Re: [Qemu-devel] [PATCH RFC] