[Qemu-devel] [PULL] pci,virtio

[Qemu-devel] [PULL] pci,virtio

[Michael S. Tsirkin]

The following changes since commit 0ffbba357c557d9fa5caf9476878a4b9c155a614:

  migration-exec: fix OpenBSD build warning (2010-06-04 20:01:07 +0000)

are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu.git for_anthony

Amit Shah (1):
      net: Fix hotplug with pci_add

Michael S. Tsirkin (2):
      virtio-net: stop vhost backend on vmstop
      virtio-net: truncating packet

 hw/virtio-net.c |   26 ++++++++++++++------------
 net.c           |    7 +++++--
 2 files changed, 19 insertions(+), 14 deletions(-)

[Qemu-devel] [PULL] pci,virtio

[Michael S. Tsirkin]

Included here is v3 virtio typesafety change - no comments
were made to v3 - I made my best to address all comment
and got no response to v3 so I assume it's OK now.

There are more optimizations in my tree but
they are a bit more scary - I'll let them
stay there as I'll be away for a week.

The following changes since commit 27dd7730582be85c7d4f680f5f71146629809c86:

  Merge remote-tracking branch 'bonzini/header-dirs' into staging (2012-12-19 17:15:39 -0600)

are available in the git repository at:


  git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_anthony

for you to fetch changes up to 89d62be9f4fb538db7f919a2be7df2544ffc02c5:

  virtio-pci: don't poll masked vectors (2012-12-26 11:49:29 +0200)

----------------------------------------------------------------
pci,virtio

This optimizes MSIX handling in virtio-pci.
Also included is pci express capability bugfix.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

----------------------------------------------------------------
Knut Omang (1):
      pcie: Fix bug in pcie_ext_cap_set_next

Michael S. Tsirkin (4):
      virtio: make bindings typesafe
      msi: add API to get notified about pending bit poll
      msix: expose access to masked/pending state
      virtio-pci: don't poll masked vectors

 hw/pci/msix.c        |  19 +++++++--
 hw/pci/msix.h        |   6 ++-
 hw/pci/pci.h         |   4 ++
 hw/pci/pcie.c        |   2 +-
 hw/s390-virtio-bus.c |  24 ++++++++---
 hw/vfio_pci.c        |   2 +-
 hw/virtio-pci.c      | 112 +++++++++++++++++++++++++++++++++++----------------
 hw/virtio.c          |   2 +-
 hw/virtio.h          |  26 ++++++------
 9 files changed, 136 insertions(+), 61 deletions(-)

Re: [Qemu-devel] [PULL] pci,virtio

[Anthony Liguori]

"Michael S. Tsirkin" <mst@redhat.com> writes:

> Included here is v3 virtio typesafety change - no comments
> were made to v3 - I made my best to address all comment
> and got no response to v3 so I assume it's OK now.
>
> There are more optimizations in my tree but
> they are a bit more scary - I'll let them
> stay there as I'll be away for a week.
>
> The following changes since commit 27dd7730582be85c7d4f680f5f71146629809c86:
>
>   Merge remote-tracking branch 'bonzini/header-dirs' into staging (2012-12-19 17:15:39 -0600)
>
> are available in the git repository at:
>
>
>   git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_anthony
>
> for you to fetch changes up to 89d62be9f4fb538db7f919a2be7df2544ffc02c5:
>
>   virtio-pci: don't poll masked vectors (2012-12-26 11:49:29 +0200)
>

Pulled. Thanks.

Regards,

Anthony Liguori

> ----------------------------------------------------------------
> pci,virtio
>
> This optimizes MSIX handling in virtio-pci.
> Also included is pci express capability bugfix.
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
>
> ----------------------------------------------------------------
> Knut Omang (1):
>       pcie: Fix bug in pcie_ext_cap_set_next
>
> Michael S. Tsirkin (4):
>       virtio: make bindings typesafe
>       msi: add API to get notified about pending bit poll
>       msix: expose access to masked/pending state
>       virtio-pci: don't poll masked vectors
>
>  hw/pci/msix.c        |  19 +++++++--
>  hw/pci/msix.h        |   6 ++-
>  hw/pci/pci.h         |   4 ++
>  hw/pci/pcie.c        |   2 +-
>  hw/s390-virtio-bus.c |  24 ++++++++---
>  hw/vfio_pci.c        |   2 +-
>  hw/virtio-pci.c      | 112 +++++++++++++++++++++++++++++++++++----------------
>  hw/virtio.c          |   2 +-
>  hw/virtio.h          |  26 ++++++------
>  9 files changed, 136 insertions(+), 61 deletions(-)

[Qemu-devel] [PULL] pci,virtio

[Michael S. Tsirkin]

The following changes since commit 8e4a424b305e29dc0e454f52df3b35577f342975:

  Revert "virtio-pci: replace byte swap hack" (2013-01-06 18:30:17 +0000)

are available in the git repository at:

  git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_anthony

for you to fetch changes up to feb9a2ab4b0260d8d680a7ffd25063dafc7ec628:

  pci-assign: Enable MSIX on device to match guest (2013-01-09 12:11:16 +0200)

----------------------------------------------------------------
pci,virtio

This further optimizes MSIX handling in virtio-pci.
Also included is pci cleanup by Paolo, and pci device
assignment fix by Alex.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

----------------------------------------------------------------
Alex Williamson (1):
      pci-assign: Enable MSIX on device to match guest

Michael S. Tsirkin (8):
      virtio: don't waste irqfds on control vqs
      msix: add api to access msix message
      kvm: add stub for update msi route
      virtio-pci: cache msix messages
      virtio: backend virtqueue notifier masking
      virtio-net: set/clear vhost_started in reverse order
      vhost: set started flag while start is in progress
      vhost: backend masking support

Paolo Bonzini (5):
      docs: move pci-ids.txt to docs/specs/
      reorganize pci-ids.txt
      virtio-9p: use symbolic constant, add to pci-ids.txt
      ivshmem: use symbolic constant for PCI ID, add to pci-ids.txt
      pci: use constants for devices under the 1B36 device ID, document them

 docs/specs/pci-ids.txt     |  50 +++++++++++
 hw/9pfs/virtio-9p-device.c |   2 +-
 hw/ivshmem.c               |   7 +-
 hw/kvm/pci-assign.c        |  17 +++-
 hw/pci/msix.c              |   2 +-
 hw/pci/msix.h              |   1 +
 hw/pci/pci.h               |   8 ++
 hw/pci_bridge_dev.c        |   8 +-
 hw/serial-pci.c            |  12 +--
 hw/vhost.c                 | 112 +++++++++++++++++++++----
 hw/vhost.h                 |  10 +++
 hw/vhost_net.c             |  27 +++++-
 hw/vhost_net.h             |   3 +
 hw/virtio-net.c            |  22 ++++-
 hw/virtio-pci.c            | 203 +++++++++++++++++++++++++++++++++++++++------
 hw/virtio-pci.h            |   2 +
 hw/virtio.h                |  15 +++-
 kvm-stub.c                 |   5 ++
 pci-ids.txt                |  31 -------
 19 files changed, 437 insertions(+), 100 deletions(-)
 create mode 100644 docs/specs/pci-ids.txt
 delete mode 100644 pci-ids.txt

Re: [Qemu-devel] [PULL] pci,virtio

[Anthony Liguori]

Pulled, thanks.

Regards,

Anthony Liguori

Re: [Qemu-devel] [PULL] pci,virtio

[Andreas Färber]

Am 13.01.2013 11:47, schrieb Michael S. Tsirkin:
> The following changes since commit 8e4a424b305e29dc0e454f52df3b35577f342975:
> 
>   Revert "virtio-pci: replace byte swap hack" (2013-01-06 18:30:17 +0000)
> 
> are available in the git repository at:
> 
>   git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_anthony
> 
> for you to fetch changes up to feb9a2ab4b0260d8d680a7ffd25063dafc7ec628:
> 
>   pci-assign: Enable MSIX on device to match guest (2013-01-09 12:11:16 +0200)
> 
> ----------------------------------------------------------------
> pci,virtio
> 
> This further optimizes MSIX handling in virtio-pci.
> Also included is pci cleanup by Paolo, and pci device
> assignment fix by Alex.
> 
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> 
> ----------------------------------------------------------------
> Alex Williamson (1):
>       pci-assign: Enable MSIX on device to match guest
> 
> Michael S. Tsirkin (8):
>       virtio: don't waste irqfds on control vqs
>       msix: add api to access msix message
>       kvm: add stub for update msi route
>       virtio-pci: cache msix messages

This broke the build with PowerKVM:

  LINK  ppc64-softmmu/qemu-system-ppc64
../hw/virtio-pci.o: In function `kvm_virtio_pci_vq_vector_unmask':
/home/andreas/QEMU/qemu/hw/virtio-pci.c:622: undefined reference to
`kvm_irqchip_update_msi_route'
collect2: error: ld returned 1 exit status
make[1]: *** [qemu-system-ppc64] Fehler 1
make: *** [subdir-ppc64-softmmu] Fehler 2

Can you supply a fix? It's not obvious to me how.

Thanks,
Andreas

>       virtio: backend virtqueue notifier masking
>       virtio-net: set/clear vhost_started in reverse order
>       vhost: set started flag while start is in progress
>       vhost: backend masking support
> 
> Paolo Bonzini (5):
>       docs: move pci-ids.txt to docs/specs/
>       reorganize pci-ids.txt
>       virtio-9p: use symbolic constant, add to pci-ids.txt
>       ivshmem: use symbolic constant for PCI ID, add to pci-ids.txt
>       pci: use constants for devices under the 1B36 device ID, document them
> 
>  docs/specs/pci-ids.txt     |  50 +++++++++++
>  hw/9pfs/virtio-9p-device.c |   2 +-
>  hw/ivshmem.c               |   7 +-
>  hw/kvm/pci-assign.c        |  17 +++-
>  hw/pci/msix.c              |   2 +-
>  hw/pci/msix.h              |   1 +
>  hw/pci/pci.h               |   8 ++
>  hw/pci_bridge_dev.c        |   8 +-
>  hw/serial-pci.c            |  12 +--
>  hw/vhost.c                 | 112 +++++++++++++++++++++----
>  hw/vhost.h                 |  10 +++
>  hw/vhost_net.c             |  27 +++++-
>  hw/vhost_net.h             |   3 +
>  hw/virtio-net.c            |  22 ++++-
>  hw/virtio-pci.c            | 203 +++++++++++++++++++++++++++++++++++++++------
>  hw/virtio-pci.h            |   2 +
>  hw/virtio.h                |  15 +++-
>  kvm-stub.c                 |   5 ++
>  pci-ids.txt                |  31 -------
>  19 files changed, 437 insertions(+), 100 deletions(-)
>  create mode 100644 docs/specs/pci-ids.txt
>  delete mode 100644 pci-ids.txt
> 


-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] [PULL] pci,virtio

[Michael S. Tsirkin]

On Tue, Jan 15, 2013 at 11:09:22AM +0100, Andreas Färber wrote:
> Am 13.01.2013 11:47, schrieb Michael S. Tsirkin:
> > The following changes since commit 8e4a424b305e29dc0e454f52df3b35577f342975:
> > 
> >   Revert "virtio-pci: replace byte swap hack" (2013-01-06 18:30:17 +0000)
> > 
> > are available in the git repository at:
> > 
> >   git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_anthony
> > 
> > for you to fetch changes up to feb9a2ab4b0260d8d680a7ffd25063dafc7ec628:
> > 
> >   pci-assign: Enable MSIX on device to match guest (2013-01-09 12:11:16 +0200)
> > 
> > ----------------------------------------------------------------
> > pci,virtio
> > 
> > This further optimizes MSIX handling in virtio-pci.
> > Also included is pci cleanup by Paolo, and pci device
> > assignment fix by Alex.
> > 
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > 
> > ----------------------------------------------------------------
> > Alex Williamson (1):
> >       pci-assign: Enable MSIX on device to match guest
> > 
> > Michael S. Tsirkin (8):
> >       virtio: don't waste irqfds on control vqs
> >       msix: add api to access msix message
> >       kvm: add stub for update msi route
> >       virtio-pci: cache msix messages
> 
> This broke the build with PowerKVM:
> 
>   LINK  ppc64-softmmu/qemu-system-ppc64
> ../hw/virtio-pci.o: In function `kvm_virtio_pci_vq_vector_unmask':
> /home/andreas/QEMU/qemu/hw/virtio-pci.c:622: undefined reference to
> `kvm_irqchip_update_msi_route'
> collect2: error: ld returned 1 exit status
> make[1]: *** [qemu-system-ppc64] Fehler 1
> make: *** [subdir-ppc64-softmmu] Fehler 2
> 
> Can you supply a fix? It's not obvious to me how.
> 
> Thanks,
> Andreas

I'm not sure how to build this. Can you tell me pls?
Maybe the below helps?

kvm: add stub for kvm_irqchip_update_msi_route

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>


diff --git a/kvm-all.c b/kvm-all.c
index fc0c6e7..bac67da 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -1182,6 +1182,11 @@ static int kvm_irqchip_assign_irqfd(KVMState *s, int fd, int virq, bool assign)
 {
     abort();
 }
+
+int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg)
+{
+    return -ENOSYS;
+}
 #endif /* !KVM_CAP_IRQ_ROUTING */
 
 int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, int virq)

Re: [Qemu-devel] [PULL] pci,virtio

[Andreas Färber]

Am 15.01.2013 17:32, schrieb Michael S. Tsirkin:
> On Tue, Jan 15, 2013 at 11:09:22AM +0100, Andreas Färber wrote:
>> Am 13.01.2013 11:47, schrieb Michael S. Tsirkin:
>>>       virtio-pci: cache msix messages
>>
>> This broke the build with PowerKVM:
>>
>>   LINK  ppc64-softmmu/qemu-system-ppc64
>> ../hw/virtio-pci.o: In function `kvm_virtio_pci_vq_vector_unmask':
>> /home/andreas/QEMU/qemu/hw/virtio-pci.c:622: undefined reference to
>> `kvm_irqchip_update_msi_route'
>> collect2: error: ld returned 1 exit status
>> make[1]: *** [qemu-system-ppc64] Fehler 1
>> make: *** [subdir-ppc64-softmmu] Fehler 2
>>
>> Can you supply a fix? It's not obvious to me how.
> 
> I'm not sure how to build this. Can you tell me pls?

Ideally Daniel would set up your pci branch to build on the revived ppc
buildbot (and on s390, thinking of virtio).

When on a Linux/ppc host the KVM headers are detected, nothing special
needs to be configured to build that code. Cross-compilation should work
as well but obviously needs the right environment; same for a ppc chroot
with linux-user+binfmt.

> Maybe the below helps?
> 
> kvm: add stub for kvm_irqchip_update_msi_route
> 
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Tested-by: Andreas Färber <afaerber@suse.de>

Great, together with Blue's bswap patch this fixes the build for me.
Thanks for the quick response!

Regards,
Andreas

> diff --git a/kvm-all.c b/kvm-all.c
> index fc0c6e7..bac67da 100644
> --- a/kvm-all.c
> +++ b/kvm-all.c
> @@ -1182,6 +1182,11 @@ static int kvm_irqchip_assign_irqfd(KVMState *s, int fd, int virq, bool assign)
>  {
>      abort();
>  }
> +
> +int kvm_irqchip_update_msi_route(KVMState *s, int virq, MSIMessage msg)
> +{
> +    return -ENOSYS;
> +}
>  #endif /* !KVM_CAP_IRQ_ROUTING */
>  
>  int kvm_irqchip_add_irqfd_notifier(KVMState *s, EventNotifier *n, int virq)

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg