing in device internal and we do not have no memory
interleave.
Actually, this bug was exported when we were enabling nvdimm in
windows guest.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-10-30 21:25 ` [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer Michael S. Tsirkin
@ 2016-10-31 9:45 ` Igor Mammedov
2016-10-31 9:52 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 9:45 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: qemu-devel, Peter Maydell, Xiao Guangrong, Paolo Bonzini,
Richard Henderson, Eduardo Habkost
On Sun, 30 Oct 2016 23:25:05 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>
> The buffer is used to save the FIT info for all the presented nvdimm
> devices which is updated after the nvdimm device is plugged or
> unplugged. In the later patch, it will be used to construct NVDIMM
> ACPI _FIT method which reflects the presented nvdimm devices after
> nvdimm hotplug
>
> As FIT buffer can not completely mapped into guest address space,
> OSPM will exit to QEMU multiple times, however, there is the race
> condition - FIT may be changed during these multiple exits, so that
> some rules are introduced:
> 1) the user should hold the @lock to access the buffer and
> 2) mark @dirty whenever the buffer is updated.
>
> @dirty is cleared for the first time OSPM gets fit buffer, if
> dirty is detected in the later access, OSPM will restart the
> access
>
> As fit should be updated after nvdimm device is successfully realized
> so that a new hotplug callback, post_hotplug, is introduced
>
> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> include/hw/hotplug.h | 10 +++++++++
> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
> hw/core/hotplug.c | 11 +++++++++
> hw/core/qdev.c | 20 +++++++++++++----
> hw/i386/acpi-build.c | 2 +-
> hw/i386/pc.c | 19 ++++++++++++++++
> 7 files changed, 124 insertions(+), 23 deletions(-)
>
> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
> index c0db869..10ca5b6 100644
> --- a/include/hw/hotplug.h
> +++ b/include/hw/hotplug.h
> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
> * @parent: Opaque parent interface.
> * @pre_plug: pre plug callback called at start of device.realize(true)
> * @plug: plug callback called at end of device.realize(true).
> + * @post_pug: post plug callback called after device is successfully plugged.
this doesn't seem to be necessary,
why hotplug_handler_plug() isn't sufficient?
> * @unplug_request: unplug request callback.
> * Used as a means to initiate device unplug for devices that
> * require asynchronous unplug handling.
> @@ -61,6 +62,7 @@ typedef struct HotplugHandlerClass {
> /* <public> */
> hotplug_fn pre_plug;
> hotplug_fn plug;
> + hotplug_fn post_plug;
> hotplug_fn unplug_request;
> hotplug_fn unplug;
> } HotplugHandlerClass;
> @@ -83,6 +85,14 @@ void hotplug_handler_pre_plug(HotplugHandler *plug_handler,
> DeviceState *plugged_dev,
> Error **errp);
>
> +/**
> + * hotplug_handler_post_plug:
> + *
> + * Call #HotplugHandlerClass.post_plug callback of @plug_handler.
> + */
> +void hotplug_handler_post_plug(HotplugHandler *plug_handler,
> + DeviceState *plugged_dev,
> + Error **errp);
>
> /**
> * hotplug_handler_unplug_request:
> diff --git a/include/hw/mem/nvdimm.h b/include/hw/mem/nvdimm.h
> index 63a2b20..33cd421 100644
> --- a/include/hw/mem/nvdimm.h
> +++ b/include/hw/mem/nvdimm.h
> @@ -98,12 +98,35 @@ typedef struct NVDIMMClass NVDIMMClass;
> #define NVDIMM_ACPI_IO_BASE 0x0a18
> #define NVDIMM_ACPI_IO_LEN 4
>
> +/*
> + * The buffer, @fit, saves the FIT info for all the presented NVDIMM
> + * devices which is updated after the NVDIMM device is plugged or
> + * unplugged.
> + *
> + * Rules to use the buffer:
> + * 1) the user should hold the @lock to access the buffer.
> + * 2) mark @dirty whenever the buffer is updated.
> + *
> + * These rules preserve NVDIMM ACPI _FIT method to read incomplete
> + * or obsolete fit info if fit update happens during multiple RFIT
> + * calls.
> + */
> +struct NvdimmFitBuffer {
> + QemuMutex lock;
> + GArray *fit;
> + bool dirty;
> +};
> +typedef struct NvdimmFitBuffer NvdimmFitBuffer;
> +
> struct AcpiNVDIMMState {
> /* detect if NVDIMM support is enabled. */
> bool is_enabled;
>
> /* the data of the fw_cfg file NVDIMM_DSM_MEM_FILE. */
> GArray *dsm_mem;
> +
> + NvdimmFitBuffer fit_buf;
> +
> /* the IO region used by OSPM to transfer control to QEMU. */
> MemoryRegion io_mr;
> };
> @@ -112,6 +135,7 @@ typedef struct AcpiNVDIMMState AcpiNVDIMMState;
> void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
> FWCfgState *fw_cfg, Object *owner);
> void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
> - BIOSLinker *linker, GArray *dsm_dma_arrea,
> + BIOSLinker *linker, AcpiNVDIMMState *state,
> uint32_t ram_slots);
> +void nvdimm_acpi_hotplug(AcpiNVDIMMState *state);
> #endif
> diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
> index b8a2e62..5f728a6 100644
> --- a/hw/acpi/nvdimm.c
> +++ b/hw/acpi/nvdimm.c
> @@ -348,8 +348,9 @@ static void nvdimm_build_structure_dcr(GArray *structures, DeviceState *dev)
> (DSM) in DSM Spec Rev1.*/);
> }
>
> -static GArray *nvdimm_build_device_structure(GSList *device_list)
> +static GArray *nvdimm_build_device_structure(void)
> {
> + GSList *device_list = nvdimm_get_plugged_device_list();
> GArray *structures = g_array_new(false, true /* clear */, 1);
>
> for (; device_list; device_list = device_list->next) {
> @@ -367,28 +368,58 @@ static GArray *nvdimm_build_device_structure(GSList *device_list)
> /* build NVDIMM Control Region Structure. */
> nvdimm_build_structure_dcr(structures, dev);
> }
> + g_slist_free(device_list);
>
> return structures;
> }
>
> -static void nvdimm_build_nfit(GSList *device_list, GArray *table_offsets,
> +static void nvdimm_init_fit_buffer(NvdimmFitBuffer *fit_buf)
> +{
> + qemu_mutex_init(&fit_buf->lock);
> + fit_buf->fit = g_array_new(false, true /* clear */, 1);
> +}
> +
> +static void nvdimm_build_fit_buffer(NvdimmFitBuffer *fit_buf)
> +{
> + qemu_mutex_lock(&fit_buf->lock);
> + g_array_free(fit_buf->fit, true);
> + fit_buf->fit = nvdimm_build_device_structure();
> + fit_buf->dirty = true;
> + qemu_mutex_unlock(&fit_buf->lock);
> +}
> +
> +void nvdimm_acpi_hotplug(AcpiNVDIMMState *state)
> +{
> + nvdimm_build_fit_buffer(&state->fit_buf);
> +}
> +
> +static void nvdimm_build_nfit(AcpiNVDIMMState *state, GArray *table_offsets,
> GArray *table_data, BIOSLinker *linker)
> {
> - GArray *structures = nvdimm_build_device_structure(device_list);
> + NvdimmFitBuffer *fit_buf = &state->fit_buf;
> unsigned int header;
>
> + qemu_mutex_lock(&fit_buf->lock);
> +
> + /* NVDIMM device is not plugged? */
> + if (!fit_buf->fit->len) {
> + goto exit;
> + }
> +
> acpi_add_table(table_offsets, table_data);
>
> /* NFIT header. */
> header = table_data->len;
> acpi_data_push(table_data, sizeof(NvdimmNfitHeader));
> /* NVDIMM device structures. */
> - g_array_append_vals(table_data, structures->data, structures->len);
> + g_array_append_vals(table_data, fit_buf->fit->data, fit_buf->fit->len);
>
> build_header(linker, table_data,
> (void *)(table_data->data + header), "NFIT",
> - sizeof(NvdimmNfitHeader) + structures->len, 1, NULL, NULL);
> - g_array_free(structures, true);
> + sizeof(NvdimmNfitHeader) + fit_buf->fit->len, 1, NULL, NULL);
> +
> +exit:
> + qemu_mutex_unlock(&fit_buf->lock);
> }
>
> struct NvdimmDsmIn {
> @@ -771,6 +802,8 @@ void nvdimm_init_acpi_state(AcpiNVDIMMState *state, MemoryRegion *io,
> acpi_data_push(state->dsm_mem, sizeof(NvdimmDsmIn));
> fw_cfg_add_file(fw_cfg, NVDIMM_DSM_MEM_FILE, state->dsm_mem->data,
> state->dsm_mem->len);
> +
> + nvdimm_init_fit_buffer(&state->fit_buf);
> }
>
> #define NVDIMM_COMMON_DSM "NCAL"
> @@ -1045,25 +1078,17 @@ static void nvdimm_build_ssdt(GArray *table_offsets, GArray *table_data,
> }
>
> void nvdimm_build_acpi(GArray *table_offsets, GArray *table_data,
> - BIOSLinker *linker, GArray *dsm_dma_arrea,
> + BIOSLinker *linker, AcpiNVDIMMState *state,
> uint32_t ram_slots)
> {
> - GSList *device_list;
> -
> - device_list = nvdimm_get_plugged_device_list();
> -
> - /* NVDIMM device is plugged. */
> - if (device_list) {
> - nvdimm_build_nfit(device_list, table_offsets, table_data, linker);
> - g_slist_free(device_list);
> - }
> + nvdimm_build_nfit(state, table_offsets, table_data, linker);
>
> /*
> * NVDIMM device is allowed to be plugged only if there is available
> * slot.
> */
> if (ram_slots) {
> - nvdimm_build_ssdt(table_offsets, table_data, linker, dsm_dma_arrea,
> + nvdimm_build_ssdt(table_offsets, table_data, linker, state->dsm_mem,
> ram_slots);
> }
> }
> diff --git a/hw/core/hotplug.c b/hw/core/hotplug.c
> index 17ac986..ab34c19 100644
> --- a/hw/core/hotplug.c
> +++ b/hw/core/hotplug.c
> @@ -35,6 +35,17 @@ void hotplug_handler_plug(HotplugHandler *plug_handler,
> }
> }
>
> +void hotplug_handler_post_plug(HotplugHandler *plug_handler,
> + DeviceState *plugged_dev,
> + Error **errp)
> +{
> + HotplugHandlerClass *hdc = HOTPLUG_HANDLER_GET_CLASS(plug_handler);
> +
> + if (hdc->post_plug) {
> + hdc->post_plug(plug_handler, plugged_dev, errp);
> + }
> +}
> +
> void hotplug_handler_unplug_request(HotplugHandler *plug_handler,
> DeviceState *plugged_dev,
> Error **errp)
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index 5783442..d835e62 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -945,10 +945,21 @@ static void device_set_realized(Object *obj, bool value, Error **errp)
> goto child_realize_fail;
> }
> }
> +
> if (dev->hotplugged) {
> device_reset(dev);
> }
> dev->pending_deleted_event = false;
> + dev->realized = value;
> +
> + if (hotplug_ctrl) {
> + hotplug_handler_post_plug(hotplug_ctrl, dev, &local_err);
> + }
> +
> + if (local_err != NULL) {
> + dev->realized = value;
> + goto post_realize_fail;
> + }
> } else if (!value && dev->realized) {
> Error **local_errp = NULL;
> QLIST_FOREACH(bus, &dev->child_bus, sibling) {
> @@ -965,13 +976,14 @@ static void device_set_realized(Object *obj, bool value, Error **errp)
> }
> dev->pending_deleted_event = true;
> DEVICE_LISTENER_CALL(unrealize, Reverse, dev);
> - }
>
> - if (local_err != NULL) {
> - goto fail;
> + if (local_err != NULL) {
> + goto fail;
> + }
> +
> + dev->realized = value;
> }
>
> - dev->realized = value;
> return;
>
> child_realize_fail:
> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> index cec4b4e..03a5386 100644
> --- a/hw/i386/acpi-build.c
> +++ b/hw/i386/acpi-build.c
> @@ -2811,7 +2811,7 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
> }
> if (pcms->acpi_nvdimm_state.is_enabled) {
> nvdimm_build_acpi(table_offsets, tables_blob, tables->linker,
> - pcms->acpi_nvdimm_state.dsm_mem, machine->ram_slots);
> + &pcms->acpi_nvdimm_state, machine->ram_slots);
> }
>
> /* Add tables supplied by user (if any) */
> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> index f56ea0f..b395717 100644
> --- a/hw/i386/pc.c
> +++ b/hw/i386/pc.c
> @@ -1721,6 +1721,16 @@ out:
> error_propagate(errp, local_err);
> }
>
> +static void pc_dimm_post_plug(HotplugHandler *hotplug_dev,
> + DeviceState *dev, Error **errp)
> +{
> + PCMachineState *pcms = PC_MACHINE(hotplug_dev);
> +
> + if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
> + nvdimm_acpi_hotplug(&pcms->acpi_nvdimm_state);
> + }
> +}
> +
> static void pc_dimm_unplug_request(HotplugHandler *hotplug_dev,
> DeviceState *dev, Error **errp)
> {
> @@ -1986,6 +1996,14 @@ static void pc_machine_device_plug_cb(HotplugHandler *hotplug_dev,
> }
> }
>
> +static void pc_machine_device_post_plug_cb(HotplugHandler *hotplug_dev,
> + DeviceState *dev, Error **errp)
> +{
> + if (object_dynamic_cast(OBJECT(dev), TYPE_PC_DIMM)) {
> + pc_dimm_post_plug(hotplug_dev, dev, errp);
> + }
> +}
> +
> static void pc_machine_device_unplug_request_cb(HotplugHandler *hotplug_dev,
> DeviceState *dev, Error **errp)
> {
> @@ -2290,6 +2308,7 @@ static void pc_machine_class_init(ObjectClass *oc, void *data)
> mc->reset = pc_machine_reset;
> hc->pre_plug = pc_machine_device_pre_plug_cb;
> hc->plug = pc_machine_device_plug_cb;
> + hc->post_plug = pc_machine_device_post_plug_cb;
> hc->unplug_request = pc_machine_device_unplug_request_cb;
> hc->unplug = pc_machine_device_unplug_cb;
> nc->nmi_monitor_handler = x86_nmi;
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (46 preceding siblings ...)
2016-10-30 21:25 ` [Qemu-devel] [PULL 47/47] acpi: fix assert failure caused by commit 35c5a52d Michael S. Tsirkin
@ 2016-10-31 9:50 ` Igor Mammedov
2016-10-31 22:48 ` Michael S. Tsirkin
2016-11-01 15:22 ` Peter Maydell
2016-11-03 16:59 ` Stefan Hajnoczi
49 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 9:50 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell
On Sun, 30 Oct 2016 23:23:18 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>
> Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>
> for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>
> acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>
> ----------------------------------------------------------------
> virtio, pc: fixes and features
>
> nvdimm hotplug support
Michael,
Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
and keep only nvdimm fixes/cleanups for now?
> virtio migration and ioeventfd rework
> virtio crypto device
> ipmi fixes
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>
> ----------------------------------------------------------------
> Corey Minyard (5):
> ipmi: Remove hotplug from IPMI BMCs
> ipmi_bmc_sim: Remove an unnecessary mutex
> ipmi: Implement shutdown via ACPI overtemp
> ipmi: Add graceful shutdown handling to the external BMC
> acpi/ipmi: Initialize the fwinfo before fetching it
>
> Cédric Le Goater (1):
> ipmi: chassis poweroff should use qemu_system_shutdown_request()
>
> Daniel P. Berrange (1):
> ipmi: fix build config variable name for ipmi_bmc_extern.o
>
> Dr. David Alan Gilbert (2):
> virtio/migration: Add VMStateDescription to VirtioDeviceClass
> virtio/migration: Migrate balloon to VMState
>
> Gonglei (12):
> cryptodev: introduce cryptodev backend interface
> cryptodev: add symmetric algorithm operation stuff
> virtio-crypto: introduce virtio_crypto.h
> cryptodev: introduce a new cryptodev backend
> virtio-crypto: add virtio crypto device emulation
> virtio-crypto-pci: add virtio crypto pci support
> virtio-crypto: set capacity of algorithms supported
> virtio-crypto: add control queue handler
> virtio-crypto: add data queue processing handler
> cryptodev: introduce an unified wrapper for crypto operation
> virtio-crypto: using bh to handle dataq's requests
> virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
>
> Haozhong Zhang (1):
> acpi: fix assert failure caused by commit 35c5a52d
>
> Paolo Bonzini (13):
> virtio: disable ioeventfd as early as possible
> virtio: move ioeventfd_disabled flag to VirtioBusState
> virtio: move ioeventfd_started flag to VirtioBusState
> virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
> virtio: introduce virtio_device_ioeventfd_enabled
> virtio-blk: always use dataplane path if ioeventfd is active
> virtio-scsi: always use dataplane path if ioeventfd is active
> Revert "virtio: Introduce virtio_add_queue_aio"
> virtio: remove set_handler argument from set_host_notifier_internal
> virtio: remove ioeventfd_disabled altogether
> virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
> virtio: inline virtio_queue_set_host_notifier_fd_handler
> virtio: inline set_host_notifier_internal
>
> Xiao Guangrong (12):
> acpi nvdimm: fix wrong buffer size returned by DSM method
> acpi nvdimm: fix OperationRegion definition
> acpi nvdimm: fix device physical address base
> acpi nvdimm: fix ARG3 conflict
> acpi nvdimm: fix Arg6 usage
> nvdimm acpi: compile nvdimm acpi code arch-independently
> acpi nvdimm: rename result_size to dsm_out_buf_siz
> nvdimm acpi: use common macros instead of magic names
> nvdimm acpi: prebuild nvdimm devices for available slots
> nvdimm acpi: introduce fit buffer
> nvdimm acpi: introduce _FIT
> pc: memhp: enable nvdimm device hotplug
>
> hw/block/dataplane/virtio-blk.h | 6 +-
> hw/s390x/virtio-ccw.h | 2 -
> hw/virtio/virtio-pci.h | 17 +-
> include/hw/acpi/acpi_dev_interface.h | 1 +
> include/hw/hotplug.h | 10 +
> include/hw/mem/nvdimm.h | 27 +-
> include/hw/virtio/virtio-bus.h | 27 +-
> include/hw/virtio/virtio-crypto.h | 101 +++
> include/hw/virtio/virtio-scsi.h | 6 +-
> include/hw/virtio/virtio.h | 15 +-
> include/standard-headers/linux/virtio_crypto.h | 429 ++++++++++++
> include/standard-headers/linux/virtio_ids.h | 2 +-
> include/sysemu/cryptodev.h | 298 ++++++++
> backends/cryptodev-builtin.c | 361 ++++++++++
> backends/cryptodev.c | 245 +++++++
> hw/acpi/ipmi.c | 1 +
> hw/acpi/memory_hotplug.c | 31 +-
> hw/acpi/nvdimm.c | 468 ++++++++++---
> hw/block/dataplane/virtio-blk.c | 73 +-
> hw/block/virtio-blk.c | 15 +-
> hw/core/hotplug.c | 11 +
> hw/core/qdev.c | 20 +-
> hw/i386/acpi-build.c | 9 +-
> hw/i386/pc.c | 31 +
> hw/ipmi/ipmi.c | 10 +-
> hw/ipmi/ipmi_bmc_extern.c | 12 +-
> hw/ipmi/ipmi_bmc_sim.c | 7 +-
> hw/mem/nvdimm.c | 4 -
> hw/s390x/virtio-ccw.c | 44 +-
> hw/scsi/virtio-scsi-dataplane.c | 56 +-
> hw/scsi/virtio-scsi.c | 24 +-
> hw/virtio/vhost.c | 5 +-
> hw/virtio/virtio-balloon.c | 31 +-
> hw/virtio/virtio-bus.c | 154 ++---
> hw/virtio/virtio-crypto-pci.c | 77 +++
> hw/virtio/virtio-crypto.c | 898 +++++++++++++++++++++++++
> hw/virtio/virtio-mmio.c | 35 +-
> hw/virtio/virtio-pci.c | 40 +-
> hw/virtio/virtio.c | 153 +++--
> tests/ipmi-bt-test.c | 2 +-
> MAINTAINERS | 13 +
> backends/Makefile.objs | 3 +
> docs/specs/acpi_mem_hotplug.txt | 3 +
> docs/specs/acpi_nvdimm.txt | 58 +-
> hw/acpi/Makefile.objs | 2 +-
> hw/ipmi/Makefile.objs | 2 +-
> hw/virtio/Makefile.objs | 2 +
> qemu-options.hx | 18 +
> 48 files changed, 3352 insertions(+), 507 deletions(-)
> create mode 100644 include/hw/virtio/virtio-crypto.h
> create mode 100644 include/standard-headers/linux/virtio_crypto.h
> create mode 100644 include/sysemu/cryptodev.h
> create mode 100644 backends/cryptodev-builtin.c
> create mode 100644 backends/cryptodev.c
> create mode 100644 hw/virtio/virtio-crypto-pci.c
> create mode 100644 hw/virtio/virtio-crypto.c
>
>
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-10-31 9:45 ` Igor Mammedov
@ 2016-10-31 9:52 ` Xiao Guangrong
2016-10-31 11:09 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-10-31 9:52 UTC (permalink / raw)
To: Igor Mammedov, Michael S. Tsirkin
Cc: qemu-devel, Peter Maydell, Paolo Bonzini, Richard Henderson,
Eduardo Habkost
On 10/31/2016 05:45 PM, Igor Mammedov wrote:
> On Sun, 30 Oct 2016 23:25:05 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>
>> The buffer is used to save the FIT info for all the presented nvdimm
>> devices which is updated after the nvdimm device is plugged or
>> unplugged. In the later patch, it will be used to construct NVDIMM
>> ACPI _FIT method which reflects the presented nvdimm devices after
>> nvdimm hotplug
>>
>> As FIT buffer can not completely mapped into guest address space,
>> OSPM will exit to QEMU multiple times, however, there is the race
>> condition - FIT may be changed during these multiple exits, so that
>> some rules are introduced:
>> 1) the user should hold the @lock to access the buffer and
>> 2) mark @dirty whenever the buffer is updated.
>>
>> @dirty is cleared for the first time OSPM gets fit buffer, if
>> dirty is detected in the later access, OSPM will restart the
>> access
>>
>> As fit should be updated after nvdimm device is successfully realized
>> so that a new hotplug callback, post_hotplug, is introduced
>>
>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>> ---
>> include/hw/hotplug.h | 10 +++++++++
>> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
>> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
>> hw/core/hotplug.c | 11 +++++++++
>> hw/core/qdev.c | 20 +++++++++++++----
>> hw/i386/acpi-build.c | 2 +-
>> hw/i386/pc.c | 19 ++++++++++++++++
>> 7 files changed, 124 insertions(+), 23 deletions(-)
>>
>> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
>> index c0db869..10ca5b6 100644
>> --- a/include/hw/hotplug.h
>> +++ b/include/hw/hotplug.h
>> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
>> * @parent: Opaque parent interface.
>> * @pre_plug: pre plug callback called at start of device.realize(true)
>> * @plug: plug callback called at end of device.realize(true).
>> + * @post_pug: post plug callback called after device is successfully plugged.
> this doesn't seem to be necessary,
> why hotplug_handler_plug() isn't sufficient?
At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
And we can not move "realized = 1" to the front of hotplug_handler_plug() as device
realize routines will check if realized is already been true.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-31 9:23 ` Xiao Guangrong
@ 2016-10-31 10:56 ` Igor Mammedov
2016-10-31 11:09 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 10:56 UTC (permalink / raw)
To: Xiao Guangrong; +Cc: Michael S. Tsirkin, qemu-devel, Peter Maydell
On Mon, 31 Oct 2016 17:23:31 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 10/31/2016 05:20 PM, Igor Mammedov wrote:
> > On Sun, 30 Oct 2016 23:24:46 +0200
> > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >
> >> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>
> >> According to ACPI 6.0 spec, "Memory Device Physical Address
> >> Region Base" in memdev is defined as "This field provides the
> >> Device Physical Address base of the region". This field should
> >> be zero in our case
> > I'm not sure that it should be a zero,
> > care to point source which tells that it should be zero?
>
> The spec says that this is the Device Physical Address, so that
> it is the device internal address, it should be zero as we do not
> reserve any thing in device internal and we do not have no memory
> interleave.
spec says (ACPI 6.1: 5.2.25.3 NVDIMM Region Mapping Structure):
"NVDIMM Physical Address Region Base":
"The base physical address within the NVDIMM of the NVDIMM region."
and nothing more than that so it's hard to come to conclusion that
it's internal address nor it is offset as you treat it here
(structure has 'Region Offset' for that).
However there is "NVDIMM Region Size" field
"In bytes.
The size of the NVDIMM region.
If SPA Range Structure Index and Interleave Ways are
both non-zero, this field shall match System Physical
Address Range Length divided by Interleave Ways."
matches SPA structure, which makes me think that
"NVDIMM Physical Address Region Base"
similarly should match "System Physical Address Range Base" from SPA.
> Actually, this bug was exported when we were enabling nvdimm in
> windows guest.
since it's rather new technology there isn't guaranty that Windows
got it right yet.
If spec is not clear we should ask for clarification and in mean time
look at existing hardware for example.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-31 10:56 ` Igor Mammedov
@ 2016-10-31 11:09 ` Xiao Guangrong
2016-10-31 13:30 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-10-31 11:09 UTC (permalink / raw)
To: Igor Mammedov; +Cc: Michael S. Tsirkin, qemu-devel, Peter Maydell
On 10/31/2016 06:56 PM, Igor Mammedov wrote:
> On Mon, 31 Oct 2016 17:23:31 +0800
> Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
>
>> On 10/31/2016 05:20 PM, Igor Mammedov wrote:
>>> On Sun, 30 Oct 2016 23:24:46 +0200
>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>
>>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>>>
>>>> According to ACPI 6.0 spec, "Memory Device Physical Address
>>>> Region Base" in memdev is defined as "This field provides the
>>>> Device Physical Address base of the region". This field should
>>>> be zero in our case
>>> I'm not sure that it should be a zero,
>>> care to point source which tells that it should be zero?
>>
>> The spec says that this is the Device Physical Address, so that
>> it is the device internal address, it should be zero as we do not
>> reserve any thing in device internal and we do not have no memory
>> interleave.
> spec says (ACPI 6.1: 5.2.25.3 NVDIMM Region Mapping Structure):
> "NVDIMM Physical Address Region Base":
> "The base physical address within the NVDIMM of the NVDIMM region."
>
> and nothing more than that so it's hard to come to conclusion that
> it's internal address nor it is offset as you treat it here
> (structure has 'Region Offset' for that).
I think it is clear as the spec says "_within_ the NVDIMM", it is
even more clear than ACPI 6.0 (5.2.25.2 Memory Device to System Physical
Address Range Mapping Structure), in that, it says:
"In bytes. This field provides the Device Physical Address
base of the region."
In the namespace spec (http://pmem.io/documents/NVDIMM_Namespace_Spec.pdf),
the explanation to DPA in page 9:
"DIMM Physical Address: A memory address from a DIMM’s perspective,
that is, the offset into the DIMM’s memory, starting
with DPA zero
as the lowest addressable byte of the DIMM.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-10-31 9:52 ` Xiao Guangrong
@ 2016-10-31 11:09 ` Igor Mammedov
2016-11-01 3:30 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 11:09 UTC (permalink / raw)
To: Xiao Guangrong
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On Mon, 31 Oct 2016 17:52:24 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 10/31/2016 05:45 PM, Igor Mammedov wrote:
> > On Sun, 30 Oct 2016 23:25:05 +0200
> > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >
> >> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>
> >> The buffer is used to save the FIT info for all the presented nvdimm
> >> devices which is updated after the nvdimm device is plugged or
> >> unplugged. In the later patch, it will be used to construct NVDIMM
> >> ACPI _FIT method which reflects the presented nvdimm devices after
> >> nvdimm hotplug
> >>
> >> As FIT buffer can not completely mapped into guest address space,
> >> OSPM will exit to QEMU multiple times, however, there is the race
> >> condition - FIT may be changed during these multiple exits, so that
> >> some rules are introduced:
> >> 1) the user should hold the @lock to access the buffer and
Could you explain why lock is needed?
> >> 2) mark @dirty whenever the buffer is updated.
> >>
> >> @dirty is cleared for the first time OSPM gets fit buffer, if
> >> dirty is detected in the later access, OSPM will restart the
> >> access
> >>
> >> As fit should be updated after nvdimm device is successfully realized
> >> so that a new hotplug callback, post_hotplug, is introduced
> >>
> >> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> >> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >> ---
> >> include/hw/hotplug.h | 10 +++++++++
> >> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
> >> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
> >> hw/core/hotplug.c | 11 +++++++++
> >> hw/core/qdev.c | 20 +++++++++++++----
> >> hw/i386/acpi-build.c | 2 +-
> >> hw/i386/pc.c | 19 ++++++++++++++++
> >> 7 files changed, 124 insertions(+), 23 deletions(-)
> >>
> >> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
> >> index c0db869..10ca5b6 100644
> >> --- a/include/hw/hotplug.h
> >> +++ b/include/hw/hotplug.h
> >> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
> >> * @parent: Opaque parent interface.
> >> * @pre_plug: pre plug callback called at start of device.realize(true)
> >> * @plug: plug callback called at end of device.realize(true).
> >> + * @post_pug: post plug callback called after device is successfully plugged.
> > this doesn't seem to be necessary,
> > why hotplug_handler_plug() isn't sufficient?
>
> At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
> however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
I suggest instead of adding redundant hook, to reuse hotplug_handler_plug()
which is called after device::realize() successfully completed and amend
nvdimm_plugged_device_list() as follow:
diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
index e486128..c6d8cbb 100644
--- a/hw/acpi/nvdimm.c
+++ b/hw/acpi/nvdimm.c
@@ -38,11 +38,7 @@ static int nvdimm_plugged_device_list(Object *obj, void *opaque)
GSList **list = opaque;
if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
- DeviceState *dev = DEVICE(obj);
-
- if (dev->realized) { /* only realized NVDIMMs matter */
- *list = g_slist_append(*list, DEVICE(obj));
- }
+ *list = g_slist_append(*list, obj);
}
object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
that should count not only already present nvdimms but also the one that's
being hotplugged.
> And we can not move "realized = 1" to the front of hotplug_handler_plug() as device
> realize routines will check if realized is already been true.
^ permalink raw reply related [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base
2016-10-31 11:09 ` Xiao Guangrong
@ 2016-10-31 13:30 ` Igor Mammedov
0 siblings, 0 replies; 87+ messages in thread
From: Igor Mammedov @ 2016-10-31 13:30 UTC (permalink / raw)
To: Xiao Guangrong; +Cc: Michael S. Tsirkin, qemu-devel, Peter Maydell
On Mon, 31 Oct 2016 19:09:07 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 10/31/2016 06:56 PM, Igor Mammedov wrote:
> > On Mon, 31 Oct 2016 17:23:31 +0800
> > Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> >
> >> On 10/31/2016 05:20 PM, Igor Mammedov wrote:
> >>> On Sun, 30 Oct 2016 23:24:46 +0200
> >>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >>>
> >>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>>>
> >>>> According to ACPI 6.0 spec, "Memory Device Physical Address
> >>>> Region Base" in memdev is defined as "This field provides the
> >>>> Device Physical Address base of the region". This field should
> >>>> be zero in our case
> >>> I'm not sure that it should be a zero,
> >>> care to point source which tells that it should be zero?
> >>
> >> The spec says that this is the Device Physical Address, so that
> >> it is the device internal address, it should be zero as we do not
> >> reserve any thing in device internal and we do not have no memory
> >> interleave.
> > spec says (ACPI 6.1: 5.2.25.3 NVDIMM Region Mapping Structure):
> > "NVDIMM Physical Address Region Base":
> > "The base physical address within the NVDIMM of the NVDIMM region."
> >
> > and nothing more than that so it's hard to come to conclusion that
> > it's internal address nor it is offset as you treat it here
> > (structure has 'Region Offset' for that).
>
> I think it is clear as the spec says "_within_ the NVDIMM", it is
> even more clear than ACPI 6.0 (5.2.25.2 Memory Device to System Physical
> Address Range Mapping Structure), in that, it says:
> "In bytes. This field provides the Device Physical Address
> base of the region."
It would be less confusing if spec would say
"starting offset of NVDIMM region within the NVDIMM"
(at least that's how I read it now after looking at 6.0 and NVDIMM_Namespace_Spec.pdf)
as "base physical address" is not defined/described in ACPI spec.
> In the namespace spec (http://pmem.io/documents/NVDIMM_Namespace_Spec.pdf),
> the explanation to DPA in page 9:
> "DIMM Physical Address: A memory address from a DIMM’s perspective,
> that is, the offset into the DIMM’s memory, starting
> with DPA zero
> as the lowest addressable byte of the DIMM.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-31 9:50 ` [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Igor Mammedov
@ 2016-10-31 22:48 ` Michael S. Tsirkin
2016-11-01 12:47 ` Peter Maydell
2016-11-01 13:21 ` Igor Mammedov
0 siblings, 2 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-10-31 22:48 UTC (permalink / raw)
To: Igor Mammedov; +Cc: qemu-devel, Peter Maydell
On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
> On Sun, 30 Oct 2016 23:23:18 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> >
> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> >
> > are available in the git repository at:
> >
> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> >
> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> >
> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> >
> > ----------------------------------------------------------------
> > virtio, pc: fixes and features
> >
> > nvdimm hotplug support
> Michael,
>
> Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
>
> and keep only nvdimm fixes/cleanups for now?
If I drop it now it won't be in the next QEMU and it seems like
a valuable feature. The comments so far are about minor style
improvements that IMO can be addressed by patches on top.
We can always revert if you see bigger issues, but let's enable the
testing of this feature.
