[Qemu-commits] [qemu/qemu] 923e2d: vhost: let SET_VRING_ENABLE message depends on pro...

[GitHub]

  Branch: refs/heads/master
  Home:   https://github.com/qemu/qemu
  Commit: 923e2d98ede7404882656aeb4364c3964a95db3d
      
https://github.com/qemu/qemu/commit/923e2d98ede7404882656aeb4364c3964a95db3d
  Author: Yuanhan Liu <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M hw/virtio/vhost-user.c

  Log Message:
  -----------
  vhost: let SET_VRING_ENABLE message depends on protocol feature

But not depend on PROTOCOL_F_MQ feature bit. So that we could use
SET_VRING_ENABLE to sign the backend on stop, even if MQ is disabled.

That's reasonable, since we will have one queue pair at least.

Signed-off-by: Yuanhan Liu <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 12b8cbac3c8243b3dd485aaebb82547aefa06adb
      
https://github.com/qemu/qemu/commit/12b8cbac3c8243b3dd485aaebb82547aefa06adb
  Author: Yuanhan Liu <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M hw/net/vhost_net.c

  Log Message:
  -----------
  vhost: don't send RESET_OWNER at stop

First of all, RESET_OWNER message is sent incorrectly, as it's sent
before GET_VRING_BASE. And the reset message would let the later call
get nothing correct.

And, sending SET_VRING_ENABLE at stop, which has already been done,
makes more sense than RESET_OWNER.

Signed-off-by: Yuanhan Liu <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: a586e65bbd017ab55fe4149dd1bcba5c3a72bcd1
      
https://github.com/qemu/qemu/commit/a586e65bbd017ab55fe4149dd1bcba5c3a72bcd1
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M docs/specs/vhost-user.txt

  Log Message:
  -----------
  vhost-user: update spec description

Clarify logging setup to make sure all clients comply in a way that is
future-proof.  Document how rings are started/stopped.

Signed-off-by: Michael S. Tsirkin <address@hidden>
Reviewed-by: Victor Kaplansky <address@hidden>


  Commit: 87656d50181e1be475303c1b88be6df0963c5bfd
      
https://github.com/qemu/qemu/commit/87656d50181e1be475303c1b88be6df0963c5bfd
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M tests/vhost-user-test.c

  Log Message:
  -----------
  vhost-user-test: support VHOST_USER_SET_VRING_ENABLE

vhost-user-test is broken now: it assumes
QEMU sends RESET_OWNER, and we stopped doing that.
Wait for ENABLE_RING with 0 instead.

Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 5421f318ecc82294ad089fd54924df787b67c971
      
https://github.com/qemu/qemu/commit/5421f318ecc82294ad089fd54924df787b67c971
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M hw/virtio/vhost-user.c

  Log Message:
  -----------
  vhost-user: print original request on error

When we get an unexpected response, print out
the original request.
Helps debug protocol errors tremendously.

Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: dc3db6adde329548771ab2addc2ef8376b2b8b32
      
https://github.com/qemu/qemu/commit/dc3db6adde329548771ab2addc2ef8376b2b8b32
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-16 (Mon, 16 Nov 2015)

  Changed paths:
    M hw/virtio/vhost-user.c

  Log Message:
  -----------
  vhost-user: start/stop all rings

We are currently only sending VRING_ENABLE message for the first ring,
that's wrong: we must start/stop them all.

Reported-by: Victor Kaplansky <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 1f8431f42d833e8914f2d16ce4a49b7b72b90db0
      
https://github.com/qemu/qemu/commit/1f8431f42d833e8914f2d16ce4a49b7b72b90db0
  Author: Bandan Das <address@hidden>
  Date:   2015-11-17 (Tue, 17 Nov 2015)

  Changed paths:
    M hw/core/machine.c
    M hw/pci-host/q35.c
    M include/hw/boards.h

  Log Message:
  -----------
  q35: Check propery to determine if iommu is set

The helper function machine_iommu() isn't necesary. We can
directly check for the property.

Signed-off-by: Bandan Das <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Bandan Das <address@hidden>


  Commit: 8d211f622b11ac2877c344f29de284d5a842d9d7
      
https://github.com/qemu/qemu/commit/8d211f622b11ac2877c344f29de284d5a842d9d7
  Author: Bandan Das <address@hidden>
  Date:   2015-11-17 (Tue, 17 Nov 2015)

  Changed paths:
    M hw/pci-host/piix.c

  Log Message:
  -----------
  i440fx: print an error message if user tries to enable iommu

There's no indication of any sort that i440fx doesn't support
"iommu=on"

Reviewed-by: Eric Blake <address@hidden>
Signed-off-by: Bandan Das <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>
Reviewed-by: Eric Blake <address@hidden>
Signed-off-by: Bandan Das <address@hidden>


  Commit: 5c93c47338dbaa8a21a8ccc9d95dc5ade3f7fa19
      
https://github.com/qemu/qemu/commit/5c93c47338dbaa8a21a8ccc9d95dc5ade3f7fa19
  Author: Victor Kaplansky <address@hidden>
  Date:   2015-11-17 (Tue, 17 Nov 2015)

  Changed paths:
    M tests/vhost-user-bridge.c

  Log Message:
  -----------
  tests/vhost-user-bridge: implement logging of dirty pages

During migration devices continue writing to the guest's memory.
The writes has to be reported to QEMU. This change implements
minimal support in vhost-user-bridge required for successful
migration of a guest with virtio-net device.

Signed-off-by: Victor Kaplansky <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 7ebcfe569211f6ff5402b558b85e2ce1e1066cf6
      
https://github.com/qemu/qemu/commit/7ebcfe569211f6ff5402b558b85e2ce1e1066cf6
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-17 (Tue, 17 Nov 2015)

  Changed paths:
    M docs/specs/vhost-user.txt

  Log Message:
  -----------
  specs/vhost-user: fix spec to match reality

We wanted to start/stop rings on VRING_ENABLE, but that is not what QEMU
does. Rather than tweaking code some more, with risk to stability, let's
just document it as it is.

We'll be  able to fix this in the future with a new protocol feature bit.

Reported-by: Victor Kaplansky <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 72018d1e1917a56d05e24aedc9f582b7c8385e19
      
https://github.com/qemu/qemu/commit/72018d1e1917a56d05e24aedc9f582b7c8385e19
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-18 (Wed, 18 Nov 2015)

  Changed paths:
    M hw/net/vhost_net.c

  Log Message:
  -----------
  vhost-user: ignore qemu-only features

Some features (such as ctrl vq) are supported
by qemu without need to communicate with the
backend.

Drop them from the feature mask so we set them
unconditionally.

Reported-by: Victor Kaplansky <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 48854f57ce3e6aa4bd13368559e5c292e1c44e49
      
https://github.com/qemu/qemu/commit/48854f57ce3e6aa4bd13368559e5c292e1c44e49
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-18 (Wed, 18 Nov 2015)

  Changed paths:
    M hw/virtio/vhost-user.c

  Log Message:
  -----------
  vhost-user: fix log size

commit 2b8819c6eee517c1582983773f8555bb3f9ed645
("vhost-user: modify SET_LOG_BASE to pass mmap size and offset")
passes log size in units of 4 byte chunks instead of the
expected size in bytes.

Fix this up.

Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: d9a3b33d2c9f996537b7f1d0246dee2d0120cefb
      
https://github.com/qemu/qemu/commit/d9a3b33d2c9f996537b7f1d0246dee2d0120cefb
  Author: Michael S. Tsirkin <address@hidden>
  Date:   2015-11-19 (Thu, 19 Nov 2015)

  Changed paths:
    M hw/acpi/core.c

  Log Message:
  -----------
  acpi: fix buffer overrun on migration

ich calls acpi_gpe_init with length ICH9_PMIO_GPE0_LEN so
ICH9_PMIO_GPE0_LEN/2 bytes are allocated, but then the full
ICH9_PMIO_GPE0_LEN bytes are migrated.

As a quick work-around, allocate twice the memory.
We'll probably want to tweak code to avoid
migrating the extra ICH9_PMIO_GPE0_LEN/2 bytes,
but that is a bit trickier to do without breaking
migration compatibility.

