* [PULL 0/6] pc,vhost: fixes
@ 2020-12-02 11:03 Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 1/6] vhost-user-scsi: Fix memleaks in vus_proc_req() Michael S. Tsirkin
` (6 more replies)
0 siblings, 7 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell
Patch 5 gave me pause but we do need patch 6 as
it's a guest triggerable assert, and it seemed
cleaner to just take the whole patchset than cherry-pick.
The following changes since commit d73c46e4a84e47ffc61b8bf7c378b1383e7316b5:
Update version for v5.2.0-rc4 release (2020-12-01 16:21:01 +0000)
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 a451089140c7b167661d0a23190e6d8bd7eae48d:
memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type (2020-12-02 05:14:29 -0500)
----------------------------------------------------------------
pc,vhost: fixes
A couple of last minute bugfixes.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
----------------------------------------------------------------
Alex Chen (1):
vhost-user-scsi: Fix memleaks in vus_proc_req()
Eugenio Pérez (5):
memory: Rename memory_region_notify_one to memory_region_notify_iommu_one
memory: Add IOMMUTLBEvent
memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType
intel_iommu: Skip page walking on device iotlb invalidations
memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type
include/exec/memory.h | 40 ++++++++------
contrib/vhost-user-scsi/vhost-user-scsi.c | 3 +-
hw/arm/smmu-common.c | 13 +++--
hw/arm/smmuv3.c | 13 +++--
hw/i386/intel_iommu.c | 92 ++++++++++++++++++-------------
hw/misc/tz-mpc.c | 32 ++++++-----
hw/ppc/spapr_iommu.c | 15 ++---
hw/s390x/s390-pci-inst.c | 27 +++++----
hw/virtio/vhost.c | 2 +-
hw/virtio/virtio-iommu.c | 30 +++++-----
softmmu/memory.c | 29 ++++++----
11 files changed, 167 insertions(+), 129 deletions(-)
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PULL 1/6] vhost-user-scsi: Fix memleaks in vus_proc_req()
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 2/6] memory: Rename memory_region_notify_one to memory_region_notify_iommu_one Michael S. Tsirkin
` (5 subsequent siblings)
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel; +Cc: Alex Chen, Peter Maydell, Raphael Norwitz, Euler Robot
From: Alex Chen <alex.chen@huawei.com>
The 'elem' is allocated memory in vu_queue_pop(), and its memory should be
freed in all error branches after vu_queue_pop().
In addition, in order to free the 'elem' memory outside of while(1) loop, move
the definition of 'elem' to the beginning of vus_proc_req().
Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Alex Chen <alex.chen@huawei.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <20201125013055.34147-1-alex.chen@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
contrib/vhost-user-scsi/vhost-user-scsi.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/contrib/vhost-user-scsi/vhost-user-scsi.c b/contrib/vhost-user-scsi/vhost-user-scsi.c
index 0f9ba4b2a2..4639440a70 100644
--- a/contrib/vhost-user-scsi/vhost-user-scsi.c
+++ b/contrib/vhost-user-scsi/vhost-user-scsi.c
@@ -232,6 +232,7 @@ static void vus_proc_req(VuDev *vu_dev, int idx)
VugDev *gdev;
VusDev *vdev_scsi;
VuVirtq *vq;
+ VuVirtqElement *elem = NULL;
assert(vu_dev);
@@ -248,7 +249,6 @@ static void vus_proc_req(VuDev *vu_dev, int idx)
g_debug("Got kicked on vq[%d]@%p", idx, vq);
while (1) {
- VuVirtqElement *elem;
VirtIOSCSICmdReq *req;
VirtIOSCSICmdResp *rsp;
@@ -288,6 +288,7 @@ static void vus_proc_req(VuDev *vu_dev, int idx)
free(elem);
}
+ free(elem);
}
static void vus_queue_set_started(VuDev *vu_dev, int idx, bool started)
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PULL 2/6] memory: Rename memory_region_notify_one to memory_region_notify_iommu_one
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 1/6] vhost-user-scsi: Fix memleaks in vus_proc_req() Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 3/6] memory: Add IOMMUTLBEvent Michael S. Tsirkin
` (4 subsequent siblings)
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Eduardo Habkost, Juan Quintela, Jason Wang,
Richard Henderson, Peter Xu, Eric Auger, qemu-arm, Paolo Bonzini,
Eugenio Pérez, David Gibson
From: Eugenio Pérez <eperezma@redhat.com>
Previous name didn't reflect the iommu operation.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-2-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/exec/memory.h | 6 +++---
hw/arm/smmu-common.c | 2 +-
hw/arm/smmuv3.c | 2 +-
hw/i386/intel_iommu.c | 4 ++--
softmmu/memory.c | 6 +++---
5 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/include/exec/memory.h b/include/exec/memory.h
index 0f3e6bcd5e..d8456ccf52 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -236,7 +236,7 @@ enum IOMMUMemoryRegionAttr {
* The IOMMU implementation must use the IOMMU notifier infrastructure
* to report whenever mappings are changed, by calling
* memory_region_notify_iommu() (or, if necessary, by calling
- * memory_region_notify_one() for each registered notifier).
+ * memory_region_notify_iommu_one() for each registered notifier).
*
* Conceptually an IOMMU provides a mapping from input address
* to an output TLB entry. If the IOMMU is aware of memory transaction
@@ -1346,7 +1346,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
IOMMUTLBEntry entry);
/**
- * memory_region_notify_one: notify a change in an IOMMU translation
+ * memory_region_notify_iommu_one: notify a change in an IOMMU translation
* entry to a single notifier
*
* This works just like memory_region_notify_iommu(), but it only
@@ -1357,7 +1357,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
* replaces all old entries for the same virtual I/O address range.
* Deleted entries have .@perm == 0.
*/
-void memory_region_notify_one(IOMMUNotifier *notifier,
+void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
IOMMUTLBEntry *entry);
/**
diff --git a/hw/arm/smmu-common.c b/hw/arm/smmu-common.c
index 3838db1395..88d2c454f0 100644
--- a/hw/arm/smmu-common.c
+++ b/hw/arm/smmu-common.c
@@ -472,7 +472,7 @@ static void smmu_unmap_notifier_range(IOMMUNotifier *n)
entry.perm = IOMMU_NONE;
entry.addr_mask = n->end - n->start;
- memory_region_notify_one(n, &entry);
+ memory_region_notify_iommu_one(n, &entry);
}
/* Unmap all notifiers attached to @mr */
diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c
index 22607c3784..273f5f7dce 100644
--- a/hw/arm/smmuv3.c
+++ b/hw/arm/smmuv3.c
@@ -828,7 +828,7 @@ static void smmuv3_notify_iova(IOMMUMemoryRegion *mr,
entry.addr_mask = num_pages * (1 << granule) - 1;
entry.perm = IOMMU_NONE;
- memory_region_notify_one(n, &entry);
+ memory_region_notify_iommu_one(n, &entry);
}
/* invalidate an asid/iova range tuple in all mr's */
diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index 70ac837733..067593b9e4 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -3497,7 +3497,7 @@ static void vtd_address_space_unmap(VTDAddressSpace *as, IOMMUNotifier *n)
/* This field is meaningless for unmap */
entry.translated_addr = 0;
- memory_region_notify_one(n, &entry);
+ memory_region_notify_iommu_one(n, &entry);
start += mask;
remain -= mask;
@@ -3535,7 +3535,7 @@ static void vtd_address_space_refresh_all(IntelIOMMUState *s)
static int vtd_replay_hook(IOMMUTLBEntry *entry, void *private)
{
- memory_region_notify_one((IOMMUNotifier *)private, entry);
+ memory_region_notify_iommu_one((IOMMUNotifier *)private, entry);
return 0;
}
diff --git a/softmmu/memory.c b/softmmu/memory.c
index 11ca94d037..44de610c72 100644
--- a/softmmu/memory.c
+++ b/softmmu/memory.c
@@ -1942,8 +1942,8 @@ void memory_region_unregister_iommu_notifier(MemoryRegion *mr,
memory_region_update_iommu_notify_flags(iommu_mr, NULL);
}
-void memory_region_notify_one(IOMMUNotifier *notifier,
- IOMMUTLBEntry *entry)
+void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
+ IOMMUTLBEntry *entry)
{
IOMMUNotifierFlag request_flags;
hwaddr entry_end = entry->iova + entry->addr_mask;
@@ -1979,7 +1979,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
IOMMU_NOTIFIER_FOREACH(iommu_notifier, iommu_mr) {
if (iommu_notifier->iommu_idx == iommu_idx) {
- memory_region_notify_one(iommu_notifier, &entry);
+ memory_region_notify_iommu_one(iommu_notifier, &entry);
}
}
}
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PULL 3/6] memory: Add IOMMUTLBEvent
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 1/6] vhost-user-scsi: Fix memleaks in vus_proc_req() Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 2/6] memory: Rename memory_region_notify_one to memory_region_notify_iommu_one Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 4/6] memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType Michael S. Tsirkin
` (3 subsequent siblings)
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Thomas Huth, Eduardo Habkost, Matthew Rosato,
Juan Quintela, Jason Wang, Cornelia Huck, Richard Henderson,
Peter Xu, Halil Pasic, Eugenio Pérez, qemu-s390x, qemu-arm,
qemu-ppc, Eric Auger, Paolo Bonzini, David Hildenbrand,
Christian Borntraeger, David Gibson
From: Eugenio Pérez <eperezma@redhat.com>
This way we can tell between regular IOMMUTLBEntry (entry of IOMMU
hardware) and notifications.
In the notifications, we set explicitly if it is a MAPs or an UNMAP,
instead of trusting in entry permissions to differentiate them.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-3-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
---
include/exec/memory.h | 27 ++++++------
hw/arm/smmu-common.c | 13 +++---
hw/arm/smmuv3.c | 13 +++---
hw/i386/intel_iommu.c | 88 ++++++++++++++++++++++------------------
hw/misc/tz-mpc.c | 32 ++++++++-------
hw/ppc/spapr_iommu.c | 15 +++----
hw/s390x/s390-pci-inst.c | 27 +++++++-----
hw/virtio/virtio-iommu.c | 30 +++++++-------
softmmu/memory.c | 20 ++++-----
9 files changed, 143 insertions(+), 122 deletions(-)
diff --git a/include/exec/memory.h b/include/exec/memory.h
index d8456ccf52..e86b5e92da 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -116,6 +116,11 @@ struct IOMMUNotifier {
};
typedef struct IOMMUNotifier IOMMUNotifier;
+typedef struct IOMMUTLBEvent {
+ IOMMUNotifierFlag type;
+ IOMMUTLBEntry entry;
+} IOMMUTLBEvent;
+
/* RAM is pre-allocated and passed into qemu_ram_alloc_from_ptr */
#define RAM_PREALLOC (1 << 0)
@@ -1326,24 +1331,18 @@ uint64_t memory_region_iommu_get_min_page_size(IOMMUMemoryRegion *iommu_mr);
/**
* memory_region_notify_iommu: notify a change in an IOMMU translation entry.
*
- * The notification type will be decided by entry.perm bits:
- *
- * - For UNMAP (cache invalidation) notifies: set entry.perm to IOMMU_NONE.
- * - For MAP (newly added entry) notifies: set entry.perm to the
- * permission of the page (which is definitely !IOMMU_NONE).
- *
* Note: for any IOMMU implementation, an in-place mapping change
* should be notified with an UNMAP followed by a MAP.
*
* @iommu_mr: the memory region that was changed
* @iommu_idx: the IOMMU index for the translation table which has changed
- * @entry: the new entry in the IOMMU translation table. The entry
- * replaces all old entries for the same virtual I/O address range.
- * Deleted entries have .@perm == 0.
+ * @event: TLB event with the new entry in the IOMMU translation table.
+ * The entry replaces all old entries for the same virtual I/O address
+ * range.
*/
void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
int iommu_idx,
- IOMMUTLBEntry entry);
+ IOMMUTLBEvent event);
/**
* memory_region_notify_iommu_one: notify a change in an IOMMU translation
@@ -1353,12 +1352,12 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
* notifies a specific notifier, not all of them.
*
* @notifier: the notifier to be notified
- * @entry: the new entry in the IOMMU translation table. The entry
- * replaces all old entries for the same virtual I/O address range.
- * Deleted entries have .@perm == 0.
+ * @event: TLB event with the new entry in the IOMMU translation table.
+ * The entry replaces all old entries for the same virtual I/O address
+ * range.
*/
void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
- IOMMUTLBEntry *entry);
+ IOMMUTLBEvent *event);
/**
* memory_region_register_iommu_notifier: register a notifier for changes to
diff --git a/hw/arm/smmu-common.c b/hw/arm/smmu-common.c
index 88d2c454f0..405d5c5325 100644
--- a/hw/arm/smmu-common.c
+++ b/hw/arm/smmu-common.c
@@ -465,14 +465,15 @@ IOMMUMemoryRegion *smmu_iommu_mr(SMMUState *s, uint32_t sid)
/* Unmap the whole notifier's range */
static void smmu_unmap_notifier_range(IOMMUNotifier *n)
{
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
- entry.target_as = &address_space_memory;
- entry.iova = n->start;
- entry.perm = IOMMU_NONE;
- entry.addr_mask = n->end - n->start;
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.target_as = &address_space_memory;
+ event.entry.iova = n->start;
+ event.entry.perm = IOMMU_NONE;
+ event.entry.addr_mask = n->end - n->start;
- memory_region_notify_iommu_one(n, &entry);
+ memory_region_notify_iommu_one(n, &event);
}
/* Unmap all notifiers attached to @mr */
diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c
index 273f5f7dce..bbca0e9f20 100644
--- a/hw/arm/smmuv3.c
+++ b/hw/arm/smmuv3.c
@@ -800,7 +800,7 @@ static void smmuv3_notify_iova(IOMMUMemoryRegion *mr,
uint8_t tg, uint64_t num_pages)
{
SMMUDevice *sdev = container_of(mr, SMMUDevice, iommu);
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
uint8_t granule = tg;
if (!tg) {
@@ -823,12 +823,13 @@ static void smmuv3_notify_iova(IOMMUMemoryRegion *mr,
granule = tt->granule_sz;
}
- entry.target_as = &address_space_memory;
- entry.iova = iova;
- entry.addr_mask = num_pages * (1 << granule) - 1;
- entry.perm = IOMMU_NONE;
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.target_as = &address_space_memory;
+ event.entry.iova = iova;
+ event.entry.addr_mask = num_pages * (1 << granule) - 1;
+ event.entry.perm = IOMMU_NONE;
- memory_region_notify_iommu_one(n, &entry);
+ memory_region_notify_iommu_one(n, &event);
}
/* invalidate an asid/iova range tuple in all mr's */
diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index 067593b9e4..56180b1c43 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -1073,7 +1073,7 @@ static int vtd_iova_to_slpte(IntelIOMMUState *s, VTDContextEntry *ce,
}
}
-typedef int (*vtd_page_walk_hook)(IOMMUTLBEntry *entry, void *private);
+typedef int (*vtd_page_walk_hook)(IOMMUTLBEvent *event, void *private);
/**
* Constant information used during page walking
@@ -1094,11 +1094,12 @@ typedef struct {
uint16_t domain_id;
} vtd_page_walk_info;
-static int vtd_page_walk_one(IOMMUTLBEntry *entry, vtd_page_walk_info *info)
+static int vtd_page_walk_one(IOMMUTLBEvent *event, vtd_page_walk_info *info)
{
VTDAddressSpace *as = info->as;
vtd_page_walk_hook hook_fn = info->hook_fn;
void *private = info->private;
+ IOMMUTLBEntry *entry = &event->entry;
DMAMap target = {
.iova = entry->iova,
.size = entry->addr_mask,
@@ -1107,7 +1108,7 @@ static int vtd_page_walk_one(IOMMUTLBEntry *entry, vtd_page_walk_info *info)
};
DMAMap *mapped = iova_tree_find(as->iova_tree, &target);
- if (entry->perm == IOMMU_NONE && !info->notify_unmap) {
+ if (event->type == IOMMU_NOTIFIER_UNMAP && !info->notify_unmap) {
trace_vtd_page_walk_one_skip_unmap(entry->iova, entry->addr_mask);
return 0;
}
@@ -1115,7 +1116,7 @@ static int vtd_page_walk_one(IOMMUTLBEntry *entry, vtd_page_walk_info *info)
assert(hook_fn);
/* Update local IOVA mapped ranges */
- if (entry->perm) {
+ if (event->type == IOMMU_NOTIFIER_MAP) {
if (mapped) {
/* If it's exactly the same translation, skip */
if (!memcmp(mapped, &target, sizeof(target))) {
@@ -1141,19 +1142,21 @@ static int vtd_page_walk_one(IOMMUTLBEntry *entry, vtd_page_walk_info *info)
int ret;
/* Emulate an UNMAP */
+ event->type = IOMMU_NOTIFIER_UNMAP;
entry->perm = IOMMU_NONE;
trace_vtd_page_walk_one(info->domain_id,
entry->iova,
entry->translated_addr,
entry->addr_mask,
entry->perm);
- ret = hook_fn(entry, private);
+ ret = hook_fn(event, private);
if (ret) {
return ret;
}
/* Drop any existing mapping */
iova_tree_remove(as->iova_tree, &target);
- /* Recover the correct permission */
+ /* Recover the correct type */
+ event->type = IOMMU_NOTIFIER_MAP;
entry->perm = cache_perm;
}
}
@@ -1170,7 +1173,7 @@ static int vtd_page_walk_one(IOMMUTLBEntry *entry, vtd_page_walk_info *info)
trace_vtd_page_walk_one(info->domain_id, entry->iova,
entry->translated_addr, entry->addr_mask,
entry->perm);
- return hook_fn(entry, private);
+ return hook_fn(event, private);
}
/**
@@ -1191,7 +1194,7 @@ static int vtd_page_walk_level(dma_addr_t addr, uint64_t start,
uint32_t offset;
uint64_t slpte;
uint64_t subpage_size, subpage_mask;
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
uint64_t iova = start;
uint64_t iova_next;
int ret = 0;
@@ -1245,13 +1248,15 @@ static int vtd_page_walk_level(dma_addr_t addr, uint64_t start,
*
* In either case, we send an IOTLB notification down.
*/
- entry.target_as = &address_space_memory;
- entry.iova = iova & subpage_mask;
- entry.perm = IOMMU_ACCESS_FLAG(read_cur, write_cur);
- entry.addr_mask = ~subpage_mask;
+ event.entry.target_as = &address_space_memory;
+ event.entry.iova = iova & subpage_mask;
+ event.entry.perm = IOMMU_ACCESS_FLAG(read_cur, write_cur);
+ event.entry.addr_mask = ~subpage_mask;
/* NOTE: this is only meaningful if entry_valid == true */
- entry.translated_addr = vtd_get_slpte_addr(slpte, info->aw);
- ret = vtd_page_walk_one(&entry, info);
+ event.entry.translated_addr = vtd_get_slpte_addr(slpte, info->aw);
+ event.type = event.entry.perm ? IOMMU_NOTIFIER_MAP :
+ IOMMU_NOTIFIER_UNMAP;
+ ret = vtd_page_walk_one(&event, info);
}
if (ret < 0) {
@@ -1430,10 +1435,10 @@ static int vtd_dev_to_context_entry(IntelIOMMUState *s, uint8_t bus_num,
return 0;
}
-static int vtd_sync_shadow_page_hook(IOMMUTLBEntry *entry,
+static int vtd_sync_shadow_page_hook(IOMMUTLBEvent *event,
void *private)
{
- memory_region_notify_iommu((IOMMUMemoryRegion *)private, 0, *entry);
+ memory_region_notify_iommu(private, 0, *event);
return 0;
}
@@ -1993,14 +1998,17 @@ static void vtd_iotlb_page_invalidate_notify(IntelIOMMUState *s,
* page tables. We just deliver the PSI down to
* invalidate caches.
*/
- IOMMUTLBEntry entry = {
- .target_as = &address_space_memory,
- .iova = addr,
- .translated_addr = 0,
- .addr_mask = size - 1,
- .perm = IOMMU_NONE,
+ IOMMUTLBEvent event = {
+ .type = IOMMU_NOTIFIER_UNMAP,
+ .entry = {
+ .target_as = &address_space_memory,
+ .iova = addr,
+ .translated_addr = 0,
+ .addr_mask = size - 1,
+ .perm = IOMMU_NONE,
+ },
};
- memory_region_notify_iommu(&vtd_as->iommu, 0, entry);
+ memory_region_notify_iommu(&vtd_as->iommu, 0, event);
}
}
}
@@ -2412,7 +2420,7 @@ static bool vtd_process_device_iotlb_desc(IntelIOMMUState *s,
VTDInvDesc *inv_desc)
{
VTDAddressSpace *vtd_dev_as;
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
struct VTDBus *vtd_bus;
hwaddr addr;
uint64_t sz;
@@ -2460,12 +2468,13 @@ static bool vtd_process_device_iotlb_desc(IntelIOMMUState *s,
sz = VTD_PAGE_SIZE;
}
- entry.target_as = &vtd_dev_as->as;
- entry.addr_mask = sz - 1;
- entry.iova = addr;
- entry.perm = IOMMU_NONE;
- entry.translated_addr = 0;
- memory_region_notify_iommu(&vtd_dev_as->iommu, 0, entry);
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.target_as = &vtd_dev_as->as;
+ event.entry.addr_mask = sz - 1;
+ event.entry.iova = addr;
+ event.entry.perm = IOMMU_NONE;
+ event.entry.translated_addr = 0;
+ memory_region_notify_iommu(&vtd_dev_as->iommu, 0, event);
done:
return true;
@@ -3485,19 +3494,20 @@ static void vtd_address_space_unmap(VTDAddressSpace *as, IOMMUNotifier *n)
size = remain = end - start + 1;
while (remain >= VTD_PAGE_SIZE) {
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
uint64_t mask = get_naturally_aligned_size(start, remain, s->aw_bits);
assert(mask);
- entry.iova = start;
- entry.addr_mask = mask - 1;
- entry.target_as = &address_space_memory;
- entry.perm = IOMMU_NONE;
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.iova = start;
+ event.entry.addr_mask = mask - 1;
+ event.entry.target_as = &address_space_memory;
+ event.entry.perm = IOMMU_NONE;
/* This field is meaningless for unmap */
- entry.translated_addr = 0;
+ event.entry.translated_addr = 0;
- memory_region_notify_iommu_one(n, &entry);
+ memory_region_notify_iommu_one(n, &event);
start += mask;
remain -= mask;
@@ -3533,9 +3543,9 @@ static void vtd_address_space_refresh_all(IntelIOMMUState *s)
vtd_switch_address_space_all(s);
}
-static int vtd_replay_hook(IOMMUTLBEntry *entry, void *private)
+static int vtd_replay_hook(IOMMUTLBEvent *event, void *private)
{
- memory_region_notify_iommu_one((IOMMUNotifier *)private, entry);
+ memory_region_notify_iommu_one(private, event);
return 0;
}
diff --git a/hw/misc/tz-mpc.c b/hw/misc/tz-mpc.c
index 98f151237f..30481e1c90 100644
--- a/hw/misc/tz-mpc.c
+++ b/hw/misc/tz-mpc.c
@@ -82,8 +82,10 @@ static void tz_mpc_iommu_notify(TZMPC *s, uint32_t lutidx,
/* Called when the LUT word at lutidx has changed from oldlut to newlut;
* must call the IOMMU notifiers for the changed blocks.
*/
- IOMMUTLBEntry entry = {
- .addr_mask = s->blocksize - 1,
+ IOMMUTLBEvent event = {
+ .entry = {
+ .addr_mask = s->blocksize - 1,
+ }
};
hwaddr addr = lutidx * s->blocksize * 32;
int i;
@@ -100,26 +102,28 @@ static void tz_mpc_iommu_notify(TZMPC *s, uint32_t lutidx,
block_is_ns = newlut & (1 << i);
trace_tz_mpc_iommu_notify(addr);
- entry.iova = addr;
- entry.translated_addr = addr;
+ event.entry.iova = addr;
+ event.entry.translated_addr = addr;
- entry.perm = IOMMU_NONE;
- memory_region_notify_iommu(&s->upstream, IOMMU_IDX_S, entry);
- memory_region_notify_iommu(&s->upstream, IOMMU_IDX_NS, entry);
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.perm = IOMMU_NONE;
+ memory_region_notify_iommu(&s->upstream, IOMMU_IDX_S, event);
+ memory_region_notify_iommu(&s->upstream, IOMMU_IDX_NS, event);
- entry.perm = IOMMU_RW;
+ event.type = IOMMU_NOTIFIER_MAP;
+ event.entry.perm = IOMMU_RW;
if (block_is_ns) {
- entry.target_as = &s->blocked_io_as;
+ event.entry.target_as = &s->blocked_io_as;
} else {
- entry.target_as = &s->downstream_as;
+ event.entry.target_as = &s->downstream_as;
}
- memory_region_notify_iommu(&s->upstream, IOMMU_IDX_S, entry);
+ memory_region_notify_iommu(&s->upstream, IOMMU_IDX_S, event);
if (block_is_ns) {
- entry.target_as = &s->downstream_as;
+ event.entry.target_as = &s->downstream_as;
} else {
- entry.target_as = &s->blocked_io_as;
+ event.entry.target_as = &s->blocked_io_as;
}
- memory_region_notify_iommu(&s->upstream, IOMMU_IDX_NS, entry);
+ memory_region_notify_iommu(&s->upstream, IOMMU_IDX_NS, event);
}
}
diff --git a/hw/ppc/spapr_iommu.c b/hw/ppc/spapr_iommu.c
index 0790239ba5..30352df00e 100644
--- a/hw/ppc/spapr_iommu.c
+++ b/hw/ppc/spapr_iommu.c
@@ -445,7 +445,7 @@ static void spapr_tce_reset(DeviceState *dev)
static target_ulong put_tce_emu(SpaprTceTable *tcet, target_ulong ioba,
target_ulong tce)
{
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
hwaddr page_mask = IOMMU_PAGE_MASK(tcet->page_shift);
unsigned long index = (ioba - tcet->bus_offset) >> tcet->page_shift;
@@ -457,12 +457,13 @@ static target_ulong put_tce_emu(SpaprTceTable *tcet, target_ulong ioba,
tcet->table[index] = tce;
- entry.target_as = &address_space_memory,
- entry.iova = (ioba - tcet->bus_offset) & page_mask;
- entry.translated_addr = tce & page_mask;
- entry.addr_mask = ~page_mask;
- entry.perm = spapr_tce_iommu_access_flags(tce);
- memory_region_notify_iommu(&tcet->iommu, 0, entry);
+ event.entry.target_as = &address_space_memory,
+ event.entry.iova = (ioba - tcet->bus_offset) & page_mask;
+ event.entry.translated_addr = tce & page_mask;
+ event.entry.addr_mask = ~page_mask;
+ event.entry.perm = spapr_tce_iommu_access_flags(tce);
+ event.type = event.entry.perm ? IOMMU_NOTIFIER_MAP : IOMMU_NOTIFIER_UNMAP;
+ memory_region_notify_iommu(&tcet->iommu, 0, event);
return H_SUCCESS;
}
diff --git a/hw/s390x/s390-pci-inst.c b/hw/s390x/s390-pci-inst.c
index 70bfd91bf7..d9e1e29f1e 100644
--- a/hw/s390x/s390-pci-inst.c
+++ b/hw/s390x/s390-pci-inst.c
@@ -602,15 +602,18 @@ static uint32_t s390_pci_update_iotlb(S390PCIIOMMU *iommu,
S390IOTLBEntry *entry)
{
S390IOTLBEntry *cache = g_hash_table_lookup(iommu->iotlb, &entry->iova);
- IOMMUTLBEntry notify = {
- .target_as = &address_space_memory,
- .iova = entry->iova,
- .translated_addr = entry->translated_addr,
- .perm = entry->perm,
- .addr_mask = ~PAGE_MASK,
+ IOMMUTLBEvent event = {
+ .type = entry->perm ? IOMMU_NOTIFIER_MAP : IOMMU_NOTIFIER_UNMAP,
+ .entry = {
+ .target_as = &address_space_memory,
+ .iova = entry->iova,
+ .translated_addr = entry->translated_addr,
+ .perm = entry->perm,
+ .addr_mask = ~PAGE_MASK,
+ },
};
- if (entry->perm == IOMMU_NONE) {
+ if (event.type == IOMMU_NOTIFIER_UNMAP) {
if (!cache) {
goto out;
}
@@ -623,9 +626,11 @@ static uint32_t s390_pci_update_iotlb(S390PCIIOMMU *iommu,
goto out;
}
- notify.perm = IOMMU_NONE;
- memory_region_notify_iommu(&iommu->iommu_mr, 0, notify);
- notify.perm = entry->perm;
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.perm = IOMMU_NONE;
+ memory_region_notify_iommu(&iommu->iommu_mr, 0, event);
+ event.type = IOMMU_NOTIFIER_MAP;
+ event.entry.perm = entry->perm;
}
cache = g_new(S390IOTLBEntry, 1);
@@ -637,7 +642,7 @@ static uint32_t s390_pci_update_iotlb(S390PCIIOMMU *iommu,
dec_dma_avail(iommu);
}
- memory_region_notify_iommu(&iommu->iommu_mr, 0, notify);
+ memory_region_notify_iommu(&iommu->iommu_mr, 0, event);
out:
return iommu->dma_limit ? iommu->dma_limit->avail : 1;
diff --git a/hw/virtio/virtio-iommu.c b/hw/virtio/virtio-iommu.c
index fc5c75d693..cea8811295 100644
--- a/hw/virtio/virtio-iommu.c
+++ b/hw/virtio/virtio-iommu.c
@@ -129,7 +129,7 @@ static void virtio_iommu_notify_map(IOMMUMemoryRegion *mr, hwaddr virt_start,
hwaddr virt_end, hwaddr paddr,
uint32_t flags)
{
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
IOMMUAccessFlags perm = IOMMU_ACCESS_FLAG(flags & VIRTIO_IOMMU_MAP_F_READ,
flags & VIRTIO_IOMMU_MAP_F_WRITE);
@@ -141,19 +141,20 @@ static void virtio_iommu_notify_map(IOMMUMemoryRegion *mr, hwaddr virt_start,
trace_virtio_iommu_notify_map(mr->parent_obj.name, virt_start, virt_end,
paddr, perm);
- entry.target_as = &address_space_memory;
- entry.addr_mask = virt_end - virt_start;
- entry.iova = virt_start;
- entry.perm = perm;
- entry.translated_addr = paddr;
+ event.type = IOMMU_NOTIFIER_MAP;
+ event.entry.target_as = &address_space_memory;
+ event.entry.addr_mask = virt_end - virt_start;
+ event.entry.iova = virt_start;
+ event.entry.perm = perm;
+ event.entry.translated_addr = paddr;
- memory_region_notify_iommu(mr, 0, entry);
+ memory_region_notify_iommu(mr, 0, event);
}
static void virtio_iommu_notify_unmap(IOMMUMemoryRegion *mr, hwaddr virt_start,
hwaddr virt_end)
{
- IOMMUTLBEntry entry;
+ IOMMUTLBEvent event;
if (!(mr->iommu_notify_flags & IOMMU_NOTIFIER_UNMAP)) {
return;
@@ -161,13 +162,14 @@ static void virtio_iommu_notify_unmap(IOMMUMemoryRegion *mr, hwaddr virt_start,
trace_virtio_iommu_notify_unmap(mr->parent_obj.name, virt_start, virt_end);
- entry.target_as = &address_space_memory;
- entry.addr_mask = virt_end - virt_start;
- entry.iova = virt_start;
- entry.perm = IOMMU_NONE;
- entry.translated_addr = 0;
+ event.type = IOMMU_NOTIFIER_UNMAP;
+ event.entry.target_as = &address_space_memory;
+ event.entry.addr_mask = virt_end - virt_start;
+ event.entry.iova = virt_start;
+ event.entry.perm = IOMMU_NONE;
+ event.entry.translated_addr = 0;
- memory_region_notify_iommu(mr, 0, entry);
+ memory_region_notify_iommu(mr, 0, event);
}
static gboolean virtio_iommu_notify_unmap_cb(gpointer key, gpointer value,
diff --git a/softmmu/memory.c b/softmmu/memory.c
index 44de610c72..6ca87e8d73 100644
--- a/softmmu/memory.c
+++ b/softmmu/memory.c
@@ -1943,11 +1943,15 @@ void memory_region_unregister_iommu_notifier(MemoryRegion *mr,
}
void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
- IOMMUTLBEntry *entry)
+ IOMMUTLBEvent *event)
{
- IOMMUNotifierFlag request_flags;
+ IOMMUTLBEntry *entry = &event->entry;
hwaddr entry_end = entry->iova + entry->addr_mask;
+ if (event->type == IOMMU_NOTIFIER_UNMAP) {
+ assert(entry->perm == IOMMU_NONE);
+ }
+
/*
* Skip the notification if the notification does not overlap
* with registered range.
@@ -1958,20 +1962,14 @@ void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
assert(entry->iova >= notifier->start && entry_end <= notifier->end);
- if (entry->perm & IOMMU_RW) {
- request_flags = IOMMU_NOTIFIER_MAP;
- } else {
- request_flags = IOMMU_NOTIFIER_UNMAP;
- }
-
- if (notifier->notifier_flags & request_flags) {
+ if (event->type & notifier->notifier_flags) {
notifier->notify(notifier, entry);
}
}
void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
int iommu_idx,
- IOMMUTLBEntry entry)
+ IOMMUTLBEvent event)
{
IOMMUNotifier *iommu_notifier;
@@ -1979,7 +1977,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,
IOMMU_NOTIFIER_FOREACH(iommu_notifier, iommu_mr) {
if (iommu_notifier->iommu_idx == iommu_idx) {
- memory_region_notify_iommu_one(iommu_notifier, &entry);
+ memory_region_notify_iommu_one(iommu_notifier, &event);
}
}
}
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PULL 4/6] memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
` (2 preceding siblings ...)
2020-12-02 11:03 ` [PULL 3/6] memory: Add IOMMUTLBEvent Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 5/6] intel_iommu: Skip page walking on device iotlb invalidations Michael S. Tsirkin
` (2 subsequent siblings)
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Eduardo Habkost, Juan Quintela, Jason Wang,
Richard Henderson, Peter Xu, Eugenio Pérez, Paolo Bonzini
From: Eugenio Pérez <eperezma@redhat.com>
This allows us to differentiate between regular IOMMU map/unmap events
and DEVIOTLB unmap. Doing so, notifiers that only need device IOTLB
invalidations will not receive regular IOMMU unmappings.