>
> > virtio migration and ioeventfd rework
> > virtio crypto device
> > ipmi fixes
> >
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >
> > ----------------------------------------------------------------
> > Corey Minyard (5):
> > ipmi: Remove hotplug from IPMI BMCs
> > ipmi_bmc_sim: Remove an unnecessary mutex
> > ipmi: Implement shutdown via ACPI overtemp
> > ipmi: Add graceful shutdown handling to the external BMC
> > acpi/ipmi: Initialize the fwinfo before fetching it
> >
> > Cédric Le Goater (1):
> > ipmi: chassis poweroff should use qemu_system_shutdown_request()
> >
> > Daniel P. Berrange (1):
> > ipmi: fix build config variable name for ipmi_bmc_extern.o
> >
> > Dr. David Alan Gilbert (2):
> > virtio/migration: Add VMStateDescription to VirtioDeviceClass
> > virtio/migration: Migrate balloon to VMState
> >
> > Gonglei (12):
> > cryptodev: introduce cryptodev backend interface
> > cryptodev: add symmetric algorithm operation stuff
> > virtio-crypto: introduce virtio_crypto.h
> > cryptodev: introduce a new cryptodev backend
> > virtio-crypto: add virtio crypto device emulation
> > virtio-crypto-pci: add virtio crypto pci support
> > virtio-crypto: set capacity of algorithms supported
> > virtio-crypto: add control queue handler
> > virtio-crypto: add data queue processing handler
> > cryptodev: introduce an unified wrapper for crypto operation
> > virtio-crypto: using bh to handle dataq's requests
> > virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
> >
> > Haozhong Zhang (1):
> > acpi: fix assert failure caused by commit 35c5a52d
> >
> > Paolo Bonzini (13):
> > virtio: disable ioeventfd as early as possible
> > virtio: move ioeventfd_disabled flag to VirtioBusState
> > virtio: move ioeventfd_started flag to VirtioBusState
> > virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
> > virtio: introduce virtio_device_ioeventfd_enabled
> > virtio-blk: always use dataplane path if ioeventfd is active
> > virtio-scsi: always use dataplane path if ioeventfd is active
> > Revert "virtio: Introduce virtio_add_queue_aio"
> > virtio: remove set_handler argument from set_host_notifier_internal
> > virtio: remove ioeventfd_disabled altogether
> > virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
> > virtio: inline virtio_queue_set_host_notifier_fd_handler
> > virtio: inline set_host_notifier_internal
> >
> > Xiao Guangrong (12):
> > acpi nvdimm: fix wrong buffer size returned by DSM method
> > acpi nvdimm: fix OperationRegion definition
> > acpi nvdimm: fix device physical address base
> > acpi nvdimm: fix ARG3 conflict
> > acpi nvdimm: fix Arg6 usage
> > nvdimm acpi: compile nvdimm acpi code arch-independently
> > acpi nvdimm: rename result_size to dsm_out_buf_siz
> > nvdimm acpi: use common macros instead of magic names
> > nvdimm acpi: prebuild nvdimm devices for available slots
> > nvdimm acpi: introduce fit buffer
> > nvdimm acpi: introduce _FIT
> > pc: memhp: enable nvdimm device hotplug
> >
> > hw/block/dataplane/virtio-blk.h | 6 +-
> > hw/s390x/virtio-ccw.h | 2 -
> > hw/virtio/virtio-pci.h | 17 +-
> > include/hw/acpi/acpi_dev_interface.h | 1 +
> > include/hw/hotplug.h | 10 +
> > include/hw/mem/nvdimm.h | 27 +-
> > include/hw/virtio/virtio-bus.h | 27 +-
> > include/hw/virtio/virtio-crypto.h | 101 +++
> > include/hw/virtio/virtio-scsi.h | 6 +-
> > include/hw/virtio/virtio.h | 15 +-
> > include/standard-headers/linux/virtio_crypto.h | 429 ++++++++++++
> > include/standard-headers/linux/virtio_ids.h | 2 +-
> > include/sysemu/cryptodev.h | 298 ++++++++
> > backends/cryptodev-builtin.c | 361 ++++++++++
> > backends/cryptodev.c | 245 +++++++
> > hw/acpi/ipmi.c | 1 +
> > hw/acpi/memory_hotplug.c | 31 +-
> > hw/acpi/nvdimm.c | 468 ++++++++++---
> > hw/block/dataplane/virtio-blk.c | 73 +-
> > hw/block/virtio-blk.c | 15 +-
> > hw/core/hotplug.c | 11 +
> > hw/core/qdev.c | 20 +-
> > hw/i386/acpi-build.c | 9 +-
> > hw/i386/pc.c | 31 +
> > hw/ipmi/ipmi.c | 10 +-
> > hw/ipmi/ipmi_bmc_extern.c | 12 +-
> > hw/ipmi/ipmi_bmc_sim.c | 7 +-
> > hw/mem/nvdimm.c | 4 -
> > hw/s390x/virtio-ccw.c | 44 +-
> > hw/scsi/virtio-scsi-dataplane.c | 56 +-
> > hw/scsi/virtio-scsi.c | 24 +-
> > hw/virtio/vhost.c | 5 +-
> > hw/virtio/virtio-balloon.c | 31 +-
> > hw/virtio/virtio-bus.c | 154 ++---
> > hw/virtio/virtio-crypto-pci.c | 77 +++
> > hw/virtio/virtio-crypto.c | 898 +++++++++++++++++++++++++
> > hw/virtio/virtio-mmio.c | 35 +-
> > hw/virtio/virtio-pci.c | 40 +-
> > hw/virtio/virtio.c | 153 +++--
> > tests/ipmi-bt-test.c | 2 +-
> > MAINTAINERS | 13 +
> > backends/Makefile.objs | 3 +
> > docs/specs/acpi_mem_hotplug.txt | 3 +
> > docs/specs/acpi_nvdimm.txt | 58 +-
> > hw/acpi/Makefile.objs | 2 +-
> > hw/ipmi/Makefile.objs | 2 +-
> > hw/virtio/Makefile.objs | 2 +
> > qemu-options.hx | 18 +
> > 48 files changed, 3352 insertions(+), 507 deletions(-)
> > create mode 100644 include/hw/virtio/virtio-crypto.h
> > create mode 100644 include/standard-headers/linux/virtio_crypto.h
> > create mode 100644 include/sysemu/cryptodev.h
> > create mode 100644 backends/cryptodev-builtin.c
> > create mode 100644 backends/cryptodev.c
> > create mode 100644 hw/virtio/virtio-crypto-pci.c
> > create mode 100644 hw/virtio/virtio-crypto.c
> >
> >
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-10-31 11:09 ` Igor Mammedov
@ 2016-11-01 3:30 ` Xiao Guangrong
2016-11-01 10:35 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-01 3:30 UTC (permalink / raw)
To: Igor Mammedov
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On 10/31/2016 07:09 PM, Igor Mammedov wrote:
> On Mon, 31 Oct 2016 17:52:24 +0800
> Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
>
>> On 10/31/2016 05:45 PM, Igor Mammedov wrote:
>>> On Sun, 30 Oct 2016 23:25:05 +0200
>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>
>>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>>>
>>>> The buffer is used to save the FIT info for all the presented nvdimm
>>>> devices which is updated after the nvdimm device is plugged or
>>>> unplugged. In the later patch, it will be used to construct NVDIMM
>>>> ACPI _FIT method which reflects the presented nvdimm devices after
>>>> nvdimm hotplug
>>>>
>>>> As FIT buffer can not completely mapped into guest address space,
>>>> OSPM will exit to QEMU multiple times, however, there is the race
>>>> condition - FIT may be changed during these multiple exits, so that
>>>> some rules are introduced:
>>>> 1) the user should hold the @lock to access the buffer and
> Could you explain why lock is needed?
Yes. These are two things need to be synced between QEMU and OSPM:
- fit buffer. QEMU products it and VM will consume it at the same time.
- dirty indicator. QEMU sets it and it will be cleared by OSPM, that means.
both sides will modify it.
>
>>>> 2) mark @dirty whenever the buffer is updated.
>>>>
>>>> @dirty is cleared for the first time OSPM gets fit buffer, if
>>>> dirty is detected in the later access, OSPM will restart the
>>>> access
>>>>
>>>> As fit should be updated after nvdimm device is successfully realized
>>>> so that a new hotplug callback, post_hotplug, is introduced
>>>>
>>>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>>> ---
>>>> include/hw/hotplug.h | 10 +++++++++
>>>> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
>>>> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
>>>> hw/core/hotplug.c | 11 +++++++++
>>>> hw/core/qdev.c | 20 +++++++++++++----
>>>> hw/i386/acpi-build.c | 2 +-
>>>> hw/i386/pc.c | 19 ++++++++++++++++
>>>> 7 files changed, 124 insertions(+), 23 deletions(-)
>>>>
>>>> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
>>>> index c0db869..10ca5b6 100644
>>>> --- a/include/hw/hotplug.h
>>>> +++ b/include/hw/hotplug.h
>>>> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
>>>> * @parent: Opaque parent interface.
>>>> * @pre_plug: pre plug callback called at start of device.realize(true)
>>>> * @plug: plug callback called at end of device.realize(true).
>>>> + * @post_pug: post plug callback called after device is successfully plugged.
>>> this doesn't seem to be necessary,
>>> why hotplug_handler_plug() isn't sufficient?
>>
>> At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
>> however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
>
> I suggest instead of adding redundant hook, to reuse hotplug_handler_plug()
> which is called after device::realize() successfully completed and amend
> nvdimm_plugged_device_list() as follow:
>
> diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
> index e486128..c6d8cbb 100644
> --- a/hw/acpi/nvdimm.c
> +++ b/hw/acpi/nvdimm.c
> @@ -38,11 +38,7 @@ static int nvdimm_plugged_device_list(Object *obj, void *opaque)
> GSList **list = opaque;
>
> if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
> - DeviceState *dev = DEVICE(obj);
> -
> - if (dev->realized) { /* only realized NVDIMMs matter */
> - *list = g_slist_append(*list, DEVICE(obj));
> - }
> + *list = g_slist_append(*list, obj);
> }
>
> object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
>
> that should count not only already present nvdimms but also the one that's
> being hotplugged.
It is not good as the device which failed to initialize or is not being plugged
(consider two nvdimm devices directly appended in QEMU command line, when we plug
the first one, both nvdimm devices will be counted in this list) is also counted in
this list...
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-11-01 3:30 ` Xiao Guangrong
@ 2016-11-01 10:35 ` Igor Mammedov
2016-11-01 13:40 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-11-01 10:35 UTC (permalink / raw)
To: Xiao Guangrong
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On Tue, 1 Nov 2016 11:30:46 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 10/31/2016 07:09 PM, Igor Mammedov wrote:
> > On Mon, 31 Oct 2016 17:52:24 +0800
> > Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> >
> >> On 10/31/2016 05:45 PM, Igor Mammedov wrote:
> >>> On Sun, 30 Oct 2016 23:25:05 +0200
> >>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >>>
> >>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>>>
> >>>> The buffer is used to save the FIT info for all the presented nvdimm
> >>>> devices which is updated after the nvdimm device is plugged or
> >>>> unplugged. In the later patch, it will be used to construct NVDIMM
> >>>> ACPI _FIT method which reflects the presented nvdimm devices after
> >>>> nvdimm hotplug
> >>>>
> >>>> As FIT buffer can not completely mapped into guest address space,
> >>>> OSPM will exit to QEMU multiple times, however, there is the race
> >>>> condition - FIT may be changed during these multiple exits, so that
> >>>> some rules are introduced:
> >>>> 1) the user should hold the @lock to access the buffer and
> > Could you explain why lock is needed?
>
> Yes. These are two things need to be synced between QEMU and OSPM:
> - fit buffer. QEMU products it and VM will consume it at the same time.
> - dirty indicator. QEMU sets it and it will be cleared by OSPM, that means.
> both sides will modify it.
I understand why 'dirty indicator' is necessary but
could you describe a concrete call flows in detail
that would cause concurrent access and need extra
lock protection.
Note that there is global lock (dubbed BQL) in QEMU,
which is taken every time guest accesses IO port or
QMP/monitor control event happens.
> >>>> 2) mark @dirty whenever the buffer is updated.
> >>>>
> >>>> @dirty is cleared for the first time OSPM gets fit buffer, if
> >>>> dirty is detected in the later access, OSPM will restart the
> >>>> access
> >>>>
> >>>> As fit should be updated after nvdimm device is successfully realized
> >>>> so that a new hotplug callback, post_hotplug, is introduced
> >>>>
> >>>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> >>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >>>> ---
> >>>> include/hw/hotplug.h | 10 +++++++++
> >>>> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
> >>>> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
> >>>> hw/core/hotplug.c | 11 +++++++++
> >>>> hw/core/qdev.c | 20 +++++++++++++----
> >>>> hw/i386/acpi-build.c | 2 +-
> >>>> hw/i386/pc.c | 19 ++++++++++++++++
> >>>> 7 files changed, 124 insertions(+), 23 deletions(-)
> >>>>
> >>>> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
> >>>> index c0db869..10ca5b6 100644
> >>>> --- a/include/hw/hotplug.h
> >>>> +++ b/include/hw/hotplug.h
> >>>> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
> >>>> * @parent: Opaque parent interface.
> >>>> * @pre_plug: pre plug callback called at start of device.realize(true)
> >>>> * @plug: plug callback called at end of device.realize(true).
> >>>> + * @post_pug: post plug callback called after device is successfully plugged.
> >>> this doesn't seem to be necessary,
> >>> why hotplug_handler_plug() isn't sufficient?
> >>
> >> At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
> >> however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
> >
> > I suggest instead of adding redundant hook, to reuse hotplug_handler_plug()
> > which is called after device::realize() successfully completed and amend
> > nvdimm_plugged_device_list() as follow:
> >
> > diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
> > index e486128..c6d8cbb 100644
> > --- a/hw/acpi/nvdimm.c
> > +++ b/hw/acpi/nvdimm.c
> > @@ -38,11 +38,7 @@ static int nvdimm_plugged_device_list(Object *obj, void *opaque)
> > GSList **list = opaque;
> >
> > if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
> > - DeviceState *dev = DEVICE(obj);
> > -
> > - if (dev->realized) { /* only realized NVDIMMs matter */
> > - *list = g_slist_append(*list, DEVICE(obj));
> > - }
> > + *list = g_slist_append(*list, obj);
> > }
> >
> > object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
> >
> > that should count not only already present nvdimms but also the one that's
> > being hotplugged.
>
> It is not good as the device which failed to initialize
See device_set_realized()
[...]
if (dc->realize) {
dc->realize(dev, &local_err);
}
if (local_err != NULL) {
goto fail;
}
DEVICE_LISTENER_CALL(realize, Forward, dev);
if (hotplug_ctrl) {
hotplug_handler_plug(hotplug_ctrl, dev, &local_err);
}
i.e. control reaches to hotplug_handler_plug() only if
call dc->realize(dev, &local_err) is successful.
> or is not being plugged
> (consider two nvdimm devices directly appended in QEMU command line, when we plug
> the first one, both nvdimm devices will be counted in this list) is also counted in
> this list...
nope,
qdev_device_add() is called sequentially (one time for each -device/device_add)
so nvdimm_plugged_device_list() sees only devices that's been added
by previous -device/device_add plus one extra device that's
being added by in progress qdev_device_add() call.
so for "-device nvdimm,id=nv1 -device nvdimm,id=2" call sequence
would look like:
1:
qdev_device_add("nvdimm,id=nv1") {
-> set parent // device becomes visible to nvdimm_get_plugged_device_list()
// this is the only time where device->realized == false
// could be observed, all other call chains would either
// see device->realized == true or not see device at all
-> realize()
-> device_set_realized()
-> nvdimm->realize()
-> hotplug_handler_plug()
nvdimm_get_plugged_device_list()
// would return list with 1 item (nv1)
// where nv1->realized == false)
}
2:
qdev_device_add("nvdimm,id=nv2") {
-> set parent // device becomes visible to nvdimm_get_plugged_device_list()
-> realize()
-> device_set_realized()
-> nvdimm->realize()
-> hotplug_handler_plug()
nvdimm_get_plugged_device_list()
// would return list with 2 items (nv1 and nv2)
// where nv1->realized == true and nv2->realized = false
}
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-31 22:48 ` Michael S. Tsirkin
@ 2016-11-01 12:47 ` Peter Maydell
2016-11-01 13:21 ` Igor Mammedov
1 sibling, 0 replies; 87+ messages in thread
From: Peter Maydell @ 2016-11-01 12:47 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: Igor Mammedov, QEMU Developers
On 31 October 2016 at 22:48, Michael S. Tsirkin <mst@redhat.com> wrote:
> On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
>> On Sun, 30 Oct 2016 23:23:18 +0200
>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>
>> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>> >
>> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>> >
>> > are available in the git repository at:
>> >
>> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>> >
>> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>> >
>> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>> >
>> > ----------------------------------------------------------------
>> > virtio, pc: fixes and features
>> >
>> > nvdimm hotplug support
>> Michael,
>>
>> Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
>>
>> and keep only nvdimm fixes/cleanups for now?
>
> If I drop it now it won't be in the next QEMU and it seems like
> a valuable feature. The comments so far are about minor style
> improvements that IMO can be addressed by patches on top.
>
> We can always revert if you see bigger issues, but let's enable the
> testing of this feature.
I'm waiting for a reply from Igor before I do anything with
this pull request.
thanks
-- PMM
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-31 22:48 ` Michael S. Tsirkin
2016-11-01 12:47 ` Peter Maydell
@ 2016-11-01 13:21 ` Igor Mammedov
2016-11-01 13:45 ` Xiao Guangrong
2016-11-01 13:55 ` Michael S. Tsirkin
1 sibling, 2 replies; 87+ messages in thread
From: Igor Mammedov @ 2016-11-01 13:21 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell
On Tue, 1 Nov 2016 00:48:11 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
> > On Sun, 30 Oct 2016 23:23:18 +0200
> > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >
> > > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> > >
> > > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> > >
> > > are available in the git repository at:
> > >
> > > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> > >
> > > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> > >
> > > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> > >
> > > ----------------------------------------------------------------
> > > virtio, pc: fixes and features
> > >
> > > nvdimm hotplug support
> > Michael,
> >
> > Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
> >
> > and keep only nvdimm fixes/cleanups for now?
>
> If I drop it now it won't be in the next QEMU and it seems like
> a valuable feature. The comments so far are about minor style
> improvements that IMO can be addressed by patches on top.
wrt nvdimm hotplug support it's not about style issues but rather
design issues: for example:
- it extends general hotplug framework unnecessarily instead of
figuring out how it works.
- adds not needed locks
maybe there is more and all of that was posted just a day before
this pull request so I haven't even had a chance to review it properly.
> We can always revert if you see bigger issues, but let's enable the
> testing of this feature.
if it didn't mess with general infrastructure, I wouldn't care much.
But it does so I'd rather avoid merging not yet ready series just for
the sake of getting it in.
I haven't reviewed 28-35 patches either but they are all cleanups/
fixes of current nvdimm code and local to it so don't mind them
getting merged.
However I suggest dropping 36-39 patches from this pull request
as not yet ready for merging.
>
>
> > > virtio migration and ioeventfd rework
> > > virtio crypto device
> > > ipmi fixes
> > >
> > > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > >
> > > ----------------------------------------------------------------
> > > Corey Minyard (5):
> > > ipmi: Remove hotplug from IPMI BMCs
> > > ipmi_bmc_sim: Remove an unnecessary mutex
> > > ipmi: Implement shutdown via ACPI overtemp
> > > ipmi: Add graceful shutdown handling to the external BMC
> > > acpi/ipmi: Initialize the fwinfo before fetching it
> > >
> > > Cédric Le Goater (1):
> > > ipmi: chassis poweroff should use qemu_system_shutdown_request()
> > >
> > > Daniel P. Berrange (1):
> > > ipmi: fix build config variable name for ipmi_bmc_extern.o
> > >
> > > Dr. David Alan Gilbert (2):
> > > virtio/migration: Add VMStateDescription to VirtioDeviceClass
> > > virtio/migration: Migrate balloon to VMState
> > >
> > > Gonglei (12):
> > > cryptodev: introduce cryptodev backend interface
> > > cryptodev: add symmetric algorithm operation stuff
> > > virtio-crypto: introduce virtio_crypto.h
> > > cryptodev: introduce a new cryptodev backend
> > > virtio-crypto: add virtio crypto device emulation
> > > virtio-crypto-pci: add virtio crypto pci support
> > > virtio-crypto: set capacity of algorithms supported
> > > virtio-crypto: add control queue handler
> > > virtio-crypto: add data queue processing handler
> > > cryptodev: introduce an unified wrapper for crypto operation
> > > virtio-crypto: using bh to handle dataq's requests
> > > virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
> > >
> > > Haozhong Zhang (1):
> > > acpi: fix assert failure caused by commit 35c5a52d
> > >
> > > Paolo Bonzini (13):
> > > virtio: disable ioeventfd as early as possible
> > > virtio: move ioeventfd_disabled flag to VirtioBusState
> > > virtio: move ioeventfd_started flag to VirtioBusState
> > > virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
> > > virtio: introduce virtio_device_ioeventfd_enabled
> > > virtio-blk: always use dataplane path if ioeventfd is active
> > > virtio-scsi: always use dataplane path if ioeventfd is active
> > > Revert "virtio: Introduce virtio_add_queue_aio"
> > > virtio: remove set_handler argument from set_host_notifier_internal
> > > virtio: remove ioeventfd_disabled altogether
> > > virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
> > > virtio: inline virtio_queue_set_host_notifier_fd_handler
> > > virtio: inline set_host_notifier_internal
> > >
> > > Xiao Guangrong (12):
> > > acpi nvdimm: fix wrong buffer size returned by DSM method
> > > acpi nvdimm: fix OperationRegion definition
> > > acpi nvdimm: fix device physical address base
> > > acpi nvdimm: fix ARG3 conflict
> > > acpi nvdimm: fix Arg6 usage
> > > nvdimm acpi: compile nvdimm acpi code arch-independently
> > > acpi nvdimm: rename result_size to dsm_out_buf_siz
> > > nvdimm acpi: use common macros instead of magic names
> > > nvdimm acpi: prebuild nvdimm devices for available slots
> > > nvdimm acpi: introduce fit buffer
> > > nvdimm acpi: introduce _FIT
> > > pc: memhp: enable nvdimm device hotplug
> > >
> > > hw/block/dataplane/virtio-blk.h | 6 +-
> > > hw/s390x/virtio-ccw.h | 2 -
> > > hw/virtio/virtio-pci.h | 17 +-
> > > include/hw/acpi/acpi_dev_interface.h | 1 +
> > > include/hw/hotplug.h | 10 +
> > > include/hw/mem/nvdimm.h | 27 +-
> > > include/hw/virtio/virtio-bus.h | 27 +-
> > > include/hw/virtio/virtio-crypto.h | 101 +++
> > > include/hw/virtio/virtio-scsi.h | 6 +-
> > > include/hw/virtio/virtio.h | 15 +-
> > > include/standard-headers/linux/virtio_crypto.h | 429 ++++++++++++
> > > include/standard-headers/linux/virtio_ids.h | 2 +-
> > > include/sysemu/cryptodev.h | 298 ++++++++
> > > backends/cryptodev-builtin.c | 361 ++++++++++
> > > backends/cryptodev.c | 245 +++++++
> > > hw/acpi/ipmi.c | 1 +
> > > hw/acpi/memory_hotplug.c | 31 +-
> > > hw/acpi/nvdimm.c | 468 ++++++++++---
> > > hw/block/dataplane/virtio-blk.c | 73 +-
> > > hw/block/virtio-blk.c | 15 +-
> > > hw/core/hotplug.c | 11 +
> > > hw/core/qdev.c | 20 +-
> > > hw/i386/acpi-build.c | 9 +-
> > > hw/i386/pc.c | 31 +
> > > hw/ipmi/ipmi.c | 10 +-
> > > hw/ipmi/ipmi_bmc_extern.c | 12 +-
> > > hw/ipmi/ipmi_bmc_sim.c | 7 +-
> > > hw/mem/nvdimm.c | 4 -
> > > hw/s390x/virtio-ccw.c | 44 +-
> > > hw/scsi/virtio-scsi-dataplane.c | 56 +-
> > > hw/scsi/virtio-scsi.c | 24 +-
> > > hw/virtio/vhost.c | 5 +-
> > > hw/virtio/virtio-balloon.c | 31 +-
> > > hw/virtio/virtio-bus.c | 154 ++---
> > > hw/virtio/virtio-crypto-pci.c | 77 +++
> > > hw/virtio/virtio-crypto.c | 898 +++++++++++++++++++++++++
> > > hw/virtio/virtio-mmio.c | 35 +-
> > > hw/virtio/virtio-pci.c | 40 +-
> > > hw/virtio/virtio.c | 153 +++--
> > > tests/ipmi-bt-test.c | 2 +-
> > > MAINTAINERS | 13 +
> > > backends/Makefile.objs | 3 +
> > > docs/specs/acpi_mem_hotplug.txt | 3 +
> > > docs/specs/acpi_nvdimm.txt | 58 +-
> > > hw/acpi/Makefile.objs | 2 +-
> > > hw/ipmi/Makefile.objs | 2 +-
> > > hw/virtio/Makefile.objs | 2 +
> > > qemu-options.hx | 18 +
> > > 48 files changed, 3352 insertions(+), 507 deletions(-)
> > > create mode 100644 include/hw/virtio/virtio-crypto.h
> > > create mode 100644 include/standard-headers/linux/virtio_crypto.h
> > > create mode 100644 include/sysemu/cryptodev.h
> > > create mode 100644 backends/cryptodev-builtin.c
> > > create mode 100644 backends/cryptodev.c
> > > create mode 100644 hw/virtio/virtio-crypto-pci.c
> > > create mode 100644 hw/virtio/virtio-crypto.c
> > >
> > >
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-11-01 10:35 ` Igor Mammedov
@ 2016-11-01 13:40 ` Xiao Guangrong
2016-11-01 13:58 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-01 13:40 UTC (permalink / raw)
To: Igor Mammedov
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On 11/01/2016 06:35 PM, Igor Mammedov wrote:
> On Tue, 1 Nov 2016 11:30:46 +0800
> Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
>
>> On 10/31/2016 07:09 PM, Igor Mammedov wrote:
>>> On Mon, 31 Oct 2016 17:52:24 +0800
>>> Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
>>>
>>>> On 10/31/2016 05:45 PM, Igor Mammedov wrote:
>>>>> On Sun, 30 Oct 2016 23:25:05 +0200
>>>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>>>
>>>>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>>>>>
>>>>>> The buffer is used to save the FIT info for all the presented nvdimm
>>>>>> devices which is updated after the nvdimm device is plugged or
>>>>>> unplugged. In the later patch, it will be used to construct NVDIMM
>>>>>> ACPI _FIT method which reflects the presented nvdimm devices after
>>>>>> nvdimm hotplug
>>>>>>
>>>>>> As FIT buffer can not completely mapped into guest address space,
>>>>>> OSPM will exit to QEMU multiple times, however, there is the race
>>>>>> condition - FIT may be changed during these multiple exits, so that
>>>>>> some rules are introduced:
>>>>>> 1) the user should hold the @lock to access the buffer and
>>> Could you explain why lock is needed?
>>
>> Yes. These are two things need to be synced between QEMU and OSPM:
>> - fit buffer. QEMU products it and VM will consume it at the same time.
>> - dirty indicator. QEMU sets it and it will be cleared by OSPM, that means.
>> both sides will modify it.
>
> I understand why 'dirty indicator' is necessary but
> could you describe a concrete call flows in detail
> that would cause concurrent access and need extra
> lock protection.
>
> Note that there is global lock (dubbed BQL) in QEMU,
> which is taken every time guest accesses IO port or
> QMP/monitor control event happens.
Ah. okay. If there is a lock to protect vm-exit handler and
monitor handler, i think it is okay to drop the lock.
>
>>>>>> 2) mark @dirty whenever the buffer is updated.
>>>>>>
>>>>>> @dirty is cleared for the first time OSPM gets fit buffer, if
>>>>>> dirty is detected in the later access, OSPM will restart the
>>>>>> access
>>>>>>
>>>>>> As fit should be updated after nvdimm device is successfully realized
>>>>>> so that a new hotplug callback, post_hotplug, is introduced
>>>>>>
>>>>>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>>>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>>>>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>>>>> ---
>>>>>> include/hw/hotplug.h | 10 +++++++++
>>>>>> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
>>>>>> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
>>>>>> hw/core/hotplug.c | 11 +++++++++
>>>>>> hw/core/qdev.c | 20 +++++++++++++----
>>>>>> hw/i386/acpi-build.c | 2 +-
>>>>>> hw/i386/pc.c | 19 ++++++++++++++++
>>>>>> 7 files changed, 124 insertions(+), 23 deletions(-)
>>>>>>
>>>>>> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
>>>>>> index c0db869..10ca5b6 100644
>>>>>> --- a/include/hw/hotplug.h
>>>>>> +++ b/include/hw/hotplug.h
>>>>>> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
>>>>>> * @parent: Opaque parent interface.
>>>>>> * @pre_plug: pre plug callback called at start of device.realize(true)
>>>>>> * @plug: plug callback called at end of device.realize(true).
>>>>>> + * @post_pug: post plug callback called after device is successfully plugged.
>>>>> this doesn't seem to be necessary,
>>>>> why hotplug_handler_plug() isn't sufficient?
>>>>
>>>> At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
>>>> however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
>>>
>>> I suggest instead of adding redundant hook, to reuse hotplug_handler_plug()
>>> which is called after device::realize() successfully completed and amend
>>> nvdimm_plugged_device_list() as follow:
>>>
>>> diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
>>> index e486128..c6d8cbb 100644
>>> --- a/hw/acpi/nvdimm.c
>>> +++ b/hw/acpi/nvdimm.c
>>> @@ -38,11 +38,7 @@ static int nvdimm_plugged_device_list(Object *obj, void *opaque)
>>> GSList **list = opaque;
>>>
>>> if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
>>> - DeviceState *dev = DEVICE(obj);
>>> -
>>> - if (dev->realized) { /* only realized NVDIMMs matter */
>>> - *list = g_slist_append(*list, DEVICE(obj));
>>> - }
>>> + *list = g_slist_append(*list, obj);
>>> }
>>>
>>> object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
>>>
>>> that should count not only already present nvdimms but also the one that's
>>> being hotplugged.
>>
>> It is not good as the device which failed to initialize
> See device_set_realized()
> [...]
> if (dc->realize) {
> dc->realize(dev, &local_err);
> }
>
> if (local_err != NULL) {
> goto fail;
> }
>
> DEVICE_LISTENER_CALL(realize, Forward, dev);
>
> if (hotplug_ctrl) {
> hotplug_handler_plug(hotplug_ctrl, dev, &local_err);
> }
>
> i.e. control reaches to hotplug_handler_plug() only if
> call dc->realize(dev, &local_err) is successful.
>
I mean, for example. if there are two devices, the first one is failed to be
realized, the second one is init-ed successfully, then can
nvdimm_plugged_device_list() get these two devices?
Based the code of pc_dimm_built_list(), i guess yes.
>> or is not being plugged
>> (consider two nvdimm devices directly appended in QEMU command line, when we plug
>> the first one, both nvdimm devices will be counted in this list) is also counted in
>> this list...
> nope,
> qdev_device_add() is called sequentially (one time for each -device/device_add)
> so nvdimm_plugged_device_list() sees only devices that's been added
> by previous -device/device_add plus one extra device that's
> being added by in progress qdev_device_add() call.
>
> so for "-device nvdimm,id=nv1 -device nvdimm,id=2" call sequence
> would look like:
> 1:
> qdev_device_add("nvdimm,id=nv1") {
> -> set parent // device becomes visible to nvdimm_get_plugged_device_list()
> // this is the only time where device->realized == false
> // could be observed, all other call chains would either
> // see device->realized == true or not see device at all
> -> realize()
> -> device_set_realized()
> -> nvdimm->realize()
> -> hotplug_handler_plug()
> nvdimm_get_plugged_device_list()
> // would return list with 1 item (nv1)
> // where nv1->realized == false)
>
> }
> 2:
> qdev_device_add("nvdimm,id=nv2") {
> -> set parent // device becomes visible to nvdimm_get_plugged_device_list()
> -> realize()
> -> device_set_realized()
> -> nvdimm->realize()
> -> hotplug_handler_plug()
> nvdimm_get_plugged_device_list()
> // would return list with 2 items (nv1 and nv2)
> // where nv1->realized == true and nv2->realized = false
> }
>
Okay.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 13:21 ` Igor Mammedov
@ 2016-11-01 13:45 ` Xiao Guangrong
2016-11-01 13:55 ` Michael S. Tsirkin
1 sibling, 0 replies; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-01 13:45 UTC (permalink / raw)
To: Igor Mammedov, Michael S. Tsirkin; +Cc: Peter Maydell, qemu-devel
On 11/01/2016 09:21 PM, Igor Mammedov wrote:
> On Tue, 1 Nov 2016 00:48:11 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
>> On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
>>> On Sun, 30 Oct 2016 23:23:18 +0200
>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>
>>>> The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>>>>
>>>> Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>>>>
>>>> are available in the git repository at:
>>>>
>>>> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>>>>
>>>> for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>>>>
>>>> acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>>>>
>>>> ----------------------------------------------------------------
>>>> virtio, pc: fixes and features
>>>>
>>>> nvdimm hotplug support
>>> Michael,
>>>
>>> Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
>>>
>>> and keep only nvdimm fixes/cleanups for now?
>>
>> If I drop it now it won't be in the next QEMU and it seems like
>> a valuable feature. The comments so far are about minor style
>> improvements that IMO can be addressed by patches on top.
> wrt nvdimm hotplug support it's not about style issues but rather
> design issues: for example:
> - it extends general hotplug framework unnecessarily instead of
> figuring out how it works.
> - adds not needed locks
> maybe there is more and all of that was posted just a day before
> this pull request so I haven't even had a chance to review it properly.
>
>
>
>> We can always revert if you see bigger issues, but let's enable the
>> testing of this feature.
> if it didn't mess with general infrastructure, I wouldn't care much.
> But it does so I'd rather avoid merging not yet ready series just for
> the sake of getting it in.
>
> I haven't reviewed 28-35 patches either but they are all cleanups/
> fixes of current nvdimm code and local to it so don't mind them
> getting merged.
>
> However I suggest dropping 36-39 patches from this pull request
> as not yet ready for merging.
Igor,
Thank for your hard work to review this patchset. Only two patches
related to general infrastructure that are:
[PULL 37/47] nvdimm acpi: introduce fit buffer
[PULL 39/47] pc: memhp: enable nvdimm device hotplug
The issue you pointed out in [PULL 37/47] can be easily improved
on the top of this patchset. Could you please look at the 39th,
if there is no big issue, could it be merged first?
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 13:21 ` Igor Mammedov
2016-11-01 13:45 ` Xiao Guangrong
@ 2016-11-01 13:55 ` Michael S. Tsirkin
2016-11-01 14:14 ` Igor Mammedov
1 sibling, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-11-01 13:55 UTC (permalink / raw)
To: Igor Mammedov; +Cc: qemu-devel, Peter Maydell
On Tue, Nov 01, 2016 at 02:21:07PM +0100, Igor Mammedov wrote:
> On Tue, 1 Nov 2016 00:48:11 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
> > On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
> > > On Sun, 30 Oct 2016 23:23:18 +0200
> > > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> > >
> > > > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> > > >
> > > > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> > > >
> > > > are available in the git repository at:
> > > >
> > > > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> > > >
> > > > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> > > >
> > > > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> > > >
> > > > ----------------------------------------------------------------
> > > > virtio, pc: fixes and features
> > > >
> > > > nvdimm hotplug support
> > > Michael,
> > >
> > > Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
> > >
> > > and keep only nvdimm fixes/cleanups for now?
> >
> > If I drop it now it won't be in the next QEMU and it seems like
> > a valuable feature. The comments so far are about minor style
> > improvements that IMO can be addressed by patches on top.
> wrt nvdimm hotplug support it's not about style issues but rather
> design issues: for example:
> - it extends general hotplug framework unnecessarily instead of
> figuring out how it works.
> - adds not needed locks
> maybe there is more and all of that was posted just a day before
> this pull request so I haven't even had a chance to review it properly.
>
>
>
> > We can always revert if you see bigger issues, but let's enable the
> > testing of this feature.
> if it didn't mess with general infrastructure, I wouldn't care much.
> But it does so I'd rather avoid merging not yet ready series just for
> the sake of getting it in.
>
> I haven't reviewed 28-35 patches either but they are all cleanups/
> fixes of current nvdimm code and local to it so don't mind them
> getting merged.
>
> However I suggest dropping 36-39 patches from this pull request
> as not yet ready for merging.
So I think it's ok to keep them from now as that should help
the feature converge faster, on the understanding the
style issues are addressed quickly.
Let's merge as is and I'll revert if this does not happen
within two weeks. Ack?
--
MST
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-11-01 13:40 ` Xiao Guangrong
@ 2016-11-01 13:58 ` Igor Mammedov
2016-11-01 15:57 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-11-01 13:58 UTC (permalink / raw)
To: Xiao Guangrong
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On Tue, 1 Nov 2016 21:40:46 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 11/01/2016 06:35 PM, Igor Mammedov wrote:
> > On Tue, 1 Nov 2016 11:30:46 +0800
> > Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> >
> >> On 10/31/2016 07:09 PM, Igor Mammedov wrote:
> >>> On Mon, 31 Oct 2016 17:52:24 +0800
> >>> Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> >>>
> >>>> On 10/31/2016 05:45 PM, Igor Mammedov wrote:
> >>>>> On Sun, 30 Oct 2016 23:25:05 +0200
> >>>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >>>>>
> >>>>>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>>>>>
> >>>>>> The buffer is used to save the FIT info for all the presented nvdimm
> >>>>>> devices which is updated after the nvdimm device is plugged or
> >>>>>> unplugged. In the later patch, it will be used to construct NVDIMM
> >>>>>> ACPI _FIT method which reflects the presented nvdimm devices after
> >>>>>> nvdimm hotplug
> >>>>>>
> >>>>>> As FIT buffer can not completely mapped into guest address space,
> >>>>>> OSPM will exit to QEMU multiple times, however, there is the race
> >>>>>> condition - FIT may be changed during these multiple exits, so that
> >>>>>> some rules are introduced:
> >>>>>> 1) the user should hold the @lock to access the buffer and
> >>> Could you explain why lock is needed?
> >>
> >> Yes. These are two things need to be synced between QEMU and OSPM:
> >> - fit buffer. QEMU products it and VM will consume it at the same time.
> >> - dirty indicator. QEMU sets it and it will be cleared by OSPM, that means.
> >> both sides will modify it.
> >
> > I understand why 'dirty indicator' is necessary but
> > could you describe a concrete call flows in detail
> > that would cause concurrent access and need extra
> > lock protection.
> >
> > Note that there is global lock (dubbed BQL) in QEMU,
> > which is taken every time guest accesses IO port or
> > QMP/monitor control event happens.
>
> Ah. okay. If there is a lock to protect vm-exit handler and
> monitor handler, i think it is okay to drop the lock.
>
> >
> >>>>>> 2) mark @dirty whenever the buffer is updated.
> >>>>>>
> >>>>>> @dirty is cleared for the first time OSPM gets fit buffer, if
> >>>>>> dirty is detected in the later access, OSPM will restart the
> >>>>>> access
> >>>>>>
> >>>>>> As fit should be updated after nvdimm device is successfully realized
> >>>>>> so that a new hotplug callback, post_hotplug, is introduced
> >>>>>>
> >>>>>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>>>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> >>>>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >>>>>> ---
> >>>>>> include/hw/hotplug.h | 10 +++++++++
> >>>>>> include/hw/mem/nvdimm.h | 26 +++++++++++++++++++++-
> >>>>>> hw/acpi/nvdimm.c | 59 +++++++++++++++++++++++++++++++++++--------------
> >>>>>> hw/core/hotplug.c | 11 +++++++++
> >>>>>> hw/core/qdev.c | 20 +++++++++++++----
> >>>>>> hw/i386/acpi-build.c | 2 +-
> >>>>>> hw/i386/pc.c | 19 ++++++++++++++++
> >>>>>> 7 files changed, 124 insertions(+), 23 deletions(-)
> >>>>>>
> >>>>>> diff --git a/include/hw/hotplug.h b/include/hw/hotplug.h
> >>>>>> index c0db869..10ca5b6 100644
> >>>>>> --- a/include/hw/hotplug.h
> >>>>>> +++ b/include/hw/hotplug.h
> >>>>>> @@ -47,6 +47,7 @@ typedef void (*hotplug_fn)(HotplugHandler *plug_handler,
> >>>>>> * @parent: Opaque parent interface.
> >>>>>> * @pre_plug: pre plug callback called at start of device.realize(true)
> >>>>>> * @plug: plug callback called at end of device.realize(true).
> >>>>>> + * @post_pug: post plug callback called after device is successfully plugged.
> >>>>> this doesn't seem to be necessary,
> >>>>> why hotplug_handler_plug() isn't sufficient?
> >>>>
> >>>> At the point of hotplug_handler_plug(), the device is not realize (realized == 0),
> >>>> however, nvdimm_get_plugged_device_list() works on realized nvdimm devices.
> >>>
> >>> I suggest instead of adding redundant hook, to reuse hotplug_handler_plug()
> >>> which is called after device::realize() successfully completed and amend
> >>> nvdimm_plugged_device_list() as follow:
> >>>
> >>> diff --git a/hw/acpi/nvdimm.c b/hw/acpi/nvdimm.c
> >>> index e486128..c6d8cbb 100644
> >>> --- a/hw/acpi/nvdimm.c
> >>> +++ b/hw/acpi/nvdimm.c
> >>> @@ -38,11 +38,7 @@ static int nvdimm_plugged_device_list(Object *obj, void *opaque)
> >>> GSList **list = opaque;
> >>>
> >>> if (object_dynamic_cast(obj, TYPE_NVDIMM)) {
> >>> - DeviceState *dev = DEVICE(obj);
> >>> -
> >>> - if (dev->realized) { /* only realized NVDIMMs matter */
> >>> - *list = g_slist_append(*list, DEVICE(obj));
> >>> - }
> >>> + *list = g_slist_append(*list, obj);
> >>> }
> >>>
> >>> object_child_foreach(obj, nvdimm_plugged_device_list, opaque);
> >>>
> >>> that should count not only already present nvdimms but also the one that's
> >>> being hotplugged.
> >>
> >> It is not good as the device which failed to initialize
> > See device_set_realized()
> > [...]
> > if (dc->realize) {
> > dc->realize(dev, &local_err);
> > }
> >
> > if (local_err != NULL) {
> > goto fail;
> > }
> >
> > DEVICE_LISTENER_CALL(realize, Forward, dev);
> >
> > if (hotplug_ctrl) {
> > hotplug_handler_plug(hotplug_ctrl, dev, &local_err);
> > }
> >
> > i.e. control reaches to hotplug_handler_plug() only if
> > call dc->realize(dev, &local_err) is successful.
> >
>
> I mean, for example. if there are two devices, the first one is failed to be
> realized, the second one is init-ed successfully, then can
> nvdimm_plugged_device_list() get these two devices?
>
> Based the code of pc_dimm_built_list(), i guess yes.
nope,
read qdev_device_add() end result:
- on success device in QOM tree
- on failure device is destroyed
>
> >> or is not being plugged
> >> (consider two nvdimm devices directly appended in QEMU command line, when we plug
> >> the first one, both nvdimm devices will be counted in this list) is also counted in
> >> this list...
> > nope,
> > qdev_device_add() is called sequentially (one time for each -device/device_add)
> > so nvdimm_plugged_device_list() sees only devices that's been added
> > by previous -device/device_add plus one extra device that's
> > being added by in progress qdev_device_add() call.
> >
> > so for "-device nvdimm,id=nv1 -device nvdimm,id=2" call sequence
> > would look like:
> > 1:
> > qdev_device_add("nvdimm,id=nv1") {
> > -> set parent // device becomes visible to nvdimm_get_plugged_device_list()
> > // this is the only time where device->realized == false
> > // could be observed, all other call chains would either
> > // see device->realized == true or not see device at all
> > -> realize()
> > -> device_set_realized()
> > -> nvdimm->realize()
> > -> hotplug_handler_plug()
> > nvdimm_get_plugged_device_list()
> > // would return list with 1 item (nv1)
> > // where nv1->realized == false)
> >
> > }
> > 2:
> > qdev_device_add("nvdimm,id=nv2") {
> > -> set parent // device becomes visible to nvdimm_get_plugged_device_list()
> > -> realize()
> > -> device_set_realized()
> > -> nvdimm->realize()
> > -> hotplug_handler_plug()
> > nvdimm_get_plugged_device_list()
> > // would return list with 2 items (nv1 and nv2)
> > // where nv1->realized == true and nv2->realized = false
> > }
> >
>
> Okay.
>
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 13:55 ` Michael S. Tsirkin
@ 2016-11-01 14:14 ` Igor Mammedov
2016-11-01 14:29 ` Peter Maydell
2016-11-01 16:03 ` Xiao Guangrong
0 siblings, 2 replies; 87+ messages in thread
From: Igor Mammedov @ 2016-11-01 14:14 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell, Xiao Guangrong
On Tue, 1 Nov 2016 15:55:36 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Tue, Nov 01, 2016 at 02:21:07PM +0100, Igor Mammedov wrote:
> > On Tue, 1 Nov 2016 00:48:11 +0200
> > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >
> > > On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
> > > > On Sun, 30 Oct 2016 23:23:18 +0200
> > > > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> > > >
> > > > > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> > > > >
> > > > > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> > > > >
> > > > > are available in the git repository at:
> > > > >
> > > > > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> > > > >
> > > > > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> > > > >
> > > > > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> > > > >
> > > > > ----------------------------------------------------------------
> > > > > virtio, pc: fixes and features
> > > > >
> > > > > nvdimm hotplug support
> > > > Michael,
> > > >
> > > > Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
> > > >
> > > > and keep only nvdimm fixes/cleanups for now?
> > >
> > > If I drop it now it won't be in the next QEMU and it seems like
> > > a valuable feature. The comments so far are about minor style
> > > improvements that IMO can be addressed by patches on top.
> > wrt nvdimm hotplug support it's not about style issues but rather
> > design issues: for example:
> > - it extends general hotplug framework unnecessarily instead of
> > figuring out how it works.
> > - adds not needed locks
> > maybe there is more and all of that was posted just a day before
> > this pull request so I haven't even had a chance to review it properly.
> >
> >
> >
> > > We can always revert if you see bigger issues, but let's enable the
> > > testing of this feature.
> > if it didn't mess with general infrastructure, I wouldn't care much.
> > But it does so I'd rather avoid merging not yet ready series just for
> > the sake of getting it in.
> >
> > I haven't reviewed 28-35 patches either but they are all cleanups/
> > fixes of current nvdimm code and local to it so don't mind them
> > getting merged.
> >
> > However I suggest dropping 36-39 patches from this pull request
> > as not yet ready for merging.
>
> So I think it's ok to keep them from now as that should help
> the feature converge faster, on the understanding the
> style issues are addressed quickly.
>
> Let's merge as is and I'll revert if this does not happen
> within two weeks. Ack?
Ack,
let's merge and revert if author won't fix issues in 1-2 weeks.
PS:
Looks like with new soft-freeze rules we just have one big
hard-freeze window and people trying to frantically squeeze
in features last minute.
I think previous soft-freeze rules worked better where
maintainers were still allowed merge features at their
discretion if series would converge in soft-freeze time
frame and be important/low risk one. Like the case here.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 14:14 ` Igor Mammedov
@ 2016-11-01 14:29 ` Peter Maydell
2016-11-01 16:03 ` Xiao Guangrong
1 sibling, 0 replies; 87+ messages in thread
From: Peter Maydell @ 2016-11-01 14:29 UTC (permalink / raw)
To: Igor Mammedov; +Cc: Michael S. Tsirkin, QEMU Developers, Xiao Guangrong
On 1 November 2016 at 14:14, Igor Mammedov <imammedo@redhat.com> wrote:
> PS:
> Looks like with new soft-freeze rules we just have one big
> hard-freeze window and people trying to frantically squeeze
> in features last minute.
> I think previous soft-freeze rules worked better where
> maintainers were still allowed merge features at their
> discretion if series would converge in soft-freeze time
> frame and be important/low risk one. Like the case here.
But you get an extra week before softfreeze deadline
(ie softfreeze is 2 weeks rather than 3), so it evens
out a bit. At a week into old-style softfreeze you
as a submaintainer should probably have been pushing
back on this as late for a new feature anyway.
thanks
-- PMM
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (47 preceding siblings ...)
2016-10-31 9:50 ` [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Igor Mammedov
@ 2016-11-01 15:22 ` Peter Maydell
2016-11-01 17:25 ` Michael S. Tsirkin
2016-11-02 4:35 ` Michael S. Tsirkin
2016-11-03 16:59 ` Stefan Hajnoczi
49 siblings, 2 replies; 87+ messages in thread
From: Peter Maydell @ 2016-11-01 15:22 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers
On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
> The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>
> Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>
> for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>
> acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>
> ----------------------------------------------------------------
> virtio, pc: fixes and features
>
> nvdimm hotplug support
> virtio migration and ioeventfd rework
> virtio crypto device
> ipmi fixes
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>
Hi; this fails to build on OSX with format string issues:
/Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
error: format specifies type 'unsign
ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
[-Werror,-Wformat]
vcrypto->max_queues, VIRTIO_QUEUE_MAX);
~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
expanded from macro 'error_setg'
(fmt), ## __VA_ARGS__)
^
Fun fact: in struct vhost_dev, max_queues is a uint64_t;
in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
it is a uint32_t...
thanks
-- PMM
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-11-01 13:58 ` Igor Mammedov
@ 2016-11-01 15:57 ` Xiao Guangrong
2016-11-01 16:26 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-01 15:57 UTC (permalink / raw)
To: Igor Mammedov
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On 11/01/2016 09:58 PM, Igor Mammedov wrote:
>>
>> I mean, for example. if there are two devices, the first one is failed to be
>> realized, the second one is init-ed successfully, then can
>> nvdimm_plugged_device_list() get these two devices?
>>
>> Based the code of pc_dimm_built_list(), i guess yes.
> nope,
> read qdev_device_add() end result:
> - on success device in QOM tree
> - on failure device is destroyed
Okay, pc_dimm_built_list let me mislead the failed case.
so this function also can be refined.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 14:14 ` Igor Mammedov
2016-11-01 14:29 ` Peter Maydell
@ 2016-11-01 16:03 ` Xiao Guangrong
1 sibling, 0 replies; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-01 16:03 UTC (permalink / raw)
To: Igor Mammedov, Michael S. Tsirkin; +Cc: Peter Maydell, qemu-devel
On 11/01/2016 10:14 PM, Igor Mammedov wrote:
> On Tue, 1 Nov 2016 15:55:36 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
>> On Tue, Nov 01, 2016 at 02:21:07PM +0100, Igor Mammedov wrote:
>>> On Tue, 1 Nov 2016 00:48:11 +0200
>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>
>>>> On Mon, Oct 31, 2016 at 10:50:31AM +0100, Igor Mammedov wrote:
>>>>> On Sun, 30 Oct 2016 23:23:18 +0200
>>>>> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>>>>>
>>>>>> The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>>>>>>
>>>>>> Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>>>>>>
>>>>>> are available in the git repository at:
>>>>>>
>>>>>> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>>>>>>
>>>>>> for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>>>>>>
>>>>>> acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>>>>>>
>>>>>> ----------------------------------------------------------------
>>>>>> virtio, pc: fixes and features
>>>>>>
>>>>>> nvdimm hotplug support
>>>>> Michael,
>>>>>
>>>>> Could you drop nvdimm hotplug from pull request (I should review at least once as it touches not only NVDIMMs but a generic hotplug infrastructure)
>>>>>
>>>>> and keep only nvdimm fixes/cleanups for now?
>>>>
>>>> If I drop it now it won't be in the next QEMU and it seems like
>>>> a valuable feature. The comments so far are about minor style
>>>> improvements that IMO can be addressed by patches on top.
>>> wrt nvdimm hotplug support it's not about style issues but rather
>>> design issues: for example:
>>> - it extends general hotplug framework unnecessarily instead of
>>> figuring out how it works.
>>> - adds not needed locks
>>> maybe there is more and all of that was posted just a day before
>>> this pull request so I haven't even had a chance to review it properly.
>>>
>>>
>>>
>>>> We can always revert if you see bigger issues, but let's enable the
>>>> testing of this feature.
>>> if it didn't mess with general infrastructure, I wouldn't care much.
>>> But it does so I'd rather avoid merging not yet ready series just for
>>> the sake of getting it in.
>>>
>>> I haven't reviewed 28-35 patches either but they are all cleanups/
>>> fixes of current nvdimm code and local to it so don't mind them
>>> getting merged.
>>>
>>> However I suggest dropping 36-39 patches from this pull request
>>> as not yet ready for merging.
>>
>> So I think it's ok to keep them from now as that should help
>> the feature converge faster, on the understanding the
>> style issues are addressed quickly.
>>
>> Let's merge as is and I'll revert if this does not happen
>> within two weeks. Ack?
> Ack,
> let's merge and revert if author won't fix issues in 1-2 weeks.
So happy to hear that. :)
I will address your comment after Peter merges these patches.
Thank you all!
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer
2016-11-01 15:57 ` Xiao Guangrong
@ 2016-11-01 16:26 ` Igor Mammedov
0 siblings, 0 replies; 87+ messages in thread
From: Igor Mammedov @ 2016-11-01 16:26 UTC (permalink / raw)
To: Xiao Guangrong
Cc: Michael S. Tsirkin, Peter Maydell, Richard Henderson, qemu-devel,
Eduardo Habkost, Paolo Bonzini
On Tue, 1 Nov 2016 23:57:55 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 11/01/2016 09:58 PM, Igor Mammedov wrote:
>
> >>
> >> I mean, for example. if there are two devices, the first one is failed to be
> >> realized, the second one is init-ed successfully, then can
> >> nvdimm_plugged_device_list() get these two devices?
> >>
> >> Based the code of pc_dimm_built_list(), i guess yes.
> > nope,
> > read qdev_device_add() end result:
> > - on success device in QOM tree
> > - on failure device is destroyed
>
> Okay, pc_dimm_built_list let me mislead the failed case.
> so this function also can be refined.
pc_dimm_built_list() uses realized flag to get already present devices
and that's not what you need.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 15:22 ` Peter Maydell
@ 2016-11-01 17:25 ` Michael S. Tsirkin
2016-11-01 17:27 ` Peter Maydell
2016-11-02 1:13 ` Gonglei (Arei)
2016-11-02 4:35 ` Michael S. Tsirkin
1 sibling, 2 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-11-01 17:25 UTC (permalink / raw)
To: Peter Maydell; +Cc: QEMU Developers
On Tue, Nov 01, 2016 at 03:22:01PM +0000, Peter Maydell wrote:
> On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> >
> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> >
> > are available in the git repository at:
> >
> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> >
> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> >
> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> >
> > ----------------------------------------------------------------
> > virtio, pc: fixes and features
> >
> > nvdimm hotplug support
> > virtio migration and ioeventfd rework
> > virtio crypto device
> > ipmi fixes
> >
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >
>
> Hi; this fails to build on OSX with format string issues:
>
> /Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
> error: format specifies type 'unsign
> ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
> [-Werror,-Wformat]
> vcrypto->max_queues, VIRTIO_QUEUE_MAX);
> ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> /Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
> expanded from macro 'error_setg'
> (fmt), ## __VA_ARGS__)
> ^
>
> Fun fact: in struct vhost_dev, max_queues is a uint64_t;
> in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
> it is a uint32_t...
>
> thanks
> -- PMM
Yes - I really think we should move that to virtio core
going forward.
Anyway, I fixed that up and pushed.
A general question - I have a couple of tweaks to tests in
my tree. Didn't push to avoid them blocking the features.
I think test tweaks are fair game after freeze - agree?
--
MST
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 17:25 ` Michael S. Tsirkin
@ 2016-11-01 17:27 ` Peter Maydell
2016-11-02 1:13 ` Gonglei (Arei)
1 sibling, 0 replies; 87+ messages in thread
From: Peter Maydell @ 2016-11-01 17:27 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers
On 1 November 2016 at 17:25, Michael S. Tsirkin <mst@redhat.com> wrote:
> A general question - I have a couple of tweaks to tests in
> my tree. Didn't push to avoid them blocking the features.
> I think test tweaks are fair game after freeze - agree?
Yes, especially if they're only tweaks, not enormous
changes.
thanks
-- PMM
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 17:25 ` Michael S. Tsirkin
2016-11-01 17:27 ` Peter Maydell
@ 2016-11-02 1:13 ` Gonglei (Arei)
1 sibling, 0 replies; 87+ messages in thread
From: Gonglei (Arei) @ 2016-11-02 1:13 UTC (permalink / raw)
To: Michael S. Tsirkin, Peter Maydell; +Cc: QEMU Developers
> -----Original Message-----
> From: Qemu-devel
> [mailto:qemu-devel-bounces+arei.gonglei=huawei.com@nongnu.org] On
> Behalf Of Michael S. Tsirkin
> Sent: Wednesday, November 02, 2016 1:26 AM
> To: Peter Maydell
> Cc: QEMU Developers
> Subject: Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
>
> On Tue, Nov 01, 2016 at 03:22:01PM +0000, Peter Maydell wrote:
> > On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
> > > The following changes since commit
> 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> > >
> > > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1'
> into staging (2016-10-28 17:59:04 +0100)
> > >
> > > are available in the git repository at:
> > >
> > > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> > >
> > > for you to fetch changes up to
> f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> > >
> > > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25
> +0200)
> > >
> > > ----------------------------------------------------------------
> > > virtio, pc: fixes and features
> > >
> > > nvdimm hotplug support
> > > virtio migration and ioeventfd rework
> > > virtio crypto device
> > > ipmi fixes
> > >
> > > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > >
> >
> > Hi; this fails to build on OSX with format string issues:
> >
> > /Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
> > error: format specifies type 'unsign
> > ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
> > [-Werror,-Wformat]
> > vcrypto->max_queues, VIRTIO_QUEUE_MAX);
> > ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > /Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
> > expanded from macro 'error_setg'
> > (fmt), ## __VA_ARGS__)
> > ^
> >
> > Fun fact: in struct vhost_dev, max_queues is a uint64_t;
> > in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
> > it is a uint32_t...
> >
> > thanks
> > -- PMM
>
> Yes - I really think we should move that to virtio core
> going forward.
> Anyway, I fixed that up and pushed.
>
Sorry about that. TBH that error log messge is copied from virtio-net.c and
I didn't notice their types are different. For virtio-crypto, it should be 'PRIu32'.
Regards,
-Gonglei
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-01 15:22 ` Peter Maydell
2016-11-01 17:25 ` Michael S. Tsirkin
@ 2016-11-02 4:35 ` Michael S. Tsirkin
2016-11-02 13:14 ` Peter Maydell
1 sibling, 1 reply; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-11-02 4:35 UTC (permalink / raw)
To: Peter Maydell; +Cc: QEMU Developers
On Tue, Nov 01, 2016 at 03:22:01PM +0000, Peter Maydell wrote:
> On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> >
> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> >
> > are available in the git repository at:
> >
> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> >
> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> >
> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> >
> > ----------------------------------------------------------------
> > virtio, pc: fixes and features
> >
> > nvdimm hotplug support
> > virtio migration and ioeventfd rework
> > virtio crypto device
> > ipmi fixes
> >
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >
>
> Hi; this fails to build on OSX with format string issues:
>
> /Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
> error: format specifies type 'unsign
> ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
> [-Werror,-Wformat]
> vcrypto->max_queues, VIRTIO_QUEUE_MAX);
> ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> /Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
> expanded from macro 'error_setg'
> (fmt), ## __VA_ARGS__)
> ^
>
> Fun fact: in struct vhost_dev, max_queues is a uint64_t;
> in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
> it is a uint32_t...
>
> thanks
> -- PMM
Just to make sure : I fixed that and pushed to
same tag. I don't think it makes sense to repost the pull
request - pls take it from
git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
--
MST
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug
2016-10-30 21:25 ` [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug Michael S. Tsirkin
@ 2016-11-02 11:19 ` Igor Mammedov
2016-11-02 16:00 ` Xiao Guangrong
0 siblings, 1 reply; 87+ messages in thread
From: Igor Mammedov @ 2016-11-02 11:19 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: qemu-devel, Peter Maydell, Xiao Guangrong, Paolo Bonzini,
Richard Henderson, Eduardo Habkost
On Sun, 30 Oct 2016 23:25:10 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>
> _GPE.E04 is dedicated for nvdimm device hotplug
>
> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> include/hw/acpi/acpi_dev_interface.h | 1 +
> hw/acpi/memory_hotplug.c | 31 +++++++++++++++++++++++--------
> hw/i386/acpi-build.c | 7 +++++++
> hw/i386/pc.c | 12 ++++++++++++
> hw/mem/nvdimm.c | 4 ----
> docs/specs/acpi_mem_hotplug.txt | 3 +++
> 6 files changed, 46 insertions(+), 12 deletions(-)
>
> diff --git a/include/hw/acpi/acpi_dev_interface.h b/include/hw/acpi/acpi_dev_interface.h
> index da4ef7f..901a4ae 100644
> --- a/include/hw/acpi/acpi_dev_interface.h
> +++ b/include/hw/acpi/acpi_dev_interface.h
> @@ -10,6 +10,7 @@ typedef enum {
> ACPI_PCI_HOTPLUG_STATUS = 2,
> ACPI_CPU_HOTPLUG_STATUS = 4,
> ACPI_MEMORY_HOTPLUG_STATUS = 8,
> + ACPI_NVDIMM_HOTPLUG_STATUS = 16,
> } AcpiEventStatusBits;
>
> #define TYPE_ACPI_DEVICE_IF "acpi-device-interface"
> diff --git a/hw/acpi/memory_hotplug.c b/hw/acpi/memory_hotplug.c
> index ec4e64b..70f6451 100644
> --- a/hw/acpi/memory_hotplug.c
> +++ b/hw/acpi/memory_hotplug.c
> @@ -2,6 +2,7 @@
> #include "hw/acpi/memory_hotplug.h"
> #include "hw/acpi/pc-hotplug.h"
> #include "hw/mem/pc-dimm.h"
> +#include "hw/mem/nvdimm.h"
> #include "hw/boards.h"
> #include "hw/qdev-core.h"
> #include "trace.h"
> @@ -232,11 +233,8 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
> DeviceState *dev, Error **errp)
> {
> MemStatus *mdev;
> - DeviceClass *dc = DEVICE_GET_CLASS(dev);
> -
> - if (!dc->hotpluggable) {
> - return;
> - }
shouldn't be removed.
> + AcpiEventStatusBits event;
> + bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
>
> mdev = acpi_memory_slot_status(mem_st, dev, errp);
> if (!mdev) {
> @@ -244,10 +242,23 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
> }
>
> mdev->dimm = dev;
> - mdev->is_enabled = true;
> +
> + /*
> + * do not set is_enabled and is_inserting if the slot is plugged with
> + * a nvdimm device to stop OSPM inquires memory region from the slot.
> + */
> + if (is_nvdimm) {
> + event = ACPI_NVDIMM_HOTPLUG_STATUS;
> + } else {
> + mdev->is_enabled = true;
> + event = ACPI_MEMORY_HOTPLUG_STATUS;
> + }
> +
> if (dev->hotplugged) {
> - mdev->is_inserting = true;
> - acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
> + if (!is_nvdimm) {
> + mdev->is_inserting = true;
> + }
> + acpi_send_event(DEVICE(hotplug_dev), event);
> }
> }
>
> @@ -262,6 +273,8 @@ void acpi_memory_unplug_request_cb(HotplugHandler *hotplug_dev,
> return;
> }
>
> + /* nvdimm device hot unplug is not supported yet. */
> + assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
that would crash guest instead of failing gracefully as it's supposed to.
> mdev->is_removing = true;
> acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
> }
> @@ -276,6 +289,8 @@ void acpi_memory_unplug_cb(MemHotplugState *mem_st,
> return;
> }
>
> + /* nvdimm device hot unplug is not supported yet. */
> + assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
ditto
> mdev->is_enabled = false;
> mdev->dimm = NULL;
> }
> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> index 03a5386..7aaa07a 100644
> --- a/hw/i386/acpi-build.c
> +++ b/hw/i386/acpi-build.c
> @@ -2069,6 +2069,13 @@ build_dsdt(GArray *table_data, BIOSLinker *linker,
> method = aml_method("_E03", 0, AML_NOTSERIALIZED);
> aml_append(method, aml_call0(MEMORY_HOTPLUG_HANDLER_PATH));
> aml_append(scope, method);
> +
> + if (pcms->acpi_nvdimm_state.is_enabled) {
> + method = aml_method("_E04", 0, AML_NOTSERIALIZED);
> + aml_append(method, aml_notify(aml_name("\\_SB.NVDR"),
> + aml_int(0x80)));
> + aml_append(scope, method);
> + }
> }
> aml_append(dsdt, scope);
>
> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> index b395717..c011552 100644
> --- a/hw/i386/pc.c
> +++ b/hw/i386/pc.c
> @@ -1744,6 +1744,12 @@ static void pc_dimm_unplug_request(HotplugHandler *hotplug_dev,
> goto out;
> }
>
> + if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
> + error_setg(&local_err,
> + "nvdimm device hot unplug is not supported yet.");
> + goto out;
> + }
> +
> hhc = HOTPLUG_HANDLER_GET_CLASS(pcms->acpi_dev);
> hhc->unplug_request(HOTPLUG_HANDLER(pcms->acpi_dev), dev, &local_err);
>
> @@ -1761,6 +1767,12 @@ static void pc_dimm_unplug(HotplugHandler *hotplug_dev,
> HotplugHandlerClass *hhc;
> Error *local_err = NULL;
>
> + if (object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM)) {
> + error_setg(&local_err,
> + "nvdimm device hot unplug is not supported yet.");
> + goto out;
> + }
> +
> hhc = HOTPLUG_HANDLER_GET_CLASS(pcms->acpi_dev);
> hhc->unplug(HOTPLUG_HANDLER(pcms->acpi_dev), dev, &local_err);
>
> diff --git a/hw/mem/nvdimm.c b/hw/mem/nvdimm.c
> index 7895805..db896b0 100644
> --- a/hw/mem/nvdimm.c
> +++ b/hw/mem/nvdimm.c
> @@ -148,13 +148,9 @@ static MemoryRegion *nvdimm_get_vmstate_memory_region(PCDIMMDevice *dimm)
>
> static void nvdimm_class_init(ObjectClass *oc, void *data)
> {
> - DeviceClass *dc = DEVICE_CLASS(oc);
> PCDIMMDeviceClass *ddc = PC_DIMM_CLASS(oc);
> NVDIMMClass *nvc = NVDIMM_CLASS(oc);
>
> - /* nvdimm hotplug has not been supported yet. */
> - dc->hotpluggable = false;
> -
> ddc->realize = nvdimm_realize;
> ddc->get_memory_region = nvdimm_get_memory_region;
> ddc->get_vmstate_memory_region = nvdimm_get_vmstate_memory_region;
> diff --git a/docs/specs/acpi_mem_hotplug.txt b/docs/specs/acpi_mem_hotplug.txt
> index 3df3620..cb26dd2 100644
> --- a/docs/specs/acpi_mem_hotplug.txt
> +++ b/docs/specs/acpi_mem_hotplug.txt
> @@ -4,6 +4,9 @@ QEMU<->ACPI BIOS memory hotplug interface
> ACPI BIOS GPE.3 handler is dedicated for notifying OS about memory hot-add
> and hot-remove events.
>
> +ACPI BIOS GPE.4 handler is dedicated for notifying OS about nvdimm device
> +hot-add and hot-remove events.
> +
> Memory hot-plug interface (IO port 0xa00-0xa17, 1-4 byte access):
> ---------------------------------------------------------------
> 0xa00:
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-02 4:35 ` Michael S. Tsirkin
@ 2016-11-02 13:14 ` Peter Maydell
2016-11-03 3:35 ` Michael S. Tsirkin
0 siblings, 1 reply; 87+ messages in thread
From: Peter Maydell @ 2016-11-02 13:14 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers, Stefan Hajnoczi
On 2 November 2016 at 04:35, Michael S. Tsirkin <mst@redhat.com> wrote:
> On Tue, Nov 01, 2016 at 03:22:01PM +0000, Peter Maydell wrote:
>> On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
>> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>> >
>> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>> >
>> > are available in the git repository at:
>> >
>> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>> >
>> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>> >
>> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>> >
>> > ----------------------------------------------------------------
>> > virtio, pc: fixes and features
>> >
>> > nvdimm hotplug support
>> > virtio migration and ioeventfd rework
>> > virtio crypto device
>> > ipmi fixes
>> >
>> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>> >
>>
>> Hi; this fails to build on OSX with format string issues:
>>
>> /Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
>> error: format specifies type 'unsign
>> ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
>> [-Werror,-Wformat]
>> vcrypto->max_queues, VIRTIO_QUEUE_MAX);
>> ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> /Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
>> expanded from macro 'error_setg'
>> (fmt), ## __VA_ARGS__)
>> ^
>>
>> Fun fact: in struct vhost_dev, max_queues is a uint64_t;
>> in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
>> it is a uint32_t...
>>
>> thanks
>> -- PMM
>
> Just to make sure : I fixed that and pushed to
> same tag. I don't think it makes sense to repost the pull
> request - pls take it from
> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
You should always repost at least the cover-letter part
of a fresh pull request, because otherwise it is likely
to not be caught by the email filters that find pull requests.
(In this case I've handed over pull request processing
to Stefan, so the question would be whether his filters
notice.)
thanks
-- PMM
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug
2016-11-02 11:19 ` Igor Mammedov
@ 2016-11-02 16:00 ` Xiao Guangrong
2016-11-03 9:41 ` Igor Mammedov
0 siblings, 1 reply; 87+ messages in thread
From: Xiao Guangrong @ 2016-11-02 16:00 UTC (permalink / raw)
To: Igor Mammedov, Michael S. Tsirkin
Cc: qemu-devel, Peter Maydell, Paolo Bonzini, Richard Henderson,
Eduardo Habkost
On 11/02/2016 07:19 PM, Igor Mammedov wrote:
> On Sun, 30 Oct 2016 23:25:10 +0200
> "Michael S. Tsirkin" <mst@redhat.com> wrote:
>
>> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>>
>> _GPE.E04 is dedicated for nvdimm device hotplug
>>
>> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>> ---
>> include/hw/acpi/acpi_dev_interface.h | 1 +
>> hw/acpi/memory_hotplug.c | 31 +++++++++++++++++++++++--------
>> hw/i386/acpi-build.c | 7 +++++++
>> hw/i386/pc.c | 12 ++++++++++++
>> hw/mem/nvdimm.c | 4 ----
>> docs/specs/acpi_mem_hotplug.txt | 3 +++
>> 6 files changed, 46 insertions(+), 12 deletions(-)
>>
>> diff --git a/include/hw/acpi/acpi_dev_interface.h b/include/hw/acpi/acpi_dev_interface.h
>> index da4ef7f..901a4ae 100644
>> --- a/include/hw/acpi/acpi_dev_interface.h
>> +++ b/include/hw/acpi/acpi_dev_interface.h
>> @@ -10,6 +10,7 @@ typedef enum {
>> ACPI_PCI_HOTPLUG_STATUS = 2,
>> ACPI_CPU_HOTPLUG_STATUS = 4,
>> ACPI_MEMORY_HOTPLUG_STATUS = 8,
>> + ACPI_NVDIMM_HOTPLUG_STATUS = 16,
>> } AcpiEventStatusBits;
>>
>> #define TYPE_ACPI_DEVICE_IF "acpi-device-interface"
>> diff --git a/hw/acpi/memory_hotplug.c b/hw/acpi/memory_hotplug.c
>> index ec4e64b..70f6451 100644
>> --- a/hw/acpi/memory_hotplug.c
>> +++ b/hw/acpi/memory_hotplug.c
>> @@ -2,6 +2,7 @@
>> #include "hw/acpi/memory_hotplug.h"
>> #include "hw/acpi/pc-hotplug.h"
>> #include "hw/mem/pc-dimm.h"
>> +#include "hw/mem/nvdimm.h"
>> #include "hw/boards.h"
>> #include "hw/qdev-core.h"
>> #include "trace.h"
>> @@ -232,11 +233,8 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
>> DeviceState *dev, Error **errp)
>> {
>> MemStatus *mdev;
>> - DeviceClass *dc = DEVICE_GET_CLASS(dev);
>> -
>> - if (!dc->hotpluggable) {
>> - return;
>> - }
> shouldn't be removed.
Well, all memory devices support hotplug now, i thought it
is reasonable to drop it, actually it was introduced by nvdimm...
>
>> + AcpiEventStatusBits event;
>> + bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
>>
>> mdev = acpi_memory_slot_status(mem_st, dev, errp);
>> if (!mdev) {
>> @@ -244,10 +242,23 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
>> }
>>
>> mdev->dimm = dev;
>> - mdev->is_enabled = true;
>> +
>> + /*
>> + * do not set is_enabled and is_inserting if the slot is plugged with
>> + * a nvdimm device to stop OSPM inquires memory region from the slot.
>> + */
>> + if (is_nvdimm) {
>> + event = ACPI_NVDIMM_HOTPLUG_STATUS;
>> + } else {
>> + mdev->is_enabled = true;
>> + event = ACPI_MEMORY_HOTPLUG_STATUS;
>> + }
>> +
>> if (dev->hotplugged) {
>> - mdev->is_inserting = true;
>> - acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
>> + if (!is_nvdimm) {
>> + mdev->is_inserting = true;
>> + }
>> + acpi_send_event(DEVICE(hotplug_dev), event);
>> }
>> }
>>
>> @@ -262,6 +273,8 @@ void acpi_memory_unplug_request_cb(HotplugHandler *hotplug_dev,
>> return;
>> }
>>
>> + /* nvdimm device hot unplug is not supported yet. */
>> + assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
> that would crash guest instead of failing gracefully as it's supposed to.
This is really a bug if it is triggered as nvdimm hot unplug is
stopped in the upper caller.
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-11-02 13:14 ` Peter Maydell
@ 2016-11-03 3:35 ` Michael S. Tsirkin
0 siblings, 0 replies; 87+ messages in thread
From: Michael S. Tsirkin @ 2016-11-03 3:35 UTC (permalink / raw)
To: Peter Maydell; +Cc: QEMU Developers, Stefan Hajnoczi
On Wed, Nov 02, 2016 at 01:14:07PM +0000, Peter Maydell wrote:
> On 2 November 2016 at 04:35, Michael S. Tsirkin <mst@redhat.com> wrote:
> > On Tue, Nov 01, 2016 at 03:22:01PM +0000, Peter Maydell wrote:
> >> On 30 October 2016 at 21:23, Michael S. Tsirkin <mst@redhat.com> wrote:
> >> > The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
> >> >
> >> > Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
> >> >
> >> > are available in the git repository at:
> >> >
> >> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
> >> >
> >> > for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
> >> >
> >> > acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
> >> >
> >> > ----------------------------------------------------------------
> >> > virtio, pc: fixes and features
> >> >
> >> > nvdimm hotplug support
> >> > virtio migration and ioeventfd rework
> >> > virtio crypto device
> >> > ipmi fixes
> >> >
> >> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >> >
> >>
> >> Hi; this fails to build on OSX with format string issues:
> >>
> >> /Users/pm215/src/qemu-for-merges/hw/virtio/virtio-crypto.c:770:20:
> >> error: format specifies type 'unsign
> >> ed short' but the argument has type 'uint32_t' (aka 'unsigned int')
> >> [-Werror,-Wformat]
> >> vcrypto->max_queues, VIRTIO_QUEUE_MAX);
> >> ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >> /Users/pm215/src/qemu-for-merges/include/qapi/error.h:163:35: note:
> >> expanded from macro 'error_setg'
> >> (fmt), ## __VA_ARGS__)
> >> ^
> >>
> >> Fun fact: in struct vhost_dev, max_queues is a uint64_t;
> >> in struct VirtIONet it is a uint16_t; and in VirtIOCrypto
> >> it is a uint32_t...
> >>
> >> thanks
> >> -- PMM
> >
> > Just to make sure : I fixed that and pushed to
> > same tag. I don't think it makes sense to repost the pull
> > request - pls take it from
> > git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>
> You should always repost at least the cover-letter part
> of a fresh pull request, because otherwise it is likely
> to not be caught by the email filters that find pull requests.
>
> (In this case I've handed over pull request processing
> to Stefan, so the question would be whether his filters
> notice.)
>
> thanks
> -- PMM
Stefan, could you confirm pls?
--
MST
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug
2016-11-02 16:00 ` Xiao Guangrong
@ 2016-11-03 9:41 ` Igor Mammedov
0 siblings, 0 replies; 87+ messages in thread
From: Igor Mammedov @ 2016-11-03 9:41 UTC (permalink / raw)
To: Xiao Guangrong
Cc: Michael S. Tsirkin, qemu-devel, Peter Maydell, Paolo Bonzini,
Richard Henderson, Eduardo Habkost
On Thu, 3 Nov 2016 00:00:45 +0800
Xiao Guangrong <guangrong.xiao@linux.intel.com> wrote:
> On 11/02/2016 07:19 PM, Igor Mammedov wrote:
> > On Sun, 30 Oct 2016 23:25:10 +0200
> > "Michael S. Tsirkin" <mst@redhat.com> wrote:
> >
> >> From: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >>
> >> _GPE.E04 is dedicated for nvdimm device hotplug
> >>
> >> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
> >> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> >> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >> ---
> >> include/hw/acpi/acpi_dev_interface.h | 1 +
> >> hw/acpi/memory_hotplug.c | 31 +++++++++++++++++++++++--------
> >> hw/i386/acpi-build.c | 7 +++++++
> >> hw/i386/pc.c | 12 ++++++++++++
> >> hw/mem/nvdimm.c | 4 ----
> >> docs/specs/acpi_mem_hotplug.txt | 3 +++
> >> 6 files changed, 46 insertions(+), 12 deletions(-)
> >>
> >> diff --git a/include/hw/acpi/acpi_dev_interface.h b/include/hw/acpi/acpi_dev_interface.h
> >> index da4ef7f..901a4ae 100644
> >> --- a/include/hw/acpi/acpi_dev_interface.h
> >> +++ b/include/hw/acpi/acpi_dev_interface.h
> >> @@ -10,6 +10,7 @@ typedef enum {
> >> ACPI_PCI_HOTPLUG_STATUS = 2,
> >> ACPI_CPU_HOTPLUG_STATUS = 4,
> >> ACPI_MEMORY_HOTPLUG_STATUS = 8,
> >> + ACPI_NVDIMM_HOTPLUG_STATUS = 16,
> >> } AcpiEventStatusBits;
> >>
> >> #define TYPE_ACPI_DEVICE_IF "acpi-device-interface"
> >> diff --git a/hw/acpi/memory_hotplug.c b/hw/acpi/memory_hotplug.c
> >> index ec4e64b..70f6451 100644
> >> --- a/hw/acpi/memory_hotplug.c
> >> +++ b/hw/acpi/memory_hotplug.c
> >> @@ -2,6 +2,7 @@
> >> #include "hw/acpi/memory_hotplug.h"
> >> #include "hw/acpi/pc-hotplug.h"
> >> #include "hw/mem/pc-dimm.h"
> >> +#include "hw/mem/nvdimm.h"
> >> #include "hw/boards.h"
> >> #include "hw/qdev-core.h"
> >> #include "trace.h"
> >> @@ -232,11 +233,8 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
> >> DeviceState *dev, Error **errp)
> >> {
> >> MemStatus *mdev;
> >> - DeviceClass *dc = DEVICE_GET_CLASS(dev);
> >> -
> >> - if (!dc->hotpluggable) {
> >> - return;
> >> - }
> > shouldn't be removed.
>
> Well, all memory devices support hotplug now, i thought it
> is reasonable to drop it, actually it was introduced by nvdimm...
this hunk could still be useful if we would have non hotpluggable pc-dimm
but yes for now you can remove it, a separate patch would be even better.
>
> >
> >> + AcpiEventStatusBits event;
> >> + bool is_nvdimm = object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM);
> >>
> >> mdev = acpi_memory_slot_status(mem_st, dev, errp);
> >> if (!mdev) {
> >> @@ -244,10 +242,23 @@ void acpi_memory_plug_cb(HotplugHandler *hotplug_dev, MemHotplugState *mem_st,
> >> }
> >>
> >> mdev->dimm = dev;
> >> - mdev->is_enabled = true;
> >> +
> >> + /*
> >> + * do not set is_enabled and is_inserting if the slot is plugged with
> >> + * a nvdimm device to stop OSPM inquires memory region from the slot.
> >> + */
> >> + if (is_nvdimm) {
> >> + event = ACPI_NVDIMM_HOTPLUG_STATUS;
> >> + } else {
> >> + mdev->is_enabled = true;
> >> + event = ACPI_MEMORY_HOTPLUG_STATUS;
> >> + }
> >> +
> >> if (dev->hotplugged) {
> >> - mdev->is_inserting = true;
> >> - acpi_send_event(DEVICE(hotplug_dev), ACPI_MEMORY_HOTPLUG_STATUS);
> >> + if (!is_nvdimm) {
> >> + mdev->is_inserting = true;
> >> + }
> >> + acpi_send_event(DEVICE(hotplug_dev), event);
> >> }
> >> }
> >>
> >> @@ -262,6 +273,8 @@ void acpi_memory_unplug_request_cb(HotplugHandler *hotplug_dev,
> >> return;
> >> }
> >>
> >> + /* nvdimm device hot unplug is not supported yet. */
> >> + assert(!object_dynamic_cast(OBJECT(dev), TYPE_NVDIMM));
> > that would crash guest instead of failing gracefully as it's supposed to.
>
> This is really a bug if it is triggered as nvdimm hot unplug is
> stopped in the upper caller.
point here is not crash VM on hotplug/unplug path if possible,
it's fragile to rely on checks somewhere else as that code
might change and cause crashes later.
just use normal error handling infrastructure for this.
i.e. error_setg()
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
` (48 preceding siblings ...)
2016-11-01 15:22 ` Peter Maydell
@ 2016-11-03 16:59 ` Stefan Hajnoczi
49 siblings, 0 replies; 87+ messages in thread
From: Stefan Hajnoczi @ 2016-11-03 16:59 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: qemu-devel, Peter Maydell
[-- Attachment #1: Type: text/plain, Size: 7493 bytes --]
On Sun, Oct 30, 2016 at 11:23:18PM +0200, Michael S. Tsirkin wrote:
> The following changes since commit 5b2ecabaeabc17f032197246c4846b9ba95ba8a6:
>
> Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20161028-1' into staging (2016-10-28 17:59:04 +0100)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
>
> for you to fetch changes up to f082ec0225bd15c71e0b4697d2df3af7bad65d7f:
>
> acpi: fix assert failure caused by commit 35c5a52d (2016-10-30 20:06:25 +0200)
>
> ----------------------------------------------------------------
> virtio, pc: fixes and features
>
> nvdimm hotplug support
> virtio migration and ioeventfd rework
> virtio crypto device
> ipmi fixes
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>
> ----------------------------------------------------------------
> Corey Minyard (5):
> ipmi: Remove hotplug from IPMI BMCs
> ipmi_bmc_sim: Remove an unnecessary mutex
> ipmi: Implement shutdown via ACPI overtemp
> ipmi: Add graceful shutdown handling to the external BMC
> acpi/ipmi: Initialize the fwinfo before fetching it
>
> Cédric Le Goater (1):
> ipmi: chassis poweroff should use qemu_system_shutdown_request()
>
> Daniel P. Berrange (1):
> ipmi: fix build config variable name for ipmi_bmc_extern.o
>
> Dr. David Alan Gilbert (2):
> virtio/migration: Add VMStateDescription to VirtioDeviceClass
> virtio/migration: Migrate balloon to VMState
>
> Gonglei (12):
> cryptodev: introduce cryptodev backend interface
> cryptodev: add symmetric algorithm operation stuff
> virtio-crypto: introduce virtio_crypto.h
> cryptodev: introduce a new cryptodev backend
> virtio-crypto: add virtio crypto device emulation
> virtio-crypto-pci: add virtio crypto pci support
> virtio-crypto: set capacity of algorithms supported
> virtio-crypto: add control queue handler
> virtio-crypto: add data queue processing handler
> cryptodev: introduce an unified wrapper for crypto operation
> virtio-crypto: using bh to handle dataq's requests
> virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer
>
> Haozhong Zhang (1):
> acpi: fix assert failure caused by commit 35c5a52d
>
> Paolo Bonzini (13):
> virtio: disable ioeventfd as early as possible
> virtio: move ioeventfd_disabled flag to VirtioBusState
> virtio: move ioeventfd_started flag to VirtioBusState
> virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass
> virtio: introduce virtio_device_ioeventfd_enabled
> virtio-blk: always use dataplane path if ioeventfd is active
> virtio-scsi: always use dataplane path if ioeventfd is active
> Revert "virtio: Introduce virtio_add_queue_aio"
> virtio: remove set_handler argument from set_host_notifier_internal
> virtio: remove ioeventfd_disabled altogether
> virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd
> virtio: inline virtio_queue_set_host_notifier_fd_handler
> virtio: inline set_host_notifier_internal
>
> Xiao Guangrong (12):
> acpi nvdimm: fix wrong buffer size returned by DSM method
> acpi nvdimm: fix OperationRegion definition
> acpi nvdimm: fix device physical address base
> acpi nvdimm: fix ARG3 conflict
> acpi nvdimm: fix Arg6 usage
> nvdimm acpi: compile nvdimm acpi code arch-independently
> acpi nvdimm: rename result_size to dsm_out_buf_siz
> nvdimm acpi: use common macros instead of magic names
> nvdimm acpi: prebuild nvdimm devices for available slots
> nvdimm acpi: introduce fit buffer
> nvdimm acpi: introduce _FIT
> pc: memhp: enable nvdimm device hotplug
>
> hw/block/dataplane/virtio-blk.h | 6 +-
> hw/s390x/virtio-ccw.h | 2 -
> hw/virtio/virtio-pci.h | 17 +-
> include/hw/acpi/acpi_dev_interface.h | 1 +
> include/hw/hotplug.h | 10 +
> include/hw/mem/nvdimm.h | 27 +-
> include/hw/virtio/virtio-bus.h | 27 +-
> include/hw/virtio/virtio-crypto.h | 101 +++
> include/hw/virtio/virtio-scsi.h | 6 +-
> include/hw/virtio/virtio.h | 15 +-
> include/standard-headers/linux/virtio_crypto.h | 429 ++++++++++++
> include/standard-headers/linux/virtio_ids.h | 2 +-
> include/sysemu/cryptodev.h | 298 ++++++++
> backends/cryptodev-builtin.c | 361 ++++++++++
> backends/cryptodev.c | 245 +++++++
> hw/acpi/ipmi.c | 1 +
> hw/acpi/memory_hotplug.c | 31 +-
> hw/acpi/nvdimm.c | 468 ++++++++++---
> hw/block/dataplane/virtio-blk.c | 73 +-
> hw/block/virtio-blk.c | 15 +-
> hw/core/hotplug.c | 11 +
> hw/core/qdev.c | 20 +-
> hw/i386/acpi-build.c | 9 +-
> hw/i386/pc.c | 31 +
> hw/ipmi/ipmi.c | 10 +-
> hw/ipmi/ipmi_bmc_extern.c | 12 +-
> hw/ipmi/ipmi_bmc_sim.c | 7 +-
> hw/mem/nvdimm.c | 4 -
> hw/s390x/virtio-ccw.c | 44 +-
> hw/scsi/virtio-scsi-dataplane.c | 56 +-
> hw/scsi/virtio-scsi.c | 24 +-
> hw/virtio/vhost.c | 5 +-
> hw/virtio/virtio-balloon.c | 31 +-
> hw/virtio/virtio-bus.c | 154 ++---
> hw/virtio/virtio-crypto-pci.c | 77 +++
> hw/virtio/virtio-crypto.c | 898 +++++++++++++++++++++++++
> hw/virtio/virtio-mmio.c | 35 +-
> hw/virtio/virtio-pci.c | 40 +-
> hw/virtio/virtio.c | 153 +++--
> tests/ipmi-bt-test.c | 2 +-
> MAINTAINERS | 13 +
> backends/Makefile.objs | 3 +
> docs/specs/acpi_mem_hotplug.txt | 3 +
> docs/specs/acpi_nvdimm.txt | 58 +-
> hw/acpi/Makefile.objs | 2 +-
> hw/ipmi/Makefile.objs | 2 +-
> hw/virtio/Makefile.objs | 2 +
> qemu-options.hx | 18 +
> 48 files changed, 3352 insertions(+), 507 deletions(-)
> create mode 100644 include/hw/virtio/virtio-crypto.h
> create mode 100644 include/standard-headers/linux/virtio_crypto.h
> create mode 100644 include/sysemu/cryptodev.h
> create mode 100644 backends/cryptodev-builtin.c
> create mode 100644 backends/cryptodev.c
> create mode 100644 hw/virtio/virtio-crypto-pci.c
> create mode 100644 hw/virtio/virtio-crypto.c
>
>
Thanks, applied latest pull request to my staging tree:
https://github.com/stefanha/qemu/commits/staging
Will be in qemu.git/master pending Travis-CI build tests.
Stefan
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-10-30 21:23 ` [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether Michael S. Tsirkin
@ 2016-11-10 14:35 ` Christian Borntraeger
2016-11-10 14:38 ` Paolo Bonzini
0 siblings, 1 reply; 87+ messages in thread
From: Christian Borntraeger @ 2016-11-10 14:35 UTC (permalink / raw)
To: Michael S. Tsirkin, qemu-devel
Cc: Cornelia Huck, Peter Maydell, Paolo Bonzini
On 10/30/2016 10:23 PM, Michael S. Tsirkin wrote:
> From: Paolo Bonzini <pbonzini@redhat.com>
>
> Now that there is not anymore a switch from the generic ioeventfd handler
> to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
> always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
> does nothing in this case. Move the invocation to vhost.c, which is the
> only place that needs it.
>
> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> include/hw/virtio/virtio-bus.h | 6 ------
> hw/virtio/vhost.c | 3 +++
> hw/virtio/virtio-bus.c | 23 ++++++++---------------
> 3 files changed, 11 insertions(+), 21 deletions(-)
This breaks vhost-net for s390/kvm after rebooting the guest. (ping fails and
ifconfig shows no packets is TXed)
Any idea?
>
> diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
> index af6b5c4..cbdf745 100644
> --- a/include/hw/virtio/virtio-bus.h
> +++ b/include/hw/virtio/virtio-bus.h
> @@ -94,12 +94,6 @@ struct VirtioBusState {
> BusState parent_obj;
>
> /*
> - * Set if the default ioeventfd handlers are disabled by vhost
> - * or dataplane.
> - */
> - bool ioeventfd_disabled;
> -
> - /*
> * Set if ioeventfd has been started.
> */
> bool ioeventfd_started;
> diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
> index 501a5f4..131f164 100644
> --- a/hw/virtio/vhost.c
> +++ b/hw/virtio/vhost.c
> @@ -1196,6 +1196,7 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
> goto fail;
> }
>
> + virtio_device_stop_ioeventfd(vdev);
> for (i = 0; i < hdev->nvqs; ++i) {
> r = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
> true);
> @@ -1215,6 +1216,7 @@ fail_vq:
> }
> assert (e >= 0);
> }
> + virtio_device_start_ioeventfd(vdev);
> fail:
> return r;
> }
> @@ -1237,6 +1239,7 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
> }
> assert (r >= 0);
> }
> + virtio_device_start_ioeventfd(vdev);
> }
>
> /* Test and clear event pending status.
> diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
> index a619445..b0e4544 100644
> --- a/hw/virtio/virtio-bus.c
> +++ b/hw/virtio/virtio-bus.c
> @@ -190,7 +190,7 @@ int virtio_bus_start_ioeventfd(VirtioBusState *bus)
> if (!k->ioeventfd_assign || !k->ioeventfd_enabled(proxy)) {
> return -ENOSYS;
> }
> - if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
> + if (bus->ioeventfd_started) {
> return 0;
> }
> r = vdc->start_ioeventfd(vdev);
> @@ -226,8 +226,8 @@ bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus)
> }
>
> /*
> - * This function switches from/to the generic ioeventfd handler.
> - * assign==false means 'use generic ioeventfd handler'.
> + * This function switches ioeventfd on/off in the device.
> + * The caller must set or clear the handlers for the EventNotifier.
> */
> int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
> {
> @@ -237,19 +237,12 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
> if (!k->ioeventfd_assign) {
> return -ENOSYS;
> }
> - bus->ioeventfd_disabled = assign;
> if (assign) {
> - /*
> - * Stop using the generic ioeventfd, we are doing eventfd handling
> - * ourselves below
> - *
> - * FIXME: We should just switch the handler and not deassign the
> - * ioeventfd.
> - * Otherwise, there's a window where we don't have an
> - * ioeventfd and we may end up with a notification where
> - * we don't expect one.
> - */
> - virtio_bus_stop_ioeventfd(bus);
> + assert(!bus->ioeventfd_started);
> + } else {
> + if (!bus->ioeventfd_started) {
> + return 0;
> + }
> }
> return set_host_notifier_internal(proxy, bus, n, assign);
> }
>
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-11-10 14:35 ` Christian Borntraeger
@ 2016-11-10 14:38 ` Paolo Bonzini
2016-11-10 14:48 ` Christian Borntraeger
0 siblings, 1 reply; 87+ messages in thread
From: Paolo Bonzini @ 2016-11-10 14:38 UTC (permalink / raw)
To: Christian Borntraeger, Michael S. Tsirkin, qemu-devel
Cc: Cornelia Huck, Peter Maydell
On 10/11/2016 15:35, Christian Borntraeger wrote:
> On 10/30/2016 10:23 PM, Michael S. Tsirkin wrote:
>> From: Paolo Bonzini <pbonzini@redhat.com>
>>
>> Now that there is not anymore a switch from the generic ioeventfd handler
>> to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
>> always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
>> does nothing in this case. Move the invocation to vhost.c, which is the
>> only place that needs it.
>>
>> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>> ---
>> include/hw/virtio/virtio-bus.h | 6 ------
>> hw/virtio/vhost.c | 3 +++
>> hw/virtio/virtio-bus.c | 23 ++++++++---------------
>> 3 files changed, 11 insertions(+), 21 deletions(-)
>
> This breaks vhost-net for s390/kvm after rebooting the guest. (ping fails and
> ifconfig shows no packets is TXed)
>
> Any idea?
Patch from Felipe:
[PATCH v2] vhost: Update 'ioeventfd_started' with host notifiers
Paolo
>>
>> diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
>> index af6b5c4..cbdf745 100644
>> --- a/include/hw/virtio/virtio-bus.h
>> +++ b/include/hw/virtio/virtio-bus.h
>> @@ -94,12 +94,6 @@ struct VirtioBusState {
>> BusState parent_obj;
>>
>> /*
>> - * Set if the default ioeventfd handlers are disabled by vhost
>> - * or dataplane.
>> - */
>> - bool ioeventfd_disabled;
>> -
>> - /*
>> * Set if ioeventfd has been started.
>> */
>> bool ioeventfd_started;
>> diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
>> index 501a5f4..131f164 100644
>> --- a/hw/virtio/vhost.c
>> +++ b/hw/virtio/vhost.c
>> @@ -1196,6 +1196,7 @@ int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
>> goto fail;
>> }
>>
>> + virtio_device_stop_ioeventfd(vdev);
>> for (i = 0; i < hdev->nvqs; ++i) {
>> r = virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), hdev->vq_index + i,
>> true);
>> @@ -1215,6 +1216,7 @@ fail_vq:
>> }
>> assert (e >= 0);
>> }
>> + virtio_device_start_ioeventfd(vdev);
>> fail:
>> return r;
>> }
>> @@ -1237,6 +1239,7 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev)
>> }
>> assert (r >= 0);
>> }
>> + virtio_device_start_ioeventfd(vdev);
>> }
>>
>> /* Test and clear event pending status.
>> diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
>> index a619445..b0e4544 100644
>> --- a/hw/virtio/virtio-bus.c
>> +++ b/hw/virtio/virtio-bus.c
>> @@ -190,7 +190,7 @@ int virtio_bus_start_ioeventfd(VirtioBusState *bus)
>> if (!k->ioeventfd_assign || !k->ioeventfd_enabled(proxy)) {
>> return -ENOSYS;
>> }
>> - if (bus->ioeventfd_started || bus->ioeventfd_disabled) {
>> + if (bus->ioeventfd_started) {
>> return 0;
>> }
>> r = vdc->start_ioeventfd(vdev);
>> @@ -226,8 +226,8 @@ bool virtio_bus_ioeventfd_enabled(VirtioBusState *bus)
>> }
>>
>> /*
>> - * This function switches from/to the generic ioeventfd handler.
>> - * assign==false means 'use generic ioeventfd handler'.
>> + * This function switches ioeventfd on/off in the device.
>> + * The caller must set or clear the handlers for the EventNotifier.
>> */
>> int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
>> {
>> @@ -237,19 +237,12 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int n, bool assign)
>> if (!k->ioeventfd_assign) {
>> return -ENOSYS;
>> }
>> - bus->ioeventfd_disabled = assign;
>> if (assign) {
>> - /*
>> - * Stop using the generic ioeventfd, we are doing eventfd handling
>> - * ourselves below
>> - *
>> - * FIXME: We should just switch the handler and not deassign the
>> - * ioeventfd.
>> - * Otherwise, there's a window where we don't have an
>> - * ioeventfd and we may end up with a notification where
>> - * we don't expect one.
>> - */
>> - virtio_bus_stop_ioeventfd(bus);
>> + assert(!bus->ioeventfd_started);
>> + } else {
>> + if (!bus->ioeventfd_started) {
>> + return 0;
>> + }
>> }
>> return set_host_notifier_internal(proxy, bus, n, assign);
>> }
>>
>
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-11-10 14:38 ` Paolo Bonzini
@ 2016-11-10 14:48 ` Christian Borntraeger
2016-11-15 8:27 ` Christian Borntraeger
0 siblings, 1 reply; 87+ messages in thread
From: Christian Borntraeger @ 2016-11-10 14:48 UTC (permalink / raw)
To: Paolo Bonzini, Michael S. Tsirkin, qemu-devel
Cc: Cornelia Huck, Peter Maydell
On 11/10/2016 03:38 PM, Paolo Bonzini wrote:
>
>
> On 10/11/2016 15:35, Christian Borntraeger wrote:
>> On 10/30/2016 10:23 PM, Michael S. Tsirkin wrote:
>>> From: Paolo Bonzini <pbonzini@redhat.com>
>>>
>>> Now that there is not anymore a switch from the generic ioeventfd handler
>>> to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
>>> always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
>>> does nothing in this case. Move the invocation to vhost.c, which is the
>>> only place that needs it.
>>>
>>> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>> ---
>>> include/hw/virtio/virtio-bus.h | 6 ------
>>> hw/virtio/vhost.c | 3 +++
>>> hw/virtio/virtio-bus.c | 23 ++++++++---------------
>>> 3 files changed, 11 insertions(+), 21 deletions(-)
>>
>> This breaks vhost-net for s390/kvm after rebooting the guest. (ping fails and
>> ifconfig shows no packets is TXed)
>>
>> Any idea?
>
> Patch from Felipe:
> [PATCH v2] vhost: Update 'ioeventfd_started' with host notifiers
>
> Paolo
Yes, that fixes the issue
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-11-10 14:48 ` Christian Borntraeger
@ 2016-11-15 8:27 ` Christian Borntraeger
2016-11-15 10:00 ` Paolo Bonzini
0 siblings, 1 reply; 87+ messages in thread
From: Christian Borntraeger @ 2016-11-15 8:27 UTC (permalink / raw)
To: Paolo Bonzini, Michael S. Tsirkin, qemu-devel
Cc: Cornelia Huck, Peter Maydell, Farhan Ali
On 11/10/2016 03:48 PM, Christian Borntraeger wrote:
> On 11/10/2016 03:38 PM, Paolo Bonzini wrote:
>>
>>
>> On 10/11/2016 15:35, Christian Borntraeger wrote:
>>> On 10/30/2016 10:23 PM, Michael S. Tsirkin wrote:
>>>> From: Paolo Bonzini <pbonzini@redhat.com>
>>>>
>>>> Now that there is not anymore a switch from the generic ioeventfd handler
>>>> to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
>>>> always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
>>>> does nothing in this case. Move the invocation to vhost.c, which is the
>>>> only place that needs it.
>>>>
>>>> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>>> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
>>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>>> ---
>>>> include/hw/virtio/virtio-bus.h | 6 ------
>>>> hw/virtio/vhost.c | 3 +++
>>>> hw/virtio/virtio-bus.c | 23 ++++++++---------------
>>>> 3 files changed, 11 insertions(+), 21 deletions(-)
>>>
>>> This breaks vhost-net for s390/kvm after rebooting the guest. (ping fails and
>>> ifconfig shows no packets is TXed)
>>>
>>> Any idea?
>>
>> Patch from Felipe:
>> [PATCH v2] vhost: Update 'ioeventfd_started' with host notifiers
>>
>> Paolo
>
> Yes, that fixes the issue
>
hmm, not quite.
This patch on top of
commit 6bbcb76301a72dc80c8d29af13d40bb9a759c9c6 MAINTAINERS: Remove obsolete stable branches
still triggers
qemu-system-s390x: hw/s390x/virtio-ccw.c:1076: virtio_ccw_remove_irqfd: Assertion `ret == 0' failed.
for some cases according to Farhan.
There are too many patches floating around, what is the latest assumed to be
final patch set? We can then test that.
Thanks
Christian
^ permalink raw reply [flat|nested] 87+ messages in thread
* Re: [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether
2016-11-15 8:27 ` Christian Borntraeger
@ 2016-11-15 10:00 ` Paolo Bonzini
0 siblings, 0 replies; 87+ messages in thread
From: Paolo Bonzini @ 2016-11-15 10:00 UTC (permalink / raw)
To: Christian Borntraeger, Michael S. Tsirkin, qemu-devel
Cc: Cornelia Huck, Peter Maydell, Farhan Ali
On 15/11/2016 09:27, Christian Borntraeger wrote:
> hmm, not quite.
> This patch on top of
> commit 6bbcb76301a72dc80c8d29af13d40bb9a759c9c6 MAINTAINERS: Remove obsolete stable branches
> still triggers
>
> qemu-system-s390x: hw/s390x/virtio-ccw.c:1076: virtio_ccw_remove_irqfd: Assertion `ret == 0' failed.
> for some cases according to Farhan.
>
> There are too many patches floating around, what is the latest assumed to be
> final patch set? We can then test that.
I'm sending it as soon as I finish testing. But for s390 it should be
just one patch, the rest only affects PCI and MMIO (and it had been
broken forever for dataplane, so the refactoring did its job by exposing
everyone to dataplane bugs, and even before the release :)).
Paolo
^ permalink raw reply [flat|nested] 87+ messages in thread
end of thread, other threads:[~2016-11-15 10:00 UTC | newest]
Thread overview: 87+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-10-30 21:23 [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 01/47] virtio/migration: Add VMStateDescription to VirtioDeviceClass Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 02/47] virtio/migration: Migrate balloon to VMState Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 03/47] virtio: disable ioeventfd as early as possible Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 04/47] virtio: move ioeventfd_disabled flag to VirtioBusState Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 05/47] virtio: move ioeventfd_started " Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 06/47] virtio: add start_ioeventfd and stop_ioeventfd to VirtioDeviceClass Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 07/47] virtio: introduce virtio_device_ioeventfd_enabled Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 08/47] virtio-blk: always use dataplane path if ioeventfd is active Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 09/47] virtio-scsi: " Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 10/47] Revert "virtio: Introduce virtio_add_queue_aio" Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 11/47] virtio: remove set_handler argument from set_host_notifier_internal Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 12/47] virtio: remove ioeventfd_disabled altogether Michael S. Tsirkin
2016-11-10 14:35 ` Christian Borntraeger
2016-11-10 14:38 ` Paolo Bonzini
2016-11-10 14:48 ` Christian Borntraeger
2016-11-15 8:27 ` Christian Borntraeger
2016-11-15 10:00 ` Paolo Bonzini
2016-10-30 21:23 ` [Qemu-devel] [PULL 13/47] virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd Michael S. Tsirkin
2016-10-30 21:23 ` [Qemu-devel] [PULL 14/47] virtio: inline virtio_queue_set_host_notifier_fd_handler Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 15/47] virtio: inline set_host_notifier_internal Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 16/47] cryptodev: introduce cryptodev backend interface Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 17/47] cryptodev: add symmetric algorithm operation stuff Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 18/47] virtio-crypto: introduce virtio_crypto.h Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 19/47] cryptodev: introduce a new cryptodev backend Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 20/47] virtio-crypto: add virtio crypto device emulation Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 21/47] virtio-crypto-pci: add virtio crypto pci support Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 22/47] virtio-crypto: set capacity of algorithms supported Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 23/47] virtio-crypto: add control queue handler Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 24/47] virtio-crypto: add data queue processing handler Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 25/47] cryptodev: introduce an unified wrapper for crypto operation Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 26/47] virtio-crypto: using bh to handle dataq's requests Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 27/47] virtio-crypto: add myself as virtio-crypto and cryptodev backends maintainer Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 28/47] acpi nvdimm: fix wrong buffer size returned by DSM method Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 29/47] acpi nvdimm: fix OperationRegion definition Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 30/47] acpi nvdimm: fix device physical address base Michael S. Tsirkin
2016-10-31 9:20 ` Igor Mammedov
2016-10-31 9:23 ` Xiao Guangrong
2016-10-31 10:56 ` Igor Mammedov
2016-10-31 11:09 ` Xiao Guangrong
2016-10-31 13:30 ` Igor Mammedov
2016-10-30 21:24 ` [Qemu-devel] [PULL 31/47] acpi nvdimm: fix ARG3 conflict Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 32/47] acpi nvdimm: fix Arg6 usage Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 33/47] nvdimm acpi: compile nvdimm acpi code arch-independently Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 34/47] acpi nvdimm: rename result_size to dsm_out_buf_siz Michael S. Tsirkin
2016-10-30 21:24 ` [Qemu-devel] [PULL 35/47] nvdimm acpi: use common macros instead of magic names Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 36/47] nvdimm acpi: prebuild nvdimm devices for available slots Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 37/47] nvdimm acpi: introduce fit buffer Michael S. Tsirkin
2016-10-31 9:45 ` Igor Mammedov
2016-10-31 9:52 ` Xiao Guangrong
2016-10-31 11:09 ` Igor Mammedov
2016-11-01 3:30 ` Xiao Guangrong
2016-11-01 10:35 ` Igor Mammedov
2016-11-01 13:40 ` Xiao Guangrong
2016-11-01 13:58 ` Igor Mammedov
2016-11-01 15:57 ` Xiao Guangrong
2016-11-01 16:26 ` Igor Mammedov
2016-10-30 21:25 ` [Qemu-devel] [PULL 38/47] nvdimm acpi: introduce _FIT Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 39/47] pc: memhp: enable nvdimm device hotplug Michael S. Tsirkin
2016-11-02 11:19 ` Igor Mammedov
2016-11-02 16:00 ` Xiao Guangrong
2016-11-03 9:41 ` Igor Mammedov
2016-10-30 21:25 ` [Qemu-devel] [PULL 40/47] ipmi: Remove hotplug from IPMI BMCs Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 41/47] ipmi_bmc_sim: Remove an unnecessary mutex Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 42/47] ipmi: chassis poweroff should use qemu_system_shutdown_request() Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 43/47] ipmi: Implement shutdown via ACPI overtemp Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 44/47] ipmi: fix build config variable name for ipmi_bmc_extern.o Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 45/47] ipmi: Add graceful shutdown handling to the external BMC Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 46/47] acpi/ipmi: Initialize the fwinfo before fetching it Michael S. Tsirkin
2016-10-30 21:25 ` [Qemu-devel] [PULL 47/47] acpi: fix assert failure caused by commit 35c5a52d Michael S. Tsirkin
2016-10-31 9:50 ` [Qemu-devel] [PULL 00/47] virtio, pc: fixes and features Igor Mammedov
2016-10-31 22:48 ` Michael S. Tsirkin
2016-11-01 12:47 ` Peter Maydell
2016-11-01 13:21 ` Igor Mammedov
2016-11-01 13:45 ` Xiao Guangrong
2016-11-01 13:55 ` Michael S. Tsirkin
2016-11-01 14:14 ` Igor Mammedov
2016-11-01 14:29 ` Peter Maydell
2016-11-01 16:03 ` Xiao Guangrong
2016-11-01 15:22 ` Peter Maydell
2016-11-01 17:25 ` Michael S. Tsirkin
2016-11-01 17:27 ` Peter Maydell
2016-11-02 1:13 ` Gonglei (Arei)
2016-11-02 4:35 ` Michael S. Tsirkin
2016-11-02 13:14 ` Peter Maydell
2016-11-03 3:35 ` Michael S. Tsirkin
2016-11-03 16:59 ` Stefan Hajnoczi
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.