Tested-by: "Dr. David Alan Gilbert" <address@hidden>
Reported-by: "Dr. David Alan Gilbert" <address@hidden>
Cc: address@hidden
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 421f4448cec3e42f8477499c5c584699e2cf656b
      
https://github.com/qemu/qemu/commit/421f4448cec3e42f8477499c5c584699e2cf656b
  Author: Marc-André Lureau <address@hidden>
  Date:   2015-11-19 (Thu, 19 Nov 2015)

  Changed paths:
    M configure
    M tests/Makefile

  Log Message:
  -----------
  tests: re-enable vhost-user-test

Commit 7fe34ca9c2e actually disabled vhost-user-test altogether,
since CONFIG_VHOST_NET is a per-target config variable.

tests/vhost-user-test is already x86/x64 softmmu specific test, in order
to enable it correctly, kvm & vhost-net are also conditions. To check
that, set CONFIG_VHOST_NET_TEST_$target when kvm is also enabled.

Since "check-qtest-x86_64-y = $(check-qtest-i386-y)", avoid duplication
when both x86 & x64 are enabled.

Other targets than x86 aren't enabled yet, and is intentionally left as
a future improvement, since I can't easily test those.

Signed-off-by: Marc-André Lureau <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 1c7ba94a184df1eddd589d5400d879568d3e5d08
      
https://github.com/qemu/qemu/commit/1c7ba94a184df1eddd589d5400d879568d3e5d08
  Author: Marc-André Lureau <address@hidden>
  Date:   2015-11-19 (Thu, 19 Nov 2015)

  Changed paths:
    M exec.c
    M vl.c

  Log Message:
  -----------
  exec: silence hugetlbfs warning under qtest

vhost-user-test prints a warning. A test should not need to run on
hugetlbfs, let's silence the warning under qtest. The
condition can't check on qtest_enabled() since vhost-user-test actually
doesn't use qtest accel. However, qtest_driver() can be used, if
qtest_init() is called early enough. For that reason, move chardev and
qtest initialization early.

Signed-off-by: Marc-André Lureau <address@hidden>
Reviewed-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>


  Commit: 348c32709fdbeb475dd072af49523cfdd75873f1
      
https://github.com/qemu/qemu/commit/348c32709fdbeb475dd072af49523cfdd75873f1
  Author: Peter Maydell <address@hidden>
  Date:   2015-11-19 (Thu, 19 Nov 2015)

  Changed paths:
    M configure
    M docs/specs/vhost-user.txt
    M exec.c
    M hw/acpi/core.c
    M hw/core/machine.c
    M hw/net/vhost_net.c
    M hw/pci-host/piix.c
    M hw/pci-host/q35.c
    M hw/virtio/vhost-user.c
    M include/hw/boards.h
    M tests/Makefile
    M tests/vhost-user-bridge.c
    M tests/vhost-user-test.c
    M vl.c

  Log Message:
  -----------
  Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

vhost, pc: fixes for 2.5

Fixes all over the place.

This also re-enables a test we disabled in 2.5 cycle
now that there's a way not to get a warning from it.

Signed-off-by: Michael S. Tsirkin <address@hidden>

# gpg: Signature made Thu 19 Nov 2015 13:27:43 GMT using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <address@hidden>"
# gpg:                 aka "Michael S. Tsirkin <address@hidden>"

* remotes/mst/tags/for_upstream:
  exec: silence hugetlbfs warning under qtest
  tests: re-enable vhost-user-test
  acpi: fix buffer overrun on migration
  vhost-user: fix log size
  vhost-user: ignore qemu-only features
  specs/vhost-user: fix spec to match reality
  tests/vhost-user-bridge: implement logging of dirty pages
  i440fx: print an error message if user tries to enable iommu
  q35: Check propery to determine if iommu is set
  vhost-user: start/stop all rings
  vhost-user: print original request on error
  vhost-user-test: support VHOST_USER_SET_VRING_ENABLE
  vhost-user: update spec description
  vhost: don't send RESET_OWNER at stop
  vhost: let SET_VRING_ENABLE message depends on protocol feature

Signed-off-by: Peter Maydell <address@hidden>


Compare: https://github.com/qemu/qemu/compare/c601a244a49f...348c32709fdb