Adapt intel and vhost to use it.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-4-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/exec/memory.h | 7 ++++++-
hw/i386/intel_iommu.c | 2 +-
hw/virtio/vhost.c | 2 +-
3 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/include/exec/memory.h b/include/exec/memory.h
index e86b5e92da..521d9901d7 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -97,9 +97,14 @@ typedef enum {
IOMMU_NOTIFIER_UNMAP = 0x1,
/* Notify entry changes (newly created entries) */
IOMMU_NOTIFIER_MAP = 0x2,
+ /* Notify changes on device IOTLB entries */
+ IOMMU_NOTIFIER_DEVIOTLB_UNMAP = 0x04,
} IOMMUNotifierFlag;
-#define IOMMU_NOTIFIER_ALL (IOMMU_NOTIFIER_MAP | IOMMU_NOTIFIER_UNMAP)
+#define IOMMU_NOTIFIER_IOTLB_EVENTS (IOMMU_NOTIFIER_MAP | IOMMU_NOTIFIER_UNMAP)
+#define IOMMU_NOTIFIER_DEVIOTLB_EVENTS IOMMU_NOTIFIER_DEVIOTLB_UNMAP
+#define IOMMU_NOTIFIER_ALL (IOMMU_NOTIFIER_IOTLB_EVENTS | \
+ IOMMU_NOTIFIER_DEVIOTLB_EVENTS)
struct IOMMUNotifier;
typedef void (*IOMMUNotify)(struct IOMMUNotifier *notifier,
diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index 56180b1c43..edc3090f91 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -2468,7 +2468,7 @@ static bool vtd_process_device_iotlb_desc(IntelIOMMUState *s,
sz = VTD_PAGE_SIZE;
}
- event.type = IOMMU_NOTIFIER_UNMAP;
+ event.type = IOMMU_NOTIFIER_DEVIOTLB_UNMAP;
event.entry.target_as = &vtd_dev_as->as;
event.entry.addr_mask = sz - 1;
event.entry.iova = addr;
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index 614ccc2bcb..28c7d78172 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -718,7 +718,7 @@ static void vhost_iommu_region_add(MemoryListener *listener,
iommu_idx = memory_region_iommu_attrs_to_index(iommu_mr,
MEMTXATTRS_UNSPECIFIED);
iommu_notifier_init(&iommu->n, vhost_iommu_unmap_notify,
- IOMMU_NOTIFIER_UNMAP,
+ IOMMU_NOTIFIER_DEVIOTLB_UNMAP,
section->offset_within_region,
int128_get64(end),
iommu_idx);
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PULL 5/6] intel_iommu: Skip page walking on device iotlb invalidations
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
` (3 preceding siblings ...)
2020-12-02 11:03 ` [PULL 4/6] memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 6/6] memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type Michael S. Tsirkin
2020-12-03 10:20 ` [PULL 0/6] pc,vhost: fixes Peter Maydell
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Eduardo Habkost, Jason Wang, Richard Henderson,
Peter Xu, Eugenio Pérez, Paolo Bonzini
From: Eugenio Pérez <eperezma@redhat.com>
Although they didn't reach the notifier because of the filtering in
memory_region_notify_iommu_one, the vt-d was still splitting huge
memory invalidations in chunks. Skipping it.
This improves performance in case of netperf with vhost-net:
* TCP_STREAM: From 1923.6Mbit/s to 2175.13Mbit/s (13%)
* TCP_RR: From 8464.73 trans/s to 8932.703333 trans/s (5.5%)
* UDP_RR: From 8562.08 trans/s to 9005.62/s (5.1%)
* UDP_STREAM: No change observed (insignificant 0.1% improvement)
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-5-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/i386/intel_iommu.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index edc3090f91..0cc71e4057 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -1478,6 +1478,10 @@ static int vtd_sync_shadow_page_table(VTDAddressSpace *vtd_as)
VTDContextEntry ce;
IOMMUNotifier *n;
+ if (!(vtd_as->iommu.iommu_notify_flags & IOMMU_NOTIFIER_IOTLB_EVENTS)) {
+ return 0;
+ }
+
ret = vtd_dev_to_context_entry(vtd_as->iommu_state,
pci_bus_num(vtd_as->bus),
vtd_as->devfn, &ce);
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PULL 6/6] memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
` (4 preceding siblings ...)
2020-12-02 11:03 ` [PULL 5/6] intel_iommu: Skip page walking on device iotlb invalidations Michael S. Tsirkin
@ 2020-12-02 11:03 ` Michael S. Tsirkin
2020-12-03 10:20 ` [PULL 0/6] pc,vhost: fixes Peter Maydell
6 siblings, 0 replies; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-02 11:03 UTC (permalink / raw)
To: qemu-devel
Cc: Peter Maydell, Juan Quintela, Jason Wang, Peter Xu,
Eugenio Pérez, Paolo Bonzini
From: Eugenio Pérez <eperezma@redhat.com>
Device IOTLB invalidations can unmap arbitrary ranges, eiter outside of
the memory region or even [0, ~0ULL] for all the space. The assertion
could be hit by a guest, and rhel7 guest effectively hit it.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-6-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
softmmu/memory.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/softmmu/memory.c b/softmmu/memory.c
index 6ca87e8d73..22bacbbc78 100644
--- a/softmmu/memory.c
+++ b/softmmu/memory.c
@@ -1947,6 +1947,7 @@ void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
{
IOMMUTLBEntry *entry = &event->entry;
hwaddr entry_end = entry->iova + entry->addr_mask;
+ IOMMUTLBEntry tmp = *entry;
if (event->type == IOMMU_NOTIFIER_UNMAP) {
assert(entry->perm == IOMMU_NONE);
@@ -1960,10 +1961,16 @@ void memory_region_notify_iommu_one(IOMMUNotifier *notifier,
return;
}
- assert(entry->iova >= notifier->start && entry_end <= notifier->end);
+ if (notifier->notifier_flags & IOMMU_NOTIFIER_DEVIOTLB_UNMAP) {
+ /* Crop (iova, addr_mask) to range */
+ tmp.iova = MAX(tmp.iova, notifier->start);
+ tmp.addr_mask = MIN(entry_end, notifier->end) - tmp.iova;
+ } else {
+ assert(entry->iova >= notifier->start && entry_end <= notifier->end);
+ }
if (event->type & notifier->notifier_flags) {
- notifier->notify(notifier, entry);
+ notifier->notify(notifier, &tmp);
}
}
--
MST
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PULL 0/6] pc,vhost: fixes
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
` (5 preceding siblings ...)
