[Qemu-devel] [PULL 00/10] arm-devs queue

[Qemu-devel] [PULL 00/10] arm-devs queue

[Peter Maydell]

Hi; here's the latest arm-devs queue, containing two things:
 * virtio-mmio
 * Soren's patches to add support for initrd with u-boot-header

These are features, but they've both had code on the list for
some time (and in particular since before softfreeze). I don't
intend to submit anything except bugfixes after this pull.

Please pull.
thanks
-- PMM


The following changes since commit 24943978cbe79634a9a8b02a20efb25b29b3ab49:

  boot-order-test: Add tests for Sun4u (2013-07-18 13:27:47 -0500)

are available in the git repository at:

  git://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-arm-devs-20130722

for you to fetch changes up to fd76663e3fe59dc7f7d5c439561b74f472727137:

  hw/arm: Use 'load_ramdisk()' for loading ramdisks w/ U-Boot header (2013-07-22 12:01:37 +0100)

----------------------------------------------------------------
arm-devs queue

----------------------------------------------------------------
Peter Maydell (8):
      device_tree: Add qemu_devtree_setprop_sized_cells() utility functions
      arm/boot: Use qemu_devtree_setprop_sized_cells()
      virtio: Add support for guest setting of queue size
      virtio: Support transports which can specify the vring alignment
      virtio: Implement MMIO based virtio transport
      arm/boot: Allow boards to modify the FDT blob
      vexpress: Make VEDBoardInfo extend arm_boot_info
      vexpress: Add virtio-mmio transports

Soren Brinkmann (2):
      hw/loader: Support ramdisk with u-boot header
      hw/arm: Use 'load_ramdisk()' for loading ramdisks w/ U-Boot header

 device_tree.c                  |   33 ++++
 hw/arm/boot.c                  |   48 +++--
 hw/arm/vexpress.c              |  128 ++++++++++--
 hw/core/loader.c               |   84 +++++---
 hw/virtio/Makefile.objs        |    1 +
 hw/virtio/virtio-mmio.c        |  421 ++++++++++++++++++++++++++++++++++++++++
 hw/virtio/virtio.c             |   40 +++-
 include/hw/arm/arm.h           |    4 +
 include/hw/loader.h            |   13 ++
 include/hw/virtio/virtio-bus.h |    6 +
 include/hw/virtio/virtio.h     |    2 +
 include/sysemu/device_tree.h   |   59 ++++++
 12 files changed, 771 insertions(+), 68 deletions(-)
 create mode 100644 hw/virtio/virtio-mmio.c

[Qemu-devel] [PULL 01/10] device_tree: Add qemu_devtree_setprop_sized_cells() utility functions

[Peter Maydell]

We already have a qemu_devtree_setprop_cells() which sets a dtb
property to an array of cells whose values are specified by varargs.
However for the fairly common case of setting a property to a list
of addresses or of address,size pairs the number of cells used by
each element in the list depends on the parent's #address-cells
and #size-cells properties. To make this easier we provide an analogous
qemu_devtree_setprop_sized_cells() macro which allows the number
of cells used by each element to be specified. This is implemented
using an underlying qemu_devtree_setprop_sized_cells_from_array()
function which takes the values and sizes as an array; this may
also be directly useful for cases where the cell contents are
constructed programmatically.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 1373977512-28932-2-git-send-email-peter.maydell@linaro.org
---
 device_tree.c                |   33 +++++++++++++++++++++++
 include/sysemu/device_tree.h |   59 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 92 insertions(+)

diff --git a/device_tree.c b/device_tree.c
index 10cf3d0..ffec99a 100644
--- a/device_tree.c
+++ b/device_tree.c
@@ -308,3 +308,36 @@ void qemu_devtree_dumpdtb(void *fdt, int size)
         exit(g_file_set_contents(dumpdtb, fdt, size, NULL) ? 0 : 1);
     }
 }
+
+int qemu_devtree_setprop_sized_cells_from_array(void *fdt,
+                                                const char *node_path,
+                                                const char *property,
+                                                int numvalues,
+                                                uint64_t *values)
+{
+    uint32_t *propcells;
+    uint64_t value;
+    int cellnum, vnum, ncells;
+    uint32_t hival;
+
+    propcells = g_new0(uint32_t, numvalues * 2);
+
+    cellnum = 0;
+    for (vnum = 0; vnum < numvalues; vnum++) {
+        ncells = values[vnum * 2];
+        if (ncells != 1 && ncells != 2) {
+            return -1;
+        }
+        value = values[vnum * 2 + 1];
+        hival = cpu_to_be32(value >> 32);
+        if (ncells > 1) {
+            propcells[cellnum++] = hival;
+        } else if (hival != 0) {
+            return -1;
+        }
+        propcells[cellnum++] = cpu_to_be32(value);
+    }
+
+    return qemu_devtree_setprop(fdt, node_path, property, propcells,
+                                cellnum * sizeof(uint32_t));
+}
diff --git a/include/sysemu/device_tree.h b/include/sysemu/device_tree.h
index f0b3f35..2b58baf 100644
--- a/include/sysemu/device_tree.h
+++ b/include/sysemu/device_tree.h
@@ -51,4 +51,63 @@ int qemu_devtree_add_subnode(void *fdt, const char *name);
 
 void qemu_devtree_dumpdtb(void *fdt, int size);
 
+/**
+ * qemu_devtree_setprop_sized_cells_from_array:
+ * @fdt: device tree blob
+ * @node_path: node to set property on
+ * @property: property to set
+ * @numvalues: number of values
+ * @values: array of number-of-cells, value pairs
+ *
+ * Set the specified property on the specified node in the device tree
+ * to be an array of cells. The values of the cells are specified via
+ * the values list, which alternates between "number of cells used by
+ * this value" and "value".
+ * number-of-cells must be either 1 or 2 (other values will result in
+ * an error being returned). If a value is too large to fit in the
+ * number of cells specified for it, an error is returned.
+ *
+ * This function is useful because device tree nodes often have cell arrays
+ * which are either lists of addresses or lists of address,size tuples, but
+ * the number of cells used for each element vary depending on the
+ * #address-cells and #size-cells properties of their parent node.
+ * If you know all your cell elements are one cell wide you can use the
+ * simpler qemu_devtree_setprop_cells(). If you're not setting up the
+ * array programmatically, qemu_devtree_setprop_sized_cells may be more
+ * convenient.
+ *
+ * Return value: 0 on success, <0 on error.
+ */
+int qemu_devtree_setprop_sized_cells_from_array(void *fdt,
+                                                const char *node_path,
+                                                const char *property,
+                                                int numvalues,
+                                                uint64_t *values);
+
+/**
+ * qemu_devtree_setprop_sized_cells:
+ * @fdt: device tree blob
+ * @node_path: node to set property on
+ * @property: property to set
+ * @...: list of number-of-cells, value pairs
+ *
+ * Set the specified property on the specified node in the device tree
+ * to be an array of cells. The values of the cells are specified via
+ * the variable arguments, which alternates between "number of cells
+ * used by this value" and "value".
+ *
+ * This is a convenience wrapper for the function
+ * qemu_devtree_setprop_sized_cells_from_array().
+ *
+ * Return value: 0 on success, <0 on error.
+ */
+#define qemu_devtree_setprop_sized_cells(fdt, node_path, property, ...)       \
+    ({                                                                        \
+        uint64_t qdt_tmp[] = { __VA_ARGS__ };                                 \
+        qemu_devtree_setprop_sized_cells_from_array(fdt, node_path,           \
+                                                    property,                 \
+                                                    ARRAY_SIZE(qdt_tmp) / 2,  \
+                                                    qdt_tmp);                 \
+    })
+
 #endif /* __DEVICE_TREE_H__ */
-- 
1.7.9.5

[Qemu-devel] [PULL 02/10] arm/boot: Use qemu_devtree_setprop_sized_cells()

[Peter Maydell]

Replace the opencoded assembly of the reg property array for the
/memory node with a call to qemu_devtree_setprop_sized_cells().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 1373977512-28932-3-git-send-email-peter.maydell@linaro.org
---
 hw/arm/boot.c |   29 ++++++++---------------------
 1 file changed, 8 insertions(+), 21 deletions(-)

diff --git a/hw/arm/boot.c b/hw/arm/boot.c
index a2e4032..1780316 100644
--- a/hw/arm/boot.c
+++ b/hw/arm/boot.c
@@ -227,12 +227,10 @@ static void set_kernel_args_old(const struct arm_boot_info *info)
 
 static int load_dtb(hwaddr addr, const struct arm_boot_info *binfo)
 {
-    uint32_t *mem_reg_property;
-    uint32_t mem_reg_propsize;
     void *fdt = NULL;
     char *filename;
     int size, rc;
-    uint32_t acells, scells, hival;
+    uint32_t acells, scells;
 
     filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, binfo->dtb_filename);
     if (!filename) {
@@ -255,29 +253,18 @@ static int load_dtb(hwaddr addr, const struct arm_boot_info *binfo)
         goto fail;
     }
 
-    mem_reg_propsize = acells + scells;
-    mem_reg_property = g_new0(uint32_t, mem_reg_propsize);
-    mem_reg_property[acells - 1] = cpu_to_be32(binfo->loader_start);
-    hival = cpu_to_be32(binfo->loader_start >> 32);
-    if (acells > 1) {
-        mem_reg_property[acells - 2] = hival;
-    } else if (hival != 0) {
-        fprintf(stderr, "qemu: dtb file not compatible with "
-                "RAM start address > 4GB\n");
-        goto fail;
-    }
-    mem_reg_property[acells + scells - 1] = cpu_to_be32(binfo->ram_size);
-    hival = cpu_to_be32(binfo->ram_size >> 32);
-    if (scells > 1) {
-        mem_reg_property[acells + scells - 2] = hival;
-    } else if (hival != 0) {
+    if (scells < 2 && binfo->ram_size >= (1ULL << 32)) {
+        /* This is user error so deserves a friendlier error message
+         * than the failure of setprop_sized_cells would provide
+         */
         fprintf(stderr, "qemu: dtb file not compatible with "
                 "RAM size > 4GB\n");
         goto fail;
     }
 
-    rc = qemu_devtree_setprop(fdt, "/memory", "reg", mem_reg_property,
-                              mem_reg_propsize * sizeof(uint32_t));
+    rc = qemu_devtree_setprop_sized_cells(fdt, "/memory", "reg",
+                                          acells, binfo->loader_start,
+                                          scells, binfo->ram_size);
     if (rc < 0) {
         fprintf(stderr, "couldn't set /memory/reg\n");
         goto fail;
-- 
1.7.9.5

[Qemu-devel] [PULL 03/10] virtio: Add support for guest setting of queue size

[Peter Maydell]

The MMIO virtio transport spec allows the guest to tell the host how
large the queue size is. Add virtio_queue_set_num() function which
implements this in the QEMU common virtio support code.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373977512-28932-4-git-send-email-peter.maydell@linaro.org
---
 hw/virtio/virtio.c         |    8 ++++++++
 include/hw/virtio/virtio.h |    1 +
 2 files changed, 9 insertions(+)

diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 8176c14..01b05f3 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -667,6 +667,14 @@ hwaddr virtio_queue_get_addr(VirtIODevice *vdev, int n)
     return vdev->vq[n].pa;
 }
 
+void virtio_queue_set_num(VirtIODevice *vdev, int n, int num)
+{
+    if (num <= VIRTQUEUE_MAX_SIZE) {
+        vdev->vq[n].vring.num = num;
+        virtqueue_init(&vdev->vq[n]);
+    }
+}
+
 int virtio_queue_get_num(VirtIODevice *vdev, int n)
 {
     return vdev->vq[n].vring.num;
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 5d1d2be..65f378d 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -200,6 +200,7 @@ void virtio_config_writew(VirtIODevice *vdev, uint32_t addr, uint32_t data);
 void virtio_config_writel(VirtIODevice *vdev, uint32_t addr, uint32_t data);
 void virtio_queue_set_addr(VirtIODevice *vdev, int n, hwaddr addr);
 hwaddr virtio_queue_get_addr(VirtIODevice *vdev, int n);
+void virtio_queue_set_num(VirtIODevice *vdev, int n, int num);
 int virtio_queue_get_num(VirtIODevice *vdev, int n);
 void virtio_queue_notify(VirtIODevice *vdev, int n);
 uint16_t virtio_queue_vector(VirtIODevice *vdev, int n);
-- 
1.7.9.5

[Qemu-devel] [PULL 04/10] virtio: Support transports which can specify the vring alignment

[Peter Maydell]

Support virtio transports which can specify the vring alignment
(ie where the guest communicates this to the host) by providing
a new virtio_queue_set_align() function. (The default alignment
remains as before.)

Transports which wish to make use of this must set the
has_variable_vring_alignment field in their VirtioBusClass
struct to true; they can then change the alignment via
virtio_queue_set_align().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373977512-28932-5-git-send-email-peter.maydell@linaro.org
---
 hw/virtio/virtio.c             |   32 +++++++++++++++++++++++++++++---
 include/hw/virtio/virtio-bus.h |    6 ++++++
 include/hw/virtio/virtio.h     |    1 +
 3 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 01b05f3..09f62c6 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -19,8 +19,11 @@
 #include "qemu/atomic.h"
 #include "hw/virtio/virtio-bus.h"
 
-/* The alignment to use between consumer and producer parts of vring.
- * x86 pagesize again. */
+/*
+ * The alignment to use between consumer and producer parts of vring.
+ * x86 pagesize again. This is the default, used by transports like PCI
+ * which don't provide a means for the guest to tell the host the alignment.
+ */
 #define VIRTIO_PCI_VRING_ALIGN         4096
 
 typedef struct VRingDesc
@@ -54,6 +57,7 @@ typedef struct VRingUsed
 typedef struct VRing
 {
     unsigned int num;
+    unsigned int align;
     hwaddr desc;
     hwaddr avail;
     hwaddr used;
@@ -93,7 +97,7 @@ static void virtqueue_init(VirtQueue *vq)
     vq->vring.avail = pa + vq->vring.num * sizeof(VRingDesc);
     vq->vring.used = vring_align(vq->vring.avail +
                                  offsetof(VRingAvail, ring[vq->vring.num]),
-                                 VIRTIO_PCI_VRING_ALIGN);
+                                 vq->vring.align);
 }
 
 static inline uint64_t vring_desc_addr(hwaddr desc_pa, int i)
@@ -687,6 +691,21 @@ int virtio_queue_get_id(VirtQueue *vq)
     return vq - &vdev->vq[0];
 }
 
+void virtio_queue_set_align(VirtIODevice *vdev, int n, int align)
+{
+    BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
+    VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
+
+    /* Check that the transport told us it was going to do this
+     * (so a buggy transport will immediately assert rather than
+     * silently failing to migrate this state)
+     */
+    assert(k->has_variable_vring_alignment);
+
+    vdev->vq[n].vring.align = align;
+    virtqueue_init(&vdev->vq[n]);
+}
+
 void virtio_queue_notify_vq(VirtQueue *vq)
 {
     if (vq->vring.desc) {
@@ -727,6 +746,7 @@ VirtQueue *virtio_add_queue(VirtIODevice *vdev, int queue_size,
         abort();
 
     vdev->vq[i].vring.num = queue_size;
+    vdev->vq[i].vring.align = VIRTIO_PCI_VRING_ALIGN;
     vdev->vq[i].handle_output = handle_output;
 
     return &vdev->vq[i];
@@ -833,6 +853,9 @@ void virtio_save(VirtIODevice *vdev, QEMUFile *f)
             break;
 
         qemu_put_be32(f, vdev->vq[i].vring.num);
+        if (k->has_variable_vring_alignment) {
+            qemu_put_be32(f, vdev->vq[i].vring.align);
+        }
         qemu_put_be64(f, vdev->vq[i].pa);
         qemu_put_be16s(f, &vdev->vq[i].last_avail_idx);
         if (k->save_queue) {
@@ -889,6 +912,9 @@ int virtio_load(VirtIODevice *vdev, QEMUFile *f)
 
     for (i = 0; i < num; i++) {
         vdev->vq[i].vring.num = qemu_get_be32(f);
+        if (k->has_variable_vring_alignment) {
+            vdev->vq[i].vring.align = qemu_get_be32(f);
+        }
         vdev->vq[i].pa = qemu_get_be64(f);
         qemu_get_be16s(f, &vdev->vq[i].last_avail_idx);
         vdev->vq[i].signalled_used_valid = false;
diff --git a/include/hw/virtio/virtio-bus.h b/include/hw/virtio/virtio-bus.h
index 9ed60f9..9217f85 100644
--- a/include/hw/virtio/virtio-bus.h
+++ b/include/hw/virtio/virtio-bus.h
@@ -62,6 +62,12 @@ typedef struct VirtioBusClass {
      * This is called by virtio-bus just before the device is unplugged.
      */
     void (*device_unplug)(DeviceState *d);
+    /*
+     * Does the transport have variable vring alignment?
+     * (ie can it ever call virtio_queue_set_align()?)
+     * Note that changing this will break migration for this transport.
+     */
+    bool has_variable_vring_alignment;
 } VirtioBusClass;
 
 struct VirtioBusState {
diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
index 65f378d..d7e9e0f 100644
--- a/include/hw/virtio/virtio.h
+++ b/include/hw/virtio/virtio.h
@@ -202,6 +202,7 @@ void virtio_queue_set_addr(VirtIODevice *vdev, int n, hwaddr addr);
 hwaddr virtio_queue_get_addr(VirtIODevice *vdev, int n);
 void virtio_queue_set_num(VirtIODevice *vdev, int n, int num);
 int virtio_queue_get_num(VirtIODevice *vdev, int n);
+void virtio_queue_set_align(VirtIODevice *vdev, int n, int align);
 void virtio_queue_notify(VirtIODevice *vdev, int n);
 uint16_t virtio_queue_vector(VirtIODevice *vdev, int n);
 void virtio_queue_set_vector(VirtIODevice *vdev, int n, uint16_t vector);
-- 
1.7.9.5

[Qemu-devel] [PULL 05/10] virtio: Implement MMIO based virtio transport

[Peter Maydell]

Add support for the generic MMIO based virtio transport.

This patch includes some fixes for bugs spotted by
Ying-Shiuan Pan <yspan@itri.org.tw>.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Message-id: 1373977512-28932-6-git-send-email-peter.maydell@linaro.org
[Fred changes: updated to new virtio-bus mechanisms]
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
[PMM changes:
 * fixed trivial makefile conflict
 * removed unused int_enable
 * host_features doesn't need migrating
 * reset guest accessible state in the reset function
 * minor style fixes like extra blank lines
 * RAZ/WI if there's no backend
 * made transport size 0x200, in line with kvmtool
 * set has_variable_vring_alignment
]
---
 hw/virtio/Makefile.objs |    1 +
 hw/virtio/virtio-mmio.c |  421 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 422 insertions(+)
 create mode 100644 hw/virtio/virtio-mmio.c

diff --git a/hw/virtio/Makefile.objs b/hw/virtio/Makefile.objs
index cbe6d51..1ba53d9 100644
--- a/hw/virtio/Makefile.objs
+++ b/hw/virtio/Makefile.objs
@@ -1,6 +1,7 @@
 common-obj-y += virtio-rng.o
 common-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o
 common-obj-y += virtio-bus.o
+common-obj-y += virtio-mmio.o
 common-obj-$(CONFIG_VIRTIO_BLK_DATA_PLANE) += dataplane/
 
 obj-y += virtio.o virtio-balloon.o 
diff --git a/hw/virtio/virtio-mmio.c b/hw/virtio/virtio-mmio.c
new file mode 100644
index 0000000..54d6679
--- /dev/null
+++ b/hw/virtio/virtio-mmio.c
@@ -0,0 +1,421 @@
+/*
+ * Virtio MMIO bindings
+ *
+ * Copyright (c) 2011 Linaro Limited
+ *
+ * Author:
+ *  Peter Maydell <peter.maydell@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "hw/sysbus.h"
+#include "hw/virtio/virtio.h"
+#include "qemu/host-utils.h"
+#include "hw/virtio/virtio-bus.h"
+
+/* #define DEBUG_VIRTIO_MMIO */
+
+#ifdef DEBUG_VIRTIO_MMIO
+
+#define DPRINTF(fmt, ...) \
+do { printf("virtio_mmio: " fmt , ## __VA_ARGS__); } while (0)
+#else
+#define DPRINTF(fmt, ...) do {} while (0)
+#endif
+
+/* QOM macros */
+/* virtio-mmio-bus */
+#define TYPE_VIRTIO_MMIO_BUS "virtio-mmio-bus"
+#define VIRTIO_MMIO_BUS(obj) \
+        OBJECT_CHECK(VirtioBusState, (obj), TYPE_VIRTIO_MMIO_BUS)
+#define VIRTIO_MMIO_BUS_GET_CLASS(obj) \
+        OBJECT_GET_CLASS(VirtioBusClass, (obj), TYPE_VIRTIO_MMIO_BUS)
+#define VIRTIO_MMIO_BUS_CLASS(klass) \
+        OBJECT_CLASS_CHECK(VirtioBusClass, (klass), TYPE_VIRTIO_MMIO_BUS)
+
+/* virtio-mmio */
+#define TYPE_VIRTIO_MMIO "virtio-mmio"
+#define VIRTIO_MMIO(obj) \
+        OBJECT_CHECK(VirtIOMMIOProxy, (obj), TYPE_VIRTIO_MMIO)
+
+/* Memory mapped register offsets */
+#define VIRTIO_MMIO_MAGIC 0x0
+#define VIRTIO_MMIO_VERSION 0x4
+#define VIRTIO_MMIO_DEVICEID 0x8
+#define VIRTIO_MMIO_VENDORID 0xc
+#define VIRTIO_MMIO_HOSTFEATURES 0x10
+#define VIRTIO_MMIO_HOSTFEATURESSEL 0x14
+#define VIRTIO_MMIO_GUESTFEATURES 0x20
+#define VIRTIO_MMIO_GUESTFEATURESSEL 0x24
+#define VIRTIO_MMIO_GUESTPAGESIZE 0x28
+#define VIRTIO_MMIO_QUEUESEL 0x30
+#define VIRTIO_MMIO_QUEUENUMMAX 0x34
+#define VIRTIO_MMIO_QUEUENUM 0x38
+#define VIRTIO_MMIO_QUEUEALIGN 0x3c
+#define VIRTIO_MMIO_QUEUEPFN 0x40
+#define VIRTIO_MMIO_QUEUENOTIFY 0x50
+#define VIRTIO_MMIO_INTERRUPTSTATUS 0x60
+#define VIRTIO_MMIO_INTERRUPTACK 0x64
+#define VIRTIO_MMIO_STATUS 0x70
+/* Device specific config space starts here */
+#define VIRTIO_MMIO_CONFIG 0x100
+
+#define VIRT_MAGIC 0x74726976 /* 'virt' */
+#define VIRT_VERSION 1
+#define VIRT_VENDOR 0x554D4551 /* 'QEMU' */
+
+typedef struct {
+    /* Generic */
+    SysBusDevice parent_obj;
+    MemoryRegion iomem;
+    qemu_irq irq;
+    uint32_t host_features;
+    /* Guest accessible state needing migration and reset */
+    uint32_t host_features_sel;
+    uint32_t guest_features_sel;
+    uint32_t guest_page_shift;
+    /* virtio-bus */
+    VirtioBusState bus;
+} VirtIOMMIOProxy;
+
+static void virtio_mmio_bus_new(VirtioBusState *bus, VirtIOMMIOProxy *dev);
+
+static uint64_t virtio_mmio_read(void *opaque, hwaddr offset, unsigned size)
+{
+    VirtIOMMIOProxy *proxy = (VirtIOMMIOProxy *)opaque;
+    VirtIODevice *vdev = proxy->bus.vdev;
+
+    DPRINTF("virtio_mmio_read offset 0x%x\n", (int)offset);
+
+    if (!vdev) {
+        /* If no backend is present, we treat most registers as
+         * read-as-zero, except for the magic number, version and
+         * vendor ID. This is not strictly sanctioned by the virtio
+         * spec, but it allows us to provide transports with no backend
+         * plugged in which don't confuse Linux's virtio code: the
+         * probe won't complain about the bad magic number, but the
+         * device ID of zero means no backend will claim it.
+         */
+        switch (offset) {
+        case VIRTIO_MMIO_MAGIC:
+            return VIRT_MAGIC;
+        case VIRTIO_MMIO_VERSION:
+            return VIRT_VERSION;
+        case VIRTIO_MMIO_VENDORID:
+            return VIRT_VENDOR;
+        default:
+            return 0;
+        }
+    }
+
+    if (offset >= VIRTIO_MMIO_CONFIG) {
+        offset -= VIRTIO_MMIO_CONFIG;
+        switch (size) {
+        case 1:
+            return virtio_config_readb(vdev, offset);
+        case 2:
+            return virtio_config_readw(vdev, offset);
+        case 4:
+            return virtio_config_readl(vdev, offset);
+        default:
+            abort();
+        }
+    }
+    if (size != 4) {
+        DPRINTF("wrong size access to register!\n");
+        return 0;
+    }
+    switch (offset) {
+    case VIRTIO_MMIO_MAGIC:
+        return VIRT_MAGIC;
+    case VIRTIO_MMIO_VERSION:
+        return VIRT_VERSION;
+    case VIRTIO_MMIO_DEVICEID:
+        return vdev->device_id;
+    case VIRTIO_MMIO_VENDORID:
+        return VIRT_VENDOR;
+    case VIRTIO_MMIO_HOSTFEATURES:
+        if (proxy->host_features_sel) {
+            return 0;
+        }
+        return proxy->host_features;
+    case VIRTIO_MMIO_QUEUENUMMAX:
+        return VIRTQUEUE_MAX_SIZE;
+    case VIRTIO_MMIO_QUEUEPFN:
+        return virtio_queue_get_addr(vdev, vdev->queue_sel)
+            >> proxy->guest_page_shift;
+    case VIRTIO_MMIO_INTERRUPTSTATUS:
+        return vdev->isr;
+    case VIRTIO_MMIO_STATUS:
+        return vdev->status;
+    case VIRTIO_MMIO_HOSTFEATURESSEL:
+    case VIRTIO_MMIO_GUESTFEATURES:
+    case VIRTIO_MMIO_GUESTFEATURESSEL:
+    case VIRTIO_MMIO_GUESTPAGESIZE:
+    case VIRTIO_MMIO_QUEUESEL:
+    case VIRTIO_MMIO_QUEUENUM:
+    case VIRTIO_MMIO_QUEUEALIGN:
+    case VIRTIO_MMIO_QUEUENOTIFY:
+    case VIRTIO_MMIO_INTERRUPTACK:
+        DPRINTF("read of write-only register\n");
+        return 0;
+    default:
+        DPRINTF("bad register offset\n");
+        return 0;
+    }
+    return 0;
+}
+
+static void virtio_mmio_write(void *opaque, hwaddr offset, uint64_t value,
+                              unsigned size)
+{
+    VirtIOMMIOProxy *proxy = (VirtIOMMIOProxy *)opaque;
+    VirtIODevice *vdev = proxy->bus.vdev;
+
+    DPRINTF("virtio_mmio_write offset 0x%x value 0x%" PRIx64 "\n",
+            (int)offset, value);
+
+    if (!vdev) {
+        /* If no backend is present, we just make all registers
+         * write-ignored. This allows us to provide transports with
+         * no backend plugged in.
+         */
+        return;
+    }
+
+    if (offset >= VIRTIO_MMIO_CONFIG) {
+        offset -= VIRTIO_MMIO_CONFIG;
+        switch (size) {
+        case 1:
+            virtio_config_writeb(vdev, offset, value);
+            break;
+        case 2:
+            virtio_config_writew(vdev, offset, value);
+            break;
+        case 4:
+            virtio_config_writel(vdev, offset, value);
+            break;
+        default:
+            abort();
+        }
+        return;
+    }
+    if (size != 4) {
+        DPRINTF("wrong size access to register!\n");
+        return;
+    }
+    switch (offset) {
+    case VIRTIO_MMIO_HOSTFEATURESSEL:
+        proxy->host_features_sel = value;
+        break;
+    case VIRTIO_MMIO_GUESTFEATURES:
+        if (!proxy->guest_features_sel) {
+            virtio_set_features(vdev, value);
+        }
+        break;
+    case VIRTIO_MMIO_GUESTFEATURESSEL:
+        proxy->guest_features_sel = value;
+        break;
+    case VIRTIO_MMIO_GUESTPAGESIZE:
+        proxy->guest_page_shift = ctz32(value);
+        if (proxy->guest_page_shift > 31) {
+            proxy->guest_page_shift = 0;
+        }
+        DPRINTF("guest page size %" PRIx64 " shift %d\n", value,
+                proxy->guest_page_shift);
+        break;
+    case VIRTIO_MMIO_QUEUESEL:
+        if (value < VIRTIO_PCI_QUEUE_MAX) {
+            vdev->queue_sel = value;
+        }
+        break;
+    case VIRTIO_MMIO_QUEUENUM:
+        DPRINTF("mmio_queue write %d max %d\n", (int)value, VIRTQUEUE_MAX_SIZE);
+        virtio_queue_set_num(vdev, vdev->queue_sel, value);
+        break;
+    case VIRTIO_MMIO_QUEUEALIGN:
+        virtio_queue_set_align(vdev, vdev->queue_sel, value);
+        break;
+    case VIRTIO_MMIO_QUEUEPFN:
+        if (value == 0) {
+            virtio_reset(vdev);
+        } else {
+            virtio_queue_set_addr(vdev, vdev->queue_sel,
+                                  value << proxy->guest_page_shift);
+        }
+        break;
+    case VIRTIO_MMIO_QUEUENOTIFY:
+        if (value < VIRTIO_PCI_QUEUE_MAX) {
+            virtio_queue_notify(vdev, value);
+        }
+        break;
+    case VIRTIO_MMIO_INTERRUPTACK:
+        vdev->isr &= ~value;
+        virtio_update_irq(vdev);
+        break;
+    case VIRTIO_MMIO_STATUS:
+        virtio_set_status(vdev, value & 0xff);
+        if (vdev->status == 0) {
+            virtio_reset(vdev);
+        }
+        break;
+    case VIRTIO_MMIO_MAGIC:
+    case VIRTIO_MMIO_VERSION:
+    case VIRTIO_MMIO_DEVICEID:
+    case VIRTIO_MMIO_VENDORID:
+    case VIRTIO_MMIO_HOSTFEATURES:
+    case VIRTIO_MMIO_QUEUENUMMAX:
+    case VIRTIO_MMIO_INTERRUPTSTATUS:
+        DPRINTF("write to readonly register\n");
+        break;
+
+    default:
+        DPRINTF("bad register offset\n");
+    }
+}
+
+static const MemoryRegionOps virtio_mem_ops = {
+    .read = virtio_mmio_read,
+    .write = virtio_mmio_write,
+    .endianness = DEVICE_NATIVE_ENDIAN,
+};
+
+static void virtio_mmio_update_irq(DeviceState *opaque, uint16_t vector)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(opaque);
+    int level;
+
+    if (!proxy->bus.vdev) {
+        return;
+    }
+    level = (proxy->bus.vdev->isr != 0);
+    DPRINTF("virtio_mmio setting IRQ %d\n", level);
+    qemu_set_irq(proxy->irq, level);
+}
+
+static unsigned int virtio_mmio_get_features(DeviceState *opaque)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(opaque);
+
+    return proxy->host_features;
+}
+
+static int virtio_mmio_load_config(DeviceState *opaque, QEMUFile *f)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(opaque);
+
+    proxy->host_features_sel = qemu_get_be32(f);
+    proxy->guest_features_sel = qemu_get_be32(f);
+    proxy->guest_page_shift = qemu_get_be32(f);
+    return 0;
+}
+
+static void virtio_mmio_save_config(DeviceState *opaque, QEMUFile *f)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(opaque);
+
+    qemu_put_be32(f, proxy->host_features_sel);
+    qemu_put_be32(f, proxy->guest_features_sel);
+    qemu_put_be32(f, proxy->guest_page_shift);
+}
+
+static void virtio_mmio_reset(DeviceState *d)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
+
+    virtio_bus_reset(&proxy->bus);
+    proxy->host_features_sel = 0;
+    proxy->guest_features_sel = 0;
+    proxy->guest_page_shift = 0;
+}
+
+/* virtio-mmio device */
+
+/* This is called by virtio-bus just after the device is plugged. */
+static void virtio_mmio_device_plugged(DeviceState *opaque)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(opaque);
+
+    proxy->host_features |= (0x1 << VIRTIO_F_NOTIFY_ON_EMPTY);
+    proxy->host_features = virtio_bus_get_vdev_features(&proxy->bus,
+                                                        proxy->host_features);
+}
+
+static void virtio_mmio_realizefn(DeviceState *d, Error **errp)
+{
+    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
+    SysBusDevice *sbd = SYS_BUS_DEVICE(d);
+
+    virtio_mmio_bus_new(&proxy->bus, proxy);
+    sysbus_init_irq(sbd, &proxy->irq);
+    memory_region_init_io(&proxy->iomem, OBJECT(d), &virtio_mem_ops, proxy,
+                          TYPE_VIRTIO_MMIO, 0x200);
+    sysbus_init_mmio(sbd, &proxy->iomem);
+}
+
+static void virtio_mmio_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->realize = virtio_mmio_realizefn;
+    dc->reset = virtio_mmio_reset;
+}
+
+static const TypeInfo virtio_mmio_info = {
+    .name          = TYPE_VIRTIO_MMIO,
+    .parent        = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(VirtIOMMIOProxy),
+    .class_init    = virtio_mmio_class_init,
+};
+
+/* virtio-mmio-bus. */
+
+static void virtio_mmio_bus_new(VirtioBusState *bus, VirtIOMMIOProxy *dev)
+{
+    DeviceState *qdev = DEVICE(dev);
+    BusState *qbus;
+
+    qbus_create_inplace((BusState *)bus, TYPE_VIRTIO_MMIO_BUS, qdev, NULL);
+    qbus = BUS(bus);
+    qbus->allow_hotplug = 0;
+}
+
+static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
+{
+    BusClass *bus_class = BUS_CLASS(klass);
+    VirtioBusClass *k = VIRTIO_BUS_CLASS(klass);
+
+    k->notify = virtio_mmio_update_irq;
+    k->save_config = virtio_mmio_save_config;
+    k->load_config = virtio_mmio_load_config;
+    k->get_features = virtio_mmio_get_features;
+    k->device_plugged = virtio_mmio_device_plugged;
+    k->has_variable_vring_alignment = true;
+    bus_class->max_dev = 1;
+}
+
+static const TypeInfo virtio_mmio_bus_info = {
+    .name          = TYPE_VIRTIO_MMIO_BUS,
+    .parent        = TYPE_VIRTIO_BUS,
+    .instance_size = sizeof(VirtioBusState),
+    .class_init    = virtio_mmio_bus_class_init,
+};
+
+static void virtio_mmio_register_types(void)
+{
+    type_register_static(&virtio_mmio_bus_info);
+    type_register_static(&virtio_mmio_info);
+}
+
+type_init(virtio_mmio_register_types)
-- 
1.7.9.5

[Qemu-devel] [PULL 06/10] arm/boot: Allow boards to modify the FDT blob

[Peter Maydell]

Add a callback hook in arm_boot_info to allow board models to
modify the device tree blob if they need to. (The major expected
use case is to add virtio-mmio nodes for virtio-mmio transports
that exist in QEMU but not in the hardware.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373977512-28932-7-git-send-email-peter.maydell@linaro.org
---
 hw/arm/boot.c        |    5 +++++
 include/hw/arm/arm.h |    4 ++++
 2 files changed, 9 insertions(+)

diff --git a/hw/arm/boot.c b/hw/arm/boot.c
index 1780316..9fbe5d2 100644
--- a/hw/arm/boot.c
+++ b/hw/arm/boot.c
@@ -294,6 +294,11 @@ static int load_dtb(hwaddr addr, const struct arm_boot_info *binfo)
             goto fail;
         }
     }
+
+    if (binfo->modify_dtb) {
+        binfo->modify_dtb(binfo, fdt);
+    }
+
     qemu_devtree_dumpdtb(fdt, size);
 
     cpu_physical_memory_write(addr, fdt, size);
diff --git a/include/hw/arm/arm.h b/include/hw/arm/arm.h
index 7b2b02d..bae87c6 100644
--- a/include/hw/arm/arm.h
+++ b/include/hw/arm/arm.h
@@ -55,6 +55,10 @@ struct arm_boot_info {
                                  const struct arm_boot_info *info);
     void (*secondary_cpu_reset_hook)(ARMCPU *cpu,
                                      const struct arm_boot_info *info);
+    /* if a board needs to be able to modify a device tree provided by
+     * the user it should implement this hook.
+     */
+    void (*modify_dtb)(const struct arm_boot_info *info, void *fdt);
     /* Used internally by arm_boot.c */
     int is_linux;
     hwaddr initrd_start;
-- 
1.7.9.5

[Qemu-devel] [PULL 07/10] vexpress: Make VEDBoardInfo extend arm_boot_info

[Peter Maydell]

Make the VEDBoardInfo struct extend arm_boot_info; this will
allow us to get at the VEDBoardInfo information inside callbacks
from arm/boot code.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373977512-28932-8-git-send-email-peter.maydell@linaro.org
---
 hw/arm/vexpress.c |   31 +++++++++++++++----------------
 1 file changed, 15 insertions(+), 16 deletions(-)

diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c
index 7d1b823..0f15337 100644
--- a/hw/arm/vexpress.c
+++ b/hw/arm/vexpress.c
@@ -36,8 +36,6 @@
 #define VEXPRESS_FLASH_SIZE (64 * 1024 * 1024)
 #define VEXPRESS_FLASH_SECT_SIZE (256 * 1024)
 
-static struct arm_boot_info vexpress_binfo;
-
 /* Address maps for peripherals:
  * the Versatile Express motherboard has two possible maps,
  * the "legacy" one (used for A9) and the "Cortex-A Series"
@@ -153,6 +151,7 @@ typedef void DBoardInitFn(const VEDBoardInfo *daughterboard,
                           qemu_irq *pic);
 
 struct VEDBoardInfo {
+    struct arm_boot_info bootinfo;
     const hwaddr *motherboard_map;
     hwaddr loader_start;
     const hwaddr gic_cpu_if_addr;
@@ -272,7 +271,7 @@ static const uint32_t a9_clocks[] = {
     66670000, /* Test chip reference clock: 66.67MHz */
 };
 
-static const VEDBoardInfo a9_daughterboard = {
+static VEDBoardInfo a9_daughterboard = {
     .motherboard_map = motherboard_legacy_map,
     .loader_start = 0x60000000,
     .gic_cpu_if_addr = 0x1e000100,
@@ -384,7 +383,7 @@ static const uint32_t a15_clocks[] = {
     40000000, /* OSCCLK8: 40MHz : DDR2 PLL reference */
 };
 
-static const VEDBoardInfo a15_daughterboard = {
+static VEDBoardInfo a15_daughterboard = {
     .motherboard_map = motherboard_aseries_map,
     .loader_start = 0x80000000,
     .gic_cpu_if_addr = 0x2c002000,
@@ -396,7 +395,7 @@ static const VEDBoardInfo a15_daughterboard = {
     .init = a15_daughterboard_init,
 };
 
-static void vexpress_common_init(const VEDBoardInfo *daughterboard,
+static void vexpress_common_init(VEDBoardInfo *daughterboard,
                                  QEMUMachineInitArgs *args)
 {
     DeviceState *dev, *sysctl, *pl041;
@@ -524,17 +523,17 @@ static void vexpress_common_init(const VEDBoardInfo *daughterboard,
 
     /* VE_DAPROM: not modelled */
 
-    vexpress_binfo.ram_size = args->ram_size;
-    vexpress_binfo.kernel_filename = args->kernel_filename;
-    vexpress_binfo.kernel_cmdline = args->kernel_cmdline;
-    vexpress_binfo.initrd_filename = args->initrd_filename;
-    vexpress_binfo.nb_cpus = smp_cpus;
-    vexpress_binfo.board_id = VEXPRESS_BOARD_ID;
-    vexpress_binfo.loader_start = daughterboard->loader_start;
-    vexpress_binfo.smp_loader_start = map[VE_SRAM];
-    vexpress_binfo.smp_bootreg_addr = map[VE_SYSREGS] + 0x30;
-    vexpress_binfo.gic_cpu_if_addr = daughterboard->gic_cpu_if_addr;
-    arm_load_kernel(ARM_CPU(first_cpu), &vexpress_binfo);
+    daughterboard->bootinfo.ram_size = args->ram_size;
+    daughterboard->bootinfo.kernel_filename = args->kernel_filename;
+    daughterboard->bootinfo.kernel_cmdline = args->kernel_cmdline;
+    daughterboard->bootinfo.initrd_filename = args->initrd_filename;
+    daughterboard->bootinfo.nb_cpus = smp_cpus;
+    daughterboard->bootinfo.board_id = VEXPRESS_BOARD_ID;
+    daughterboard->bootinfo.loader_start = daughterboard->loader_start;
+    daughterboard->bootinfo.smp_loader_start = map[VE_SRAM];
+    daughterboard->bootinfo.smp_bootreg_addr = map[VE_SYSREGS] + 0x30;
+    daughterboard->bootinfo.gic_cpu_if_addr = daughterboard->gic_cpu_if_addr;
+    arm_load_kernel(ARM_CPU(first_cpu), &daughterboard->bootinfo);
 }
 
 static void vexpress_a9_init(QEMUMachineInitArgs *args)
-- 
1.7.9.5

[Qemu-devel] [PULL 08/10] vexpress: Add virtio-mmio transports

[Peter Maydell]

Add some virtio-mmio transports to the vexpress board model,
together with a modify_dtb hook which adds them to the device
tree so that the kernel will probe for them. We put them
in a reserved area of the address map.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373977512-28932-9-git-send-email-peter.maydell@linaro.org
---
 hw/arm/vexpress.c |   99 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 99 insertions(+)

diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c
index 0f15337..9586e38 100644
--- a/hw/arm/vexpress.c
+++ b/hw/arm/vexpress.c
@@ -31,11 +31,18 @@
 #include "exec/address-spaces.h"
 #include "sysemu/blockdev.h"
 #include "hw/block/flash.h"
+#include "sysemu/device_tree.h"
+#include <libfdt.h>
 
 #define VEXPRESS_BOARD_ID 0x8e0
 #define VEXPRESS_FLASH_SIZE (64 * 1024 * 1024)
 #define VEXPRESS_FLASH_SECT_SIZE (256 * 1024)
 
+/* Number of virtio transports to create (0..8; limited by
+ * number of available IRQ lines).
+ */
+#define NUM_VIRTIO_TRANSPORTS 4
+
 /* Address maps for peripherals:
  * the Versatile Express motherboard has two possible maps,
  * the "legacy" one (used for A9) and the "Cortex-A Series"
@@ -71,6 +78,7 @@ enum {
     VE_ETHERNET,
     VE_USB,
     VE_DAPROM,
+    VE_VIRTIO,
 };
 
 static hwaddr motherboard_legacy_map[] = {
@@ -89,6 +97,7 @@ static hwaddr motherboard_legacy_map[] = {
     [VE_WDT] = 0x1000f000,
     [VE_TIMER01] = 0x10011000,
     [VE_TIMER23] = 0x10012000,
+    [VE_VIRTIO] = 0x10013000,
     [VE_SERIALDVI] = 0x10016000,
     [VE_RTC] = 0x10017000,
     [VE_COMPACTFLASH] = 0x1001a000,
@@ -135,6 +144,7 @@ static hwaddr motherboard_aseries_map[] = {
     [VE_WDT] = 0x1c0f0000,
     [VE_TIMER01] = 0x1c110000,
     [VE_TIMER23] = 0x1c120000,
+    [VE_VIRTIO] = 0x1c130000,
     [VE_SERIALDVI] = 0x1c160000,
     [VE_RTC] = 0x1c170000,
     [VE_COMPACTFLASH] = 0x1c1a0000,
@@ -395,6 +405,85 @@ static VEDBoardInfo a15_daughterboard = {
     .init = a15_daughterboard_init,
 };
 
+static int add_virtio_mmio_node(void *fdt, uint32_t acells, uint32_t scells,
+                                hwaddr addr, hwaddr size, uint32_t intc,
+                                int irq)
+{
+    /* Add a virtio_mmio node to the device tree blob:
+     *   virtio_mmio@ADDRESS {
+     *       compatible = "virtio,mmio";
+     *       reg = <ADDRESS, SIZE>;
+     *       interrupt-parent = <&intc>;
+     *       interrupts = <0, irq, 1>;
+     *   }
+     * (Note that the format of the interrupts property is dependent on the
+     * interrupt controller that interrupt-parent points to; these are for
+     * the ARM GIC and indicate an SPI interrupt, rising-edge-triggered.)
+     */
+    int rc;
+    char *nodename = g_strdup_printf("/virtio_mmio@%" PRIx64, addr);
+
+    rc = qemu_devtree_add_subnode(fdt, nodename);
+    rc |= qemu_devtree_setprop_string(fdt, nodename,
+                                      "compatible", "virtio,mmio");
+    rc |= qemu_devtree_setprop_sized_cells(fdt, nodename, "reg",
+                                           acells, addr, scells, size);
+    qemu_devtree_setprop_cells(fdt, nodename, "interrupt-parent", intc);
+    qemu_devtree_setprop_cells(fdt, nodename, "interrupts", 0, irq, 1);
+    g_free(nodename);
+    if (rc) {
+        return -1;
+    }
+    return 0;
+}
+
+static uint32_t find_int_controller(void *fdt)
+{
+    /* Find the FDT node corresponding to the interrupt controller
+     * for virtio-mmio devices. We do this by scanning the fdt for
+     * a node with the right compatibility, since we know there is
+     * only one GIC on a vexpress board.
+     * We return the phandle of the node, or 0 if none was found.
+     */
+    const char *compat = "arm,cortex-a9-gic";
+    int offset;
+
+    offset = fdt_node_offset_by_compatible(fdt, -1, compat);
+    if (offset >= 0) {
+        return fdt_get_phandle(fdt, offset);
+    }
+    return 0;
+}
+
+static void vexpress_modify_dtb(const struct arm_boot_info *info, void *fdt)
+{
+    uint32_t acells, scells, intc;
+    const VEDBoardInfo *daughterboard = (const VEDBoardInfo *)info;
+
+    acells = qemu_devtree_getprop_cell(fdt, "/", "#address-cells");
+    scells = qemu_devtree_getprop_cell(fdt, "/", "#size-cells");
+    intc = find_int_controller(fdt);
+    if (!intc) {
+        /* Not fatal, we just won't provide virtio. This will
+         * happen with older device tree blobs.
+         */
+        fprintf(stderr, "QEMU: warning: couldn't find interrupt controller in "
+                "dtb; will not include virtio-mmio devices in the dtb.\n");
+    } else {
+        int i;
+        const hwaddr *map = daughterboard->motherboard_map;
+
+        /* We iterate backwards here because adding nodes
+         * to the dtb puts them in last-first.
+         */
+        for (i = NUM_VIRTIO_TRANSPORTS - 1; i >= 0; i--) {
+            add_virtio_mmio_node(fdt, acells, scells,
+                                 map[VE_VIRTIO] + 0x200 * i,
+                                 0x200, intc, 40 + i);
+        }
+    }
+}
+
 static void vexpress_common_init(VEDBoardInfo *daughterboard,
                                  QEMUMachineInitArgs *args)
 {
@@ -523,6 +612,15 @@ static void vexpress_common_init(VEDBoardInfo *daughterboard,
 
     /* VE_DAPROM: not modelled */
 
+    /* Create mmio transports, so the user can create virtio backends
+     * (which will be automatically plugged in to the transports). If
+     * no backend is created the transport will just sit harmlessly idle.
+     */
+    for (i = 0; i < NUM_VIRTIO_TRANSPORTS; i++) {
+        sysbus_create_simple("virtio-mmio", map[VE_VIRTIO] + 0x200 * i,
+                             pic[40 + i]);
+    }
+
     daughterboard->bootinfo.ram_size = args->ram_size;
     daughterboard->bootinfo.kernel_filename = args->kernel_filename;
     daughterboard->bootinfo.kernel_cmdline = args->kernel_cmdline;
@@ -533,6 +631,7 @@ static void vexpress_common_init(VEDBoardInfo *daughterboard,
     daughterboard->bootinfo.smp_loader_start = map[VE_SRAM];
     daughterboard->bootinfo.smp_bootreg_addr = map[VE_SYSREGS] + 0x30;
     daughterboard->bootinfo.gic_cpu_if_addr = daughterboard->gic_cpu_if_addr;
+    daughterboard->bootinfo.modify_dtb = vexpress_modify_dtb;
     arm_load_kernel(ARM_CPU(first_cpu), &daughterboard->bootinfo);
 }
 
-- 
1.7.9.5

[Qemu-devel] [PULL 09/10] hw/loader: Support ramdisk with u-boot header

[Peter Maydell]

From: Soren Brinkmann <soren.brinkmann@xilinx.com>

Introduce 'load_ramdisk()' which can load "normal" ramdisks and ramdisks
with a u-boot header.
To enable this and leverage synergies 'load_uimage()' is refactored to
accomodate this additional use case.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373323202-17083-2-git-send-email-soren.brinkmann@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/core/loader.c    |   84 ++++++++++++++++++++++++++++++++++++---------------
 include/hw/loader.h |   13 ++++++++
 2 files changed, 72 insertions(+), 25 deletions(-)

diff --git a/hw/core/loader.c b/hw/core/loader.c
index 44311ff..c3c28cf 100644
--- a/hw/core/loader.c
+++ b/hw/core/loader.c
@@ -434,15 +434,17 @@ static ssize_t gunzip(void *dst, size_t dstlen, uint8_t *src,
 }
 
 /* Load a U-Boot image.  */
-int load_uimage(const char *filename, hwaddr *ep,
-                hwaddr *loadaddr, int *is_linux)
+static int load_uboot_image(const char *filename, hwaddr *ep, hwaddr *loadaddr,
+                            int *is_linux, uint8_t image_type)
 {
     int fd;
     int size;
+    hwaddr address;
     uboot_image_header_t h;
     uboot_image_header_t *hdr = &h;
     uint8_t *data = NULL;
     int ret = -1;
+    int do_uncompress = 0;
 
     fd = open(filename, O_RDONLY | O_BINARY);
     if (fd < 0)
@@ -457,32 +459,55 @@ int load_uimage(const char *filename, hwaddr *ep,
     if (hdr->ih_magic != IH_MAGIC)
         goto out;
 
-    /* TODO: Implement other image types.  */
-    if (hdr->ih_type != IH_TYPE_KERNEL) {
-        fprintf(stderr, "Can only load u-boot image type \"kernel\"\n");
+    if (hdr->ih_type != image_type) {
+        fprintf(stderr, "Wrong image type %d, expected %d\n", hdr->ih_type,
+                image_type);
         goto out;
     }
 
-    switch (hdr->ih_comp) {
-    case IH_COMP_NONE:
-    case IH_COMP_GZIP:
+    /* TODO: Implement other image types.  */
+    switch (hdr->ih_type) {
+    case IH_TYPE_KERNEL:
+        address = hdr->ih_load;
+        if (loadaddr) {
+            *loadaddr = hdr->ih_load;
+        }
+
+        switch (hdr->ih_comp) {
+        case IH_COMP_NONE:
+            break;
+        case IH_COMP_GZIP:
+            do_uncompress = 1;
+            break;
+        default:
+            fprintf(stderr,
+                    "Unable to load u-boot images with compression type %d\n",
+                    hdr->ih_comp);
+            goto out;
+        }
+
+        if (ep) {
+            *ep = hdr->ih_ep;
+        }
+
+        /* TODO: Check CPU type.  */
+        if (is_linux) {
+            if (hdr->ih_os == IH_OS_LINUX) {
+                *is_linux = 1;
+            } else {
+                *is_linux = 0;
+            }
+        }
+
+        break;
+    case IH_TYPE_RAMDISK:
+        address = *loadaddr;
         break;
     default:
-        fprintf(stderr,
-                "Unable to load u-boot images with compression type %d\n",
-                hdr->ih_comp);
+        fprintf(stderr, "Unsupported u-boot image type %d\n", hdr->ih_type);
         goto out;
     }
 
-    /* TODO: Check CPU type.  */
-    if (is_linux) {
-        if (hdr->ih_os == IH_OS_LINUX)
-            *is_linux = 1;
-        else
-            *is_linux = 0;
-    }
-
-    *ep = hdr->ih_ep;
     data = g_malloc(hdr->ih_size);
 
     if (read(fd, data, hdr->ih_size) != hdr->ih_size) {
@@ -490,7 +515,7 @@ int load_uimage(const char *filename, hwaddr *ep,
         goto out;
     }
 
-    if (hdr->ih_comp == IH_COMP_GZIP) {
+    if (do_uncompress) {
         uint8_t *compressed_data;
         size_t max_bytes;
         ssize_t bytes;
@@ -508,10 +533,7 @@ int load_uimage(const char *filename, hwaddr *ep,
         hdr->ih_size = bytes;
     }
 
-    rom_add_blob_fixed(filename, data, hdr->ih_size, hdr->ih_load);
-
-    if (loadaddr)
-        *loadaddr = hdr->ih_load;
+    rom_add_blob_fixed(filename, data, hdr->ih_size, address);
 
     ret = hdr->ih_size;
 
@@ -522,6 +544,18 @@ out:
     return ret;
 }
 
+int load_uimage(const char *filename, hwaddr *ep, hwaddr *loadaddr,
+                int *is_linux)
+{
+    return load_uboot_image(filename, ep, loadaddr, is_linux, IH_TYPE_KERNEL);
+}
+
+/* Load a ramdisk.  */
+int load_ramdisk(const char *filename, hwaddr addr, uint64_t max_sz)
+{
+    return load_uboot_image(filename, NULL, &addr, NULL, IH_TYPE_RAMDISK);
+}
+
 /*
  * Functions for reboot-persistent memory regions.
  *  - used for vga bios and option roms.
diff --git a/include/hw/loader.h b/include/hw/loader.h
index 15d4cc9..eb9c9a3 100644
--- a/include/hw/loader.h
+++ b/include/hw/loader.h
@@ -17,6 +17,19 @@ int load_aout(const char *filename, hwaddr addr, int max_sz,
 int load_uimage(const char *filename, hwaddr *ep,
                 hwaddr *loadaddr, int *is_linux);
 
+/**
+ * load_ramdisk:
+ * @filename: Path to the ramdisk image
+ * @addr: Memory address to load the ramdisk to
+ * @max_sz: Maximum allowed ramdisk size (for non-u-boot ramdisks)
+ *
+ * Load a ramdisk image with U-Boot header to the specified memory
+ * address.
+ *
+ * Returns the size of the loaded image on success, -1 otherwise.
+ */
+int load_ramdisk(const char *filename, hwaddr addr, uint64_t max_sz);
+
 ssize_t read_targphys(const char *name,
                       int fd, hwaddr dst_addr, size_t nbytes);
 void pstrcpy_targphys(const char *name,
-- 
1.7.9.5

[Qemu-devel] [PULL 10/10] hw/arm: Use 'load_ramdisk()' for loading ramdisks w/ U-Boot header

[Peter Maydell]

From: Soren Brinkmann <soren.brinkmann@xilinx.com>

The load_ramdisk function is used to load ramdisk featuring a U-Boot
header.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1373323202-17083-3-git-send-email-soren.brinkmann@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/arm/boot.c |   14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/hw/arm/boot.c b/hw/arm/boot.c
index 9fbe5d2..2cbeefd 100644
--- a/hw/arm/boot.c
+++ b/hw/arm/boot.c
@@ -411,10 +411,16 @@ void arm_load_kernel(ARMCPU *cpu, struct arm_boot_info *info)
     info->entry = entry;
     if (is_linux) {
         if (info->initrd_filename) {
-            initrd_size = load_image_targphys(info->initrd_filename,
-                                              info->initrd_start,
-                                              info->ram_size -
-                                              info->initrd_start);
+            initrd_size = load_ramdisk(info->initrd_filename,
+                                       info->initrd_start,
+                                       info->ram_size -
+                                       info->initrd_start);
+            if (initrd_size < 0) {
+                initrd_size = load_image_targphys(info->initrd_filename,
+                                                  info->initrd_start,
+                                                  info->ram_size -
+                                                  info->initrd_start);
+            }
             if (initrd_size < 0) {
                 fprintf(stderr, "qemu: could not load initrd '%s'\n",
                         info->initrd_filename);
-- 
1.7.9.5

[Qemu-devel] [PULL 00/10] arm-devs queue

[Peter Maydell]

Usual arm-devs pullreq; this one contains my changes to allow the
vexpress-a15 to accept more than 2GB of RAM, the Exynos I2C
device and a couple of minor bugfix patches.

thanks
-- PMM

The following changes since commit 61dc008f3529fa74a63aad1907438dad857e255a:

  Revert "audio: Make PC speaker audio card available by default" (2012-07-19 18:25:52 -0500)

are available in the git repository at:
  git://git.linaro.org/people/pmaydell/qemu-arm.git arm-devs.for-upstream

Mitsyanko Igor (1):
      exynos4210: add Exynos4210 i2c implementation

Oleg Ogurtsov (2):
      hw/exynos4210_rtc.c: Fix calculating for value of year
      hw/exynos4210_rtc.c: remove unnecessary code

Peter Maydell (7):
      hw/pl011.c: Avoid crash on read when no chr backend present
      hw/arm_boot.c: Make ram_size a uint64_t
      hw/arm_boot.c: Consistently use ram_size from arm_boot_info struct
      hw/arm_boot.c: Check for RAM sizes exceeding ATAGS capacity
      device_tree: Add support for reading device tree properties
      hw/arm_boot.c: Support DTBs which use 64 bit addresses
      hw/vexpress.c: Allow >4GB of RAM for Cortex-A15 daughterboard

 device_tree.c        |   30 +++++
 device_tree.h        |    4 +
 hw/arm-misc.h        |    2 +-
 hw/arm/Makefile.objs |    2 +-
 hw/arm_boot.c        |   46 ++++++-
 hw/exynos4210.c      |   27 ++++
 hw/exynos4210.h      |    3 +
 hw/exynos4210_i2c.c  |  334 ++++++++++++++++++++++++++++++++++++++++++++++++++
 hw/exynos4210_rtc.c  |    7 +-
 hw/pl011.c           |    4 +-
 hw/vexpress.c        |   13 ++-
 11 files changed, 456 insertions(+), 16 deletions(-)
 create mode 100644 hw/exynos4210_i2c.c

Re: [Qemu-devel] [PULL 00/10] arm-devs queue

[Blue Swirl]

On Wed, Jul 4, 2012 at 11:14 AM, Peter Maydell <peter.maydell@linaro.org> wrote:
> Usual arm-devs pullreq; this contains some minor exynos fixes
> and the iMX31 model. Please pull.

Thanks, pulled.

>
> thanks
> -- PMM
>
>
> The following changes since commit 71ea2e016131a9fcde6f1ffd3e0e34a64c21f593:
>
>   bsd-user: fix build (2012-06-28 20:28:36 +0000)
>
> are available in the git repository at:
>   git://git.linaro.org/people/pmaydell/qemu-arm.git arm-devs.for-upstream
>
> Evgeny Voevodin (2):
>       hw/exynos4210_pwm.c: Fix STOP status in tick handler.
>       hw/exynos4210.c: Fix misleading initialization of IROM mirror
>
> Oleg Ogurtsov (1):
>       Exynos4: add RTC device
>
> Peter Chubb (5):
>       i.MX: UART support
>       i.MX31: Clock Control Module
>       i.MX31: Timers
>       i.MX31: Interrupt Controller
>       i.MX31: KZM-ARM11-01 evaluation board
>
> Peter Maydell (1):
>       hw/cadence_gem: Make rx_desc_addr and tx_desc_addr uint32_t
>
> Stanislav Vorobiov (1):
>       ARM: hw/exynos4210_mct.c: Fix a bug which hangs Linux kernel.
>
>  MAINTAINERS          |    6 +
>  hw/arm/Makefile.objs |    3 +
>  hw/cadence_gem.c     |    4 +-
>  hw/exynos4210.c      |   10 +-
>  hw/exynos4210_mct.c  |    4 -
>  hw/exynos4210_pwm.c  |    2 +-
>  hw/exynos4210_rtc.c  |  595 +++++++++++++++++++++++++++++++++++++++++++
>  hw/imx.h             |   34 +++
>  hw/imx_avic.c        |  408 +++++++++++++++++++++++++++++
>  hw/imx_ccm.c         |  321 +++++++++++++++++++++++
>  hw/imx_serial.c      |  467 ++++++++++++++++++++++++++++++++++
>  hw/imx_timer.c       |  689 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  hw/kzm.c             |  154 +++++++++++
>  13 files changed, 2689 insertions(+), 8 deletions(-)
>  create mode 100644 hw/exynos4210_rtc.c
>  create mode 100644 hw/imx.h
>  create mode 100644 hw/imx_avic.c
>  create mode 100644 hw/imx_ccm.c
>  create mode 100644 hw/imx_serial.c
>  create mode 100644 hw/imx_timer.c
>  create mode 100644 hw/kzm.c

[Qemu-devel] [PULL 00/10] arm-devs queue

[Peter Maydell]

Usual arm-devs pullreq; this contains some minor exynos fixes
and the iMX31 model. Please pull.

thanks
-- PMM


The following changes since commit 71ea2e016131a9fcde6f1ffd3e0e34a64c21f593:

  bsd-user: fix build (2012-06-28 20:28:36 +0000)

are available in the git repository at:
  git://git.linaro.org/people/pmaydell/qemu-arm.git arm-devs.for-upstream

Evgeny Voevodin (2):
      hw/exynos4210_pwm.c: Fix STOP status in tick handler.
      hw/exynos4210.c: Fix misleading initialization of IROM mirror

Oleg Ogurtsov (1):
      Exynos4: add RTC device

Peter Chubb (5):
      i.MX: UART support
      i.MX31: Clock Control Module
      i.MX31: Timers
      i.MX31: Interrupt Controller
      i.MX31: KZM-ARM11-01 evaluation board

Peter Maydell (1):
      hw/cadence_gem: Make rx_desc_addr and tx_desc_addr uint32_t

Stanislav Vorobiov (1):
      ARM: hw/exynos4210_mct.c: Fix a bug which hangs Linux kernel.

 MAINTAINERS          |    6 +
 hw/arm/Makefile.objs |    3 +
 hw/cadence_gem.c     |    4 +-
 hw/exynos4210.c      |   10 +-
 hw/exynos4210_mct.c  |    4 -
 hw/exynos4210_pwm.c  |    2 +-
 hw/exynos4210_rtc.c  |  595 +++++++++++++++++++++++++++++++++++++++++++
 hw/imx.h             |   34 +++
 hw/imx_avic.c        |  408 +++++++++++++++++++++++++++++
 hw/imx_ccm.c         |  321 +++++++++++++++++++++++
 hw/imx_serial.c      |  467 ++++++++++++++++++++++++++++++++++
 hw/imx_timer.c       |  689 ++++++++++++++++++++++++++++++++++++++++++++++++++
 hw/kzm.c             |  154 +++++++++++
 13 files changed, 2689 insertions(+), 8 deletions(-)
 create mode 100644 hw/exynos4210_rtc.c
 create mode 100644 hw/imx.h
 create mode 100644 hw/imx_avic.c
 create mode 100644 hw/imx_ccm.c
 create mode 100644 hw/imx_serial.c
 create mode 100644 hw/imx_timer.c
 create mode 100644 hw/kzm.c