[Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Michael S. Tsirkin]

Current support for bus master (clearing OK bit) together with the need to
support guests which do not enable PCI bus mastering, leads to extra state in
VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
migration for the case when guests use the device before setting DRIVER_OK.

Rip out this code, and replace it:
-   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
    so just drop it for latest machine type.
-   For compat machine types, set PCI_COMMAND if DRIVER_OK
    is set.

As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
to a new common header.

Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---

Alexander, could you pls ack me merging this?
Thanks!


changes from v2:
    drop default = -1 from ppc - was a typo, reported by Greg

 hw/virtio/virtio-pci.h |  5 +++++
 include/hw/compat.h    | 16 ++++++++++++++++
 include/hw/i386/pc.h   | 10 ++--------
 hw/i386/pc_piix.c      |  2 +-
 hw/i386/pc_q35.c       |  2 +-
 hw/ppc/spapr.c         |  7 +++++++
 hw/virtio/virtio-pci.c | 29 +++++++++++------------------
 7 files changed, 43 insertions(+), 28 deletions(-)
 create mode 100644 include/hw/compat.h

diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h
index 1cea157..8873b6d 100644
--- a/hw/virtio/virtio-pci.h
+++ b/hw/virtio/virtio-pci.h
@@ -53,6 +53,11 @@ typedef struct VirtioBusClass VirtioPCIBusClass;
 #define VIRTIO_PCI_BUS_CLASS(klass) \
         OBJECT_CLASS_CHECK(VirtioPCIBusClass, klass, TYPE_VIRTIO_PCI_BUS)
 
+/* Need to activate work-arounds for buggy guests at vmstate load. */
+#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT  0
+#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION \
+    (1 << VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT)
+
 /* Performance improves when virtqueue kick processing is decoupled from the
  * vcpu thread using ioeventfd for some devices. */
 #define VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT 1
diff --git a/include/hw/compat.h b/include/hw/compat.h
new file mode 100644
index 0000000..47f6ff5
--- /dev/null
+++ b/include/hw/compat.h
@@ -0,0 +1,16 @@
+#ifndef HW_COMPAT_H
+#define HW_COMPAT_H
+
+#define HW_COMPAT_2_1 \
+        {\
+            .driver   = "intel-hda",\
+            .property = "old_msi_addr",\
+            .value    = "on",\
+        },\
+        {\
+            .driver   = "virtio-pci",\
+            .property = "virtio-pci-bus-master-bug-migration",\
+            .value    = "on",\
+        }
+
+#endif /* HW_COMPAT_H */
diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index bae023a..82ad046 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -14,6 +14,7 @@
 #include "sysemu/sysemu.h"
 #include "hw/pci/pci.h"
 #include "hw/boards.h"
+#include "hw/compat.h"
 
 #define HPET_INTCAP "hpet-intcap"
 
@@ -307,15 +308,8 @@ int e820_add_entry(uint64_t, uint64_t, uint32_t);
 int e820_get_num_entries(void);
 bool e820_get_entry(int, uint32_t, uint64_t *, uint64_t *);
 
-#define PC_COMPAT_2_1 \
-        {\
-            .driver   = "intel-hda",\
-            .property = "old_msi_addr",\
-            .value    = "on",\
-        }
-
 #define PC_COMPAT_2_0 \
-        PC_COMPAT_2_1, \
+        HW_COMPAT_2_1, \
         {\
             .driver   = "virtio-scsi-pci",\
             .property = "any_layout",\
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 553afdd..a1634ab 100644
--- a/hw/i386/pc_piix.c
+++ b/hw/i386/pc_piix.c
@@ -502,7 +502,7 @@ static QEMUMachine pc_i440fx_machine_v2_1 = {
     .name = "pc-i440fx-2.1",
     .init = pc_init_pci,
     .compat_props = (GlobalProperty[]) {
-        PC_COMPAT_2_1,
+        HW_COMPAT_2_1,
         { /* end of list */ }
     },
 };
diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
index a199043..f330f7a 100644
--- a/hw/i386/pc_q35.c
+++ b/hw/i386/pc_q35.c
@@ -365,7 +365,7 @@ static QEMUMachine pc_q35_machine_v2_1 = {
     .name = "pc-q35-2.1",
     .init = pc_q35_init,
     .compat_props = (GlobalProperty[]) {
-        PC_COMPAT_2_1,
+        HW_COMPAT_2_1,
         { /* end of list */ }
     },
 };
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 2becc9f..623f626 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -57,6 +57,8 @@
 #include "trace.h"
 #include "hw/nmi.h"
 
+#include "hw/compat.h"
+
 #include <libfdt.h>
 
 /* SLOF memory layout:
@@ -1689,10 +1691,15 @@ static const TypeInfo spapr_machine_info = {
 static void spapr_machine_2_1_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    static GlobalProperty compat_props[] = {
+        HW_COMPAT_2_1,
+        { /* end of list */ }
+    };
 
     mc->name = "pseries-2.1";
     mc->desc = "pSeries Logical Partition (PAPR compliant) v2.1";
     mc->is_default = 0;
+    mc->compat_props = compat_props;
 }
 
 static const TypeInfo spapr_machine_2_1_info = {
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index a827cd4..a499a3c 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -86,9 +86,6 @@
  * 12 is historical, and due to x86 page size. */
 #define VIRTIO_PCI_QUEUE_ADDR_SHIFT    12
 
-/* Flags track per-device state like workarounds for quirks in older guests. */
-#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG  (1 << 0)
-
 static void virtio_pci_bus_new(VirtioBusState *bus, size_t bus_size,
                                VirtIOPCIProxy *dev);
 
@@ -323,14 +320,6 @@ static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
                                      proxy->pci_dev.config[PCI_COMMAND] |
                                      PCI_COMMAND_MASTER, 1);
         }
-
-        /* Linux before 2.6.34 sets the device as OK without enabling
-           the PCI device bus master bit. In this case we need to disable
-           some safety checks. */
-        if ((val & VIRTIO_CONFIG_S_DRIVER_OK) &&
-            !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
-            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
-        }
         break;
     case VIRTIO_MSI_CONFIG_VECTOR:
         msix_vector_unuse(&proxy->pci_dev, vdev->config_vector);
@@ -483,8 +472,7 @@ static void virtio_write_config(PCIDevice *pci_dev, uint32_t address,
     pci_default_write_config(pci_dev, address, val, len);
 
     if (range_covers_byte(address, len, PCI_COMMAND) &&
-        !(pci_dev->config[PCI_COMMAND] & PCI_COMMAND_MASTER) &&
-        !(proxy->flags & VIRTIO_PCI_FLAG_BUS_MASTER_BUG)) {
+        !(pci_dev->config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
         virtio_pci_stop_ioeventfd(proxy);
         virtio_set_status(vdev, vdev->status & ~VIRTIO_CONFIG_S_DRIVER_OK);
     }
@@ -895,11 +883,15 @@ static void virtio_pci_vmstate_change(DeviceState *d, bool running)
     VirtIODevice *vdev = virtio_bus_get_device(&proxy->bus);
 
     if (running) {
-        /* Try to find out if the guest has bus master disabled, but is
-           in ready state. Then we have a buggy guest OS. */
-        if ((vdev->status & VIRTIO_CONFIG_S_DRIVER_OK) &&
+        /* Old QEMU versions did not set bus master enable on status write.
+         * Detect DRIVER set and enable it.
+         */
+        if ((proxy->flags & VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION) &&
+            (vdev->status & VIRTIO_CONFIG_S_DRIVER) &&
             !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
-            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
+            pci_default_write_config(&proxy->pci_dev, PCI_COMMAND,
+                                     proxy->pci_dev.config[PCI_COMMAND] |
+                                     PCI_COMMAND_MASTER, 1);
         }
         virtio_pci_start_ioeventfd(proxy);
     } else {
@@ -1040,10 +1032,11 @@ static void virtio_pci_reset(DeviceState *qdev)
     virtio_pci_stop_ioeventfd(proxy);
     virtio_bus_reset(bus);
     msix_unuse_all_vectors(&proxy->pci_dev);
-    proxy->flags &= ~VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
 }
 
 static Property virtio_pci_properties[] = {
+    DEFINE_PROP_BIT("virtio-pci-bus-master-bug-migration", VirtIOPCIProxy, flags,
+                    VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT, false),
     DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
     DEFINE_PROP_END_OF_LIST(),
 };
-- 
MST

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Greg Kurz]

On Mon, 20 Oct 2014 09:58:50 +0300
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> Current support for bus master (clearing OK bit) together with the need to
> support guests which do not enable PCI bus mastering, leads to extra state in
> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> migration for the case when guests use the device before setting DRIVER_OK.
> 
> Rip out this code, and replace it:
> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
>     so just drop it for latest machine type.
> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
>     is set.
> 
> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> to a new common header.
> 
> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> 
> Alexander, could you pls ack me merging this?
> Thanks!
> 

This patch conflicts with the following commit in Alex's ppc-next branch:

commit 6dd65a940532eaac42a302c30964b0d42967e187
Author: David Gibson <david@gibson.dropbear.id.au>
Date:   Mon Sep 8 15:30:31 2014 +1000

    spapr: Cleanup machine naming conventions, and prepare for 2.2 release

Easy to fix though.

Alex,

I can send the rebased version if you want.

Cheers.

--
Greg

> 
> changes from v2:
>     drop default = -1 from ppc - was a typo, reported by Greg
> 
>  hw/virtio/virtio-pci.h |  5 +++++
>  include/hw/compat.h    | 16 ++++++++++++++++
>  include/hw/i386/pc.h   | 10 ++--------
>  hw/i386/pc_piix.c      |  2 +-
>  hw/i386/pc_q35.c       |  2 +-
>  hw/ppc/spapr.c         |  7 +++++++
>  hw/virtio/virtio-pci.c | 29 +++++++++++------------------
>  7 files changed, 43 insertions(+), 28 deletions(-)
>  create mode 100644 include/hw/compat.h
> 
> diff --git a/hw/virtio/virtio-pci.h b/hw/virtio/virtio-pci.h
> index 1cea157..8873b6d 100644
> --- a/hw/virtio/virtio-pci.h
> +++ b/hw/virtio/virtio-pci.h
> @@ -53,6 +53,11 @@ typedef struct VirtioBusClass VirtioPCIBusClass;
>  #define VIRTIO_PCI_BUS_CLASS(klass) \
>          OBJECT_CLASS_CHECK(VirtioPCIBusClass, klass, TYPE_VIRTIO_PCI_BUS)
> 
> +/* Need to activate work-arounds for buggy guests at vmstate load. */
> +#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT  0
> +#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION \
> +    (1 << VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT)
> +
>  /* Performance improves when virtqueue kick processing is decoupled from the
>   * vcpu thread using ioeventfd for some devices. */
>  #define VIRTIO_PCI_FLAG_USE_IOEVENTFD_BIT 1
> diff --git a/include/hw/compat.h b/include/hw/compat.h
> new file mode 100644
> index 0000000..47f6ff5
> --- /dev/null
> +++ b/include/hw/compat.h
> @@ -0,0 +1,16 @@
> +#ifndef HW_COMPAT_H
> +#define HW_COMPAT_H
> +
> +#define HW_COMPAT_2_1 \
> +        {\
> +            .driver   = "intel-hda",\
> +            .property = "old_msi_addr",\
> +            .value    = "on",\
> +        },\
> +        {\
> +            .driver   = "virtio-pci",\
> +            .property = "virtio-pci-bus-master-bug-migration",\
> +            .value    = "on",\
> +        }
> +
> +#endif /* HW_COMPAT_H */
> diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
> index bae023a..82ad046 100644
> --- a/include/hw/i386/pc.h
> +++ b/include/hw/i386/pc.h
> @@ -14,6 +14,7 @@
>  #include "sysemu/sysemu.h"
>  #include "hw/pci/pci.h"
>  #include "hw/boards.h"
> +#include "hw/compat.h"
> 
>  #define HPET_INTCAP "hpet-intcap"
> 
> @@ -307,15 +308,8 @@ int e820_add_entry(uint64_t, uint64_t, uint32_t);
>  int e820_get_num_entries(void);
>  bool e820_get_entry(int, uint32_t, uint64_t *, uint64_t *);
> 
> -#define PC_COMPAT_2_1 \
> -        {\
> -            .driver   = "intel-hda",\
> -            .property = "old_msi_addr",\
> -            .value    = "on",\
> -        }
> -
>  #define PC_COMPAT_2_0 \
> -        PC_COMPAT_2_1, \
> +        HW_COMPAT_2_1, \
>          {\
>              .driver   = "virtio-scsi-pci",\
>              .property = "any_layout",\
> diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
> index 553afdd..a1634ab 100644
> --- a/hw/i386/pc_piix.c
> +++ b/hw/i386/pc_piix.c
> @@ -502,7 +502,7 @@ static QEMUMachine pc_i440fx_machine_v2_1 = {
>      .name = "pc-i440fx-2.1",
>      .init = pc_init_pci,
>      .compat_props = (GlobalProperty[]) {
> -        PC_COMPAT_2_1,
> +        HW_COMPAT_2_1,
>          { /* end of list */ }
>      },
>  };
> diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
> index a199043..f330f7a 100644
> --- a/hw/i386/pc_q35.c
> +++ b/hw/i386/pc_q35.c
> @@ -365,7 +365,7 @@ static QEMUMachine pc_q35_machine_v2_1 = {
>      .name = "pc-q35-2.1",
>      .init = pc_q35_init,
>      .compat_props = (GlobalProperty[]) {
> -        PC_COMPAT_2_1,
> +        HW_COMPAT_2_1,
>          { /* end of list */ }
>      },
>  };
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index 2becc9f..623f626 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -57,6 +57,8 @@
>  #include "trace.h"
>  #include "hw/nmi.h"
> 
> +#include "hw/compat.h"
> +
>  #include <libfdt.h>
> 
>  /* SLOF memory layout:
> @@ -1689,10 +1691,15 @@ static const TypeInfo spapr_machine_info = {
>  static void spapr_machine_2_1_class_init(ObjectClass *oc, void *data)
>  {
>      MachineClass *mc = MACHINE_CLASS(oc);
> +    static GlobalProperty compat_props[] = {
> +        HW_COMPAT_2_1,
> +        { /* end of list */ }1
> +    };
> 
>      mc->name = "pseries-2.1";
>      mc->desc = "pSeries Logical Partition (PAPR compliant) v2.1";
>      mc->is_default = 0;
> +    mc->compat_props = compat_props;
>  }
> 
>  static const TypeInfo spapr_machine_2_1_info = {
> diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
> index a827cd4..a499a3c 100644
> --- a/hw/virtio/virtio-pci.c
> +++ b/hw/virtio/virtio-pci.c
> @@ -86,9 +86,6 @@
>   * 12 is historical, and due to x86 page size. */
>  #define VIRTIO_PCI_QUEUE_ADDR_SHIFT    12
> 
> -/* Flags track per-device state like workarounds for quirks in older guests. */
> -#define VIRTIO_PCI_FLAG_BUS_MASTER_BUG  (1 << 0)
> -
>  static void virtio_pci_bus_new(VirtioBusState *bus, size_t bus_size,
>                                 VirtIOPCIProxy *dev);
> 
> @@ -323,14 +320,6 @@ static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val)
>                                       proxy->pci_dev.config[PCI_COMMAND] |
>                                       PCI_COMMAND_MASTER, 1);
>          }
> -
> -        /* Linux before 2.6.34 sets the device as OK without enabling
> -           the PCI device bus master bit. In this case we need to disable
> -           some safety checks. */
> -        if ((val & VIRTIO_CONFIG_S_DRIVER_OK) &&
> -            !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
> -            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
> -        }
>          break;
>      case VIRTIO_MSI_CONFIG_VECTOR:
>          msix_vector_unuse(&proxy->pci_dev, vdev->config_vector);
> @@ -483,8 +472,7 @@ static void virtio_write_config(PCIDevice *pci_dev, uint32_t address,
>      pci_default_write_config(pci_dev, address, val, len);
> 
>      if (range_covers_byte(address, len, PCI_COMMAND) &&
> -        !(pci_dev->config[PCI_COMMAND] & PCI_COMMAND_MASTER) &&1
> -        !(proxy->flags & VIRTIO_PCI_FLAG_BUS_MASTER_BUG)) {
> +        !(pci_dev->config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
>          virtio_pci_stop_ioeventfd(proxy);
>          virtio_set_status(vdev, vdev->status & ~VIRTIO_CONFIG_S_DRIVER_OK);
>      }
> @@ -895,11 +883,15 @@ static void virtio_pci_vmstate_change(DeviceState *d, bool running)
>      VirtIODevice *vdev = virtio_bus_get_device(&proxy->bus);
> 
>      if (running) {
> -        /* Try to find out if the guest has bus master disabled, but is
> -           in ready state. Then we have a buggy guest OS. */
> -        if ((vdev->status & VIRTIO_CONFIG_S_DRIVER_OK) &&
> +        /* Old QEMU versions did not set bus master enable on status write.
> +         * Detect DRIVER set and enable it.
> +         */
> +        if ((proxy->flags & VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION) &&
> +            (vdev->status & VIRTIO_CONFIG_S_DRIVER) &&
>              !(proxy->pci_dev.config[PCI_COMMAND] & PCI_COMMAND_MASTER)) {
> -            proxy->flags |= VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
> +            pci_default_write_config(&proxy->pci_dev, PCI_COMMAND,
> +                                     proxy->pci_dev.config[PCI_COMMAND] |
> +                                     PCI_COMMAND_MASTER, 1);
>          }
>          virtio_pci_start_ioeventfd(proxy);
>      } else {
> @@ -1040,10 +1032,11 @@ static void virtio_pci_reset(DeviceState *qdev)
>      virtio_pci_stop_ioeventfd(proxy);
>      virtio_bus_reset(bus);
>      msix_unuse_all_vectors(&proxy->pci_dev);
> -    proxy->flags &= ~VIRTIO_PCI_FLAG_BUS_MASTER_BUG;
>  }
> 
>  static Property virtio_pci_properties[] = {
> +    DEFINE_PROP_BIT("virtio-pci-bus-master-bug-migration", VirtIOPCIProxy, flags,
> +                    VIRTIO_PCI_FLAG_BUS_MASTER_BUG_MIGRATION_BIT, false),
>      DEFINE_VIRTIO_COMMON_FEATURES(VirtIOPCIProxy, host_features),
>      DEFINE_PROP_END_OF_LIST(),
>  };