2020-12-02 11:03 ` [PULL 6/6] memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type Michael S. Tsirkin
@ 2020-12-03 10:20 ` Peter Maydell
2020-12-03 10:58 ` Michael S. Tsirkin
6 siblings, 1 reply; 11+ messages in thread
From: Peter Maydell @ 2020-12-03 10:20 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers
On Wed, 2 Dec 2020 at 11:03, Michael S. Tsirkin <mst@redhat.com> wrote:
>
> Patch 5 gave me pause but we do need patch 6 as
> it's a guest triggerable assert, and it seemed
> cleaner to just take the whole patchset than cherry-pick.
Is this only "fixes a guest triggerable assert"? If so, that's
not sufficiently important to require an rc5 at this point.
thanks
-- PMM
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PULL 0/6] pc,vhost: fixes
2020-12-03 10:20 ` [PULL 0/6] pc,vhost: fixes Peter Maydell
@ 2020-12-03 10:58 ` Michael S. Tsirkin
2020-12-03 11:02 ` Peter Maydell
0 siblings, 1 reply; 11+ messages in thread
From: Michael S. Tsirkin @ 2020-12-03 10:58 UTC (permalink / raw)
To: Peter Maydell; +Cc: QEMU Developers
On Thu, Dec 03, 2020 at 10:20:03AM +0000, Peter Maydell wrote:
> On Wed, 2 Dec 2020 at 11:03, Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > Patch 5 gave me pause but we do need patch 6 as
> > it's a guest triggerable assert, and it seemed
> > cleaner to just take the whole patchset than cherry-pick.
>
> Is this only "fixes a guest triggerable assert"?
My understanding is that without the patches a rhel7 guest triggers
the assert if started with vtd enabled and virtio-net with
iommu_platform=on.
https://bugs.launchpad.net/qemu/+bug/1885175
> If so, that's
> not sufficiently important to require an rc5 at this point.
>
> thanks
> -- PMM
For sure most people don't use vtd ...
--
MST
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PULL 0/6] pc,vhost: fixes
2020-12-03 10:58 ` Michael S. Tsirkin
@ 2020-12-03 11:02 ` Peter Maydell
2020-12-03 11:57 ` Peter Maydell
0 siblings, 1 reply; 11+ messages in thread
From: Peter Maydell @ 2020-12-03 11:02 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers
On Thu, 3 Dec 2020 at 10:59, Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Thu, Dec 03, 2020 at 10:20:03AM +0000, Peter Maydell wrote:
> > On Wed, 2 Dec 2020 at 11:03, Michael S. Tsirkin <mst@redhat.com> wrote:
> > >
> > > Patch 5 gave me pause but we do need patch 6 as
> > > it's a guest triggerable assert, and it seemed
> > > cleaner to just take the whole patchset than cherry-pick.
> >
> > Is this only "fixes a guest triggerable assert"?
>
> My understanding is that without the patches a rhel7 guest triggers
> the assert if started with vtd enabled and virtio-net with
> iommu_platform=on.
>
> https://bugs.launchpad.net/qemu/+bug/1885175
Bug reported in June ? Is this a regression since 5.1, or
was it this way in 5.1 as well?
thanks
-- PMM
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PULL 0/6] pc,vhost: fixes
2020-12-03 11:02 ` Peter Maydell
@ 2020-12-03 11:57 ` Peter Maydell
0 siblings, 0 replies; 11+ messages in thread
From: Peter Maydell @ 2020-12-03 11:57 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: QEMU Developers
On Thu, 3 Dec 2020 at 11:02, Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Thu, 3 Dec 2020 at 10:59, Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > On Thu, Dec 03, 2020 at 10:20:03AM +0000, Peter Maydell wrote:
> > > On Wed, 2 Dec 2020 at 11:03, Michael S. Tsirkin <mst@redhat.com> wrote:
> > > >
> > > > Patch 5 gave me pause but we do need patch 6 as
> > > > it's a guest triggerable assert, and it seemed
> > > > cleaner to just take the whole patchset than cherry-pick.
> > >
> > > Is this only "fixes a guest triggerable assert"?
> >
> > My understanding is that without the patches a rhel7 guest triggers
> > the assert if started with vtd enabled and virtio-net with
> > iommu_platform=on.
> >
> > https://bugs.launchpad.net/qemu/+bug/1885175
>
> Bug reported in June ? Is this a regression since 5.1, or
> was it this way in 5.1 as well?
MST confirmed on IRC that this isn't a regression since 5.1,
so we'll fix it for 6.0.
thanks
-- PMM
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2020-12-03 11:59 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-12-02 11:03 [PULL 0/6] pc,vhost: fixes Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 1/6] vhost-user-scsi: Fix memleaks in vus_proc_req() Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 2/6] memory: Rename memory_region_notify_one to memory_region_notify_iommu_one Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 3/6] memory: Add IOMMUTLBEvent Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 4/6] memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 5/6] intel_iommu: Skip page walking on device iotlb invalidations Michael S. Tsirkin
2020-12-02 11:03 ` [PULL 6/6] memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type Michael S. Tsirkin
2020-12-03 10:20 ` [PULL 0/6] pc,vhost: fixes Peter Maydell
2020-12-03 10:58 ` Michael S. Tsirkin
2020-12-03 11:02 ` Peter Maydell
2020-12-03 11:57 ` Peter Maydell
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).