Re: [Qemu-devel] [PULL v2 0/7] Block patches

[Peter Maydell]

On 9 November 2015 at 17:50, Marc-André Lureau <address@hidden> wrote:
> Hi
>
> ----- Original Message -----
>> On 9 November 2015 at 15:09, Marc-André Lureau <address@hidden> wrote:
>> >> On 9 November 2015 at 12:51, Peter Maydell <address@hidden>
>> >> wrote:
>> >> Marc-André, can you look into why the ivshmem tests might be
>> >> intermittently
>> >> failing like this, please?
>> >
>> > Is this with an slow or emulated host? It could be that the 5s timeout
>> > is not enough?
>>
>> This is with the 32-bit build on a 64-bit ARM server box. So it's
>> not the fastest machine in the world, but it's not bad either.
>> It will be using TCG, obviously.
>>
>> A test which takes 5 seconds to run isn't ideal from a "keep
>> the make-check time down" perspective either.
>
> I can imagine a test starting a server thread and 2 qemu instances would take 
> more than 5s on such configuration then.
>
> Could you try timing the test a few times to confirm this?

address@hidden:~/qemu/build/all-a64$ time
QTEST_QEMU_BINARY=i386-softmmu/qemu-system-i386
QTEST_QEMU_IMG=qemu-img MALLOC_PERTURB_=${MALLOC_PERTURB_:-$((RANDOM %
255 + 1))} gtester -k --verbose -m=quick tests/ivshmem-test
TEST: tests/ivshmem-test... (pid=10893)
  /i386/ivshmem/single:                                                OK
  /i386/ivshmem/pair:                                                  OK
  /i386/ivshmem/server:                                                OK
  /i386/ivshmem/hotplug:                                               OK
  /i386/ivshmem/memdev:                                                OK
PASS: tests/ivshmem-test

real    0m11.945s
user    0m11.020s
sys     0m0.310s

(almost all of the runtime seems to be in the "pair" subtest).

thanks
-- PMM