[Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Philippe Mathieu-Daudé]

[-- Attachment #1: Type: text/plain, Size: 673 bytes --]

Hi,

Travis CI hit this error testing commit 853f8ca13cd6d256:

$ clang --version
Apple LLVM version 8.1.0 (clang-802.0.42)
Target: x86_64-apple-darwin16.7.0

$ ./configure
C compiler        clang
host CPU          x86_64
...
TPM support       yes
libssh2 support   no
TPM passthrough   no
TPM emulator      yes

  GTESTER check-qtest-i386
Unexpected error in qio_channel_socket_readv() at io/channel-socket.c:494:
**
ERROR:tests/tpm-emu.c:27:tpm_emu_test_wait_cond: code should not be reached
make: *** [check-qtest-i386] Error 1
make: *** Waiting for unfinished jobs....

see https://travis-ci.org/qemu/qemu/jobs/376289241

Regards,

Phil.


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Peter Maydell]

On 8 May 2018 at 13:41, Philippe Mathieu-Daudé <f4bug@amsat.org> wrote:
> Hi,
>
> Travis CI hit this error testing commit 853f8ca13cd6d256:
>
> $ clang --version
> Apple LLVM version 8.1.0 (clang-802.0.42)
> Target: x86_64-apple-darwin16.7.0
>
> $ ./configure
> C compiler        clang
> host CPU          x86_64
> ...
> TPM support       yes
> libssh2 support   no
> TPM passthrough   no
> TPM emulator      yes
>
>   GTESTER check-qtest-i386
> Unexpected error in qio_channel_socket_readv() at io/channel-socket.c:494:
> **
> ERROR:tests/tpm-emu.c:27:tpm_emu_test_wait_cond: code should not be reached
> make: *** [check-qtest-i386] Error 1
> make: *** Waiting for unfinished jobs....
>
> see https://travis-ci.org/qemu/qemu/jobs/376289241

I see this on my OSX test box from time to time as well.

thanks
-- PMM

Re: [Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Daniel P. Berrangé]

On Tue, May 08, 2018 at 09:41:05AM -0300, Philippe Mathieu-Daudé wrote:
> Hi,
> 
> Travis CI hit this error testing commit 853f8ca13cd6d256:
> 
> $ clang --version
> Apple LLVM version 8.1.0 (clang-802.0.42)
> Target: x86_64-apple-darwin16.7.0
> 
> $ ./configure
> C compiler        clang
> host CPU          x86_64
> ...
> TPM support       yes
> libssh2 support   no
> TPM passthrough   no
> TPM emulator      yes
> 
>   GTESTER check-qtest-i386
> Unexpected error in qio_channel_socket_readv() at io/channel-socket.c:494:
> **
> ERROR:tests/tpm-emu.c:27:tpm_emu_test_wait_cond: code should not be reached
> make: *** [check-qtest-i386] Error 1
> make: *** Waiting for unfinished jobs....

This test has a couple of threads running, and passes &error_abort to
the qio_channel_socket_readv() function. So I expect that one thread
is closing the socket connection, and the other thread has not yet
shutdown cleanly, so tries to readv on a dead socket triggering the
error_abort.  IOW,  qio_channel_socket_readv is likely just the symptom
of brokeness somewhere else in the test, not the cause.


Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: [Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Peter Maydell]

On 8 May 2018 at 14:48, Daniel P. Berrangé <berrange@redhat.com> wrote:
> This test has a couple of threads running, and passes &error_abort to
> the qio_channel_socket_readv() function. So I expect that one thread
> is closing the socket connection, and the other thread has not yet
> shutdown cleanly, so tries to readv on a dead socket triggering the
> error_abort.  IOW,  qio_channel_socket_readv is likely just the symptom
> of brokeness somewhere else in the test, not the cause.

That sounds familiar...didn't we have a thread about this a while back?

thanks
-- PMM

Re: [Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Daniel P. Berrangé]

On Tue, May 08, 2018 at 02:54:39PM +0100, Peter Maydell wrote:
> On 8 May 2018 at 14:48, Daniel P. Berrangé <berrange@redhat.com> wrote:
> > This test has a couple of threads running, and passes &error_abort to
> > the qio_channel_socket_readv() function. So I expect that one thread
> > is closing the socket connection, and the other thread has not yet
> > shutdown cleanly, so tries to readv on a dead socket triggering the
> > error_abort.  IOW,  qio_channel_socket_readv is likely just the symptom
> > of brokeness somewhere else in the test, not the cause.
> 
> That sounds familiar...didn't we have a thread about this a while back?

Yeah, I don't recall if it was related to this particular test, or
another similar situation in a different test.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: [Qemu-devel] tpm_emu_test_wait_cond: code should not be reached

[Stefan Berger]

On 05/08/2018 10:03 AM, Daniel P. Berrangé wrote:
> On Tue, May 08, 2018 at 02:54:39PM +0100, Peter Maydell wrote:
>> On 8 May 2018 at 14:48, Daniel P. Berrangé <berrange@redhat.com> wrote:
>>> This test has a couple of threads running, and passes &error_abort to
>>> the qio_channel_socket_readv() function. So I expect that one thread
>>> is closing the socket connection, and the other thread has not yet
>>> shutdown cleanly, so tries to readv on a dead socket triggering the
>>> error_abort.  IOW,  qio_channel_socket_readv is likely just the symptom
>>> of brokeness somewhere else in the test, not the cause.
>> That sounds familiar...didn't we have a thread about this a while back?
> Yeah, I don't recall if it was related to this particular test, or
> another similar situation in a different test.

My bad. I didn't upstream this one here:

https://lists.gnu.org/archive/html/qemu-devel/2018-03/msg04783.html

I'll send it to qemu-trivial, unless someone objects.

    Stefan

>
> Regards,
> Daniel