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Alexander Graf]

On 20.10.14 08:58, Michael S. Tsirkin wrote:
> Current support for bus master (clearing OK bit) together with the need to
> support guests which do not enable PCI bus mastering, leads to extra state in
> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> migration for the case when guests use the device before setting DRIVER_OK.
> 
> Rip out this code, and replace it:
> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
>     so just drop it for latest machine type.
> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
>     is set.
> 
> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> to a new common header.
> 
> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> 
> Alexander, could you pls ack me merging this?
> Thanks!

Have you tried whether this works with old kernel versions? We
introduced the broken flag for very specific old kernel versions that
got their flags wrong IIRC.


Alex

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Greg Kurz]

On Mon, 20 Oct 2014 18:08:56 +0200
Alexander Graf <agraf@suse.de> wrote:

> 
> 
> On 20.10.14 08:58, Michael S. Tsirkin wrote:
> > Current support for bus master (clearing OK bit) together with the need to
> > support guests which do not enable PCI bus mastering, leads to extra state in
> > VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> > migration for the case when guests use the device before setting DRIVER_OK.
> > 
> > Rip out this code, and replace it:
> > -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
> >     so just drop it for latest machine type.
> > -   For compat machine types, set PCI_COMMAND if DRIVER_OK
> >     is set.
> > 
> > As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> > to a new common header.
> > 
> > Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> > 
> > Alexander, could you pls ack me merging this?
> > Thanks!
> 
> Have you tried whether this works with old kernel versions? We
> introduced the broken flag for very specific old kernel versions that
> got their flags wrong IIRC.
> 

This is a follow-up to the following commit:

commit e43c0b2ea5574efb0bedebf6a7d05916eefeba52
Author: Michael S. Tsirkin <mst@redhat.com>
Date:   Thu Sep 11 18:45:33 2014 +0200

    virtio-pci: enable bus master for old guests

The goal was to support *again* buggy guests that do DMA without
enabling bus master, since QEMU now uses the bus master address
space for delivering MSI/MSI-x messages.

We were missing some bits to support cross version migration.
Thanks to Michael's patch, I could migrate a buggy RHEL 6.5 guest
from QEMU v2.1 to QEMU master.

> 
> Alex
> 

Cheers.

--
Greg

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Michael S. Tsirkin]

On Mon, Oct 20, 2014 at 06:08:56PM +0200, Alexander Graf wrote:
> 
> 
> On 20.10.14 08:58, Michael S. Tsirkin wrote:
> > Current support for bus master (clearing OK bit) together with the need to
> > support guests which do not enable PCI bus mastering, leads to extra state in
> > VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> > migration for the case when guests use the device before setting DRIVER_OK.
> > 
> > Rip out this code, and replace it:
> > -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
> >     so just drop it for latest machine type.
> > -   For compat machine types, set PCI_COMMAND if DRIVER_OK
> >     is set.
> > 
> > As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> > to a new common header.
> > 
> > Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> > 
> > Alexander, could you pls ack me merging this?
> > Thanks!
> 
> Have you tried whether this works with old kernel versions? We
> introduced the broken flag for very specific old kernel versions that
> got their flags wrong IIRC.
> 
> 
> Alex

Yes. The point is that since 2.1 the broken flag isn't effective
anymore: when bus master is off, PCI blocks outgoing transactions and
setting internal flags isn't going to help.

Instead, we detect a buggy guest and immediately enable bus master for
it - no need for a separate flag.



-- 
MST

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Alexander Graf]

On 21.10.14 12:16, Michael S. Tsirkin wrote:
> On Mon, Oct 20, 2014 at 06:08:56PM +0200, Alexander Graf wrote:
>>
>>
>> On 20.10.14 08:58, Michael S. Tsirkin wrote:
>>> Current support for bus master (clearing OK bit) together with the need to
>>> support guests which do not enable PCI bus mastering, leads to extra state in
>>> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
>>> migration for the case when guests use the device before setting DRIVER_OK.
>>>
>>> Rip out this code, and replace it:
>>> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
>>>     so just drop it for latest machine type.
>>> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
>>>     is set.
>>>
>>> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
>>> to a new common header.
>>>
>>> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
>>> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>> ---
>>>
>>> Alexander, could you pls ack me merging this?
>>> Thanks!
>>
>> Have you tried whether this works with old kernel versions? We
>> introduced the broken flag for very specific old kernel versions that
>> got their flags wrong IIRC.
>>
>>
>> Alex
> 
> Yes. The point is that since 2.1 the broken flag isn't effective
> anymore: when bus master is off, PCI blocks outgoing transactions and
> setting internal flags isn't going to help.
> 
> Instead, we detect a buggy guest and immediately enable bus master for
> it - no need for a separate flag.

Ok, works for me then.


Alex

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Michael S. Tsirkin]

On Tue, Oct 21, 2014 at 01:35:39PM +0200, Alexander Graf wrote:
> 
> 
> On 21.10.14 12:16, Michael S. Tsirkin wrote:
> > On Mon, Oct 20, 2014 at 06:08:56PM +0200, Alexander Graf wrote:
> >>
> >>
> >> On 20.10.14 08:58, Michael S. Tsirkin wrote:
> >>> Current support for bus master (clearing OK bit) together with the need to
> >>> support guests which do not enable PCI bus mastering, leads to extra state in
> >>> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> >>> migration for the case when guests use the device before setting DRIVER_OK.
> >>>
> >>> Rip out this code, and replace it:
> >>> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
> >>>     so just drop it for latest machine type.
> >>> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
> >>>     is set.
> >>>
> >>> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> >>> to a new common header.
> >>>
> >>> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> >>> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> >>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> >>> ---
> >>>
> >>> Alexander, could you pls ack me merging this?
> >>> Thanks!
> >>
> >> Have you tried whether this works with old kernel versions? We
> >> introduced the broken flag for very specific old kernel versions that
> >> got their flags wrong IIRC.
> >>
> >>
> >> Alex
> > 
> > Yes. The point is that since 2.1 the broken flag isn't effective
> > anymore: when bus master is off, PCI blocks outgoing transactions and
> > setting internal flags isn't going to help.
> > 
> > Instead, we detect a buggy guest and immediately enable bus master for
> > it - no need for a separate flag.
> 
> Ok, works for me then.
> 
> 
> Alex

cool.
Gerd tells me there's a conflict with your next tree:
would you like for me to wait until it's merged,
or go ahead and send pull request and rebase yours on top?

-- 
MST

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Alexander Graf]

On 21.10.14 14:10, Michael S. Tsirkin wrote:
> On Tue, Oct 21, 2014 at 01:35:39PM +0200, Alexander Graf wrote:
>>
>>
>> On 21.10.14 12:16, Michael S. Tsirkin wrote:
>>> On Mon, Oct 20, 2014 at 06:08:56PM +0200, Alexander Graf wrote:
>>>>
>>>>
>>>> On 20.10.14 08:58, Michael S. Tsirkin wrote:
>>>>> Current support for bus master (clearing OK bit) together with the need to
>>>>> support guests which do not enable PCI bus mastering, leads to extra state in
>>>>> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
>>>>> migration for the case when guests use the device before setting DRIVER_OK.
>>>>>
>>>>> Rip out this code, and replace it:
>>>>> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
>>>>>     so just drop it for latest machine type.
>>>>> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
>>>>>     is set.
>>>>>
>>>>> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
>>>>> to a new common header.
>>>>>
>>>>> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
>>>>> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
>>>>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>>>>> ---
>>>>>
>>>>> Alexander, could you pls ack me merging this?
>>>>> Thanks!
>>>>
>>>> Have you tried whether this works with old kernel versions? We
>>>> introduced the broken flag for very specific old kernel versions that
>>>> got their flags wrong IIRC.
>>>>
>>>>
>>>> Alex
>>>
>>> Yes. The point is that since 2.1 the broken flag isn't effective
>>> anymore: when bus master is off, PCI blocks outgoing transactions and
>>> setting internal flags isn't going to help.
>>>
>>> Instead, we detect a buggy guest and immediately enable bus master for
>>> it - no need for a separate flag.
>>
>> Ok, works for me then.
>>
>>
>> Alex
> 
> cool.
> Gerd tells me there's a conflict with your next tree:
> would you like for me to wait until it's merged,
> or go ahead and send pull request and rebase yours on top?

Whoever sends it first wins I'd say ;)


Alex

Re: [Qemu-devel] [PATCH v3] virtio-pci: fix migration for pci bus master

[Greg Kurz]

On Tue, 21 Oct 2014 15:10:22 +0300
"Michael S. Tsirkin" <mst@redhat.com> wrote:

> On Tue, Oct 21, 2014 at 01:35:39PM +0200, Alexander Graf wrote:
> > 
> > 
> > On 21.10.14 12:16, Michael S. Tsirkin wrote:
> > > On Mon, Oct 20, 2014 at 06:08:56PM +0200, Alexander Graf wrote:
> > >>
> > >>
> > >> On 20.10.14 08:58, Michael S. Tsirkin wrote:
> > >>> Current support for bus master (clearing OK bit) together with the need to
> > >>> support guests which do not enable PCI bus mastering, leads to extra state in
> > >>> VIRTIO_PCI_FLAG_BUS_MASTER_BUG bit, which isn't robust in case of cross-version
> > >>> migration for the case when guests use the device before setting DRIVER_OK.
> > >>>
> > >>> Rip out this code, and replace it:
> > >>> -   Modern QEMU doesn't need VIRTIO_PCI_FLAG_BUS_MASTER_BUG
> > >>>     so just drop it for latest machine type.
> > >>> -   For compat machine types, set PCI_COMMAND if DRIVER_OK
> > >>>     is set.
> > >>>
> > >>> As this is needed for 2.1 for both pc and ppc, move PC_COMPAT macros from pc.h
> > >>> to a new common header.
> > >>>
> > >>> Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > >>> Tested-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> > >>> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > >>> ---
> > >>>
> > >>> Alexander, could you pls ack me merging this?
> > >>> Thanks!
> > >>
> > >> Have you tried whether this works with old kernel versions? We
> > >> introduced the broken flag for very specific old kernel versions that
> > >> got their flags wrong IIRC.
> > >>
> > >>
> > >> Alex
> > > 
> > > Yes. The point is that since 2.1 the broken flag isn't effective
> > > anymore: when bus master is off, PCI blocks outgoing transactions and
> > > setting internal flags isn't going to help.
> > > 
> > > Instead, we detect a buggy guest and immediately enable bus master for
> > > it - no need for a separate flag.
> > 
> > Ok, works for me then.
> > 
> > 
> > Alex
> 
> cool.
> Gerd tells me there's a conflict with your next tree:
> would you like for me to wait until it's merged,
> or go ahead and send pull request and rebase yours on top?
> 

This patch also needs to be rebased on current master actually
because PC_COMPAT_2_1 in include/hw/i386/pc.h got changed.

Cheers.

--
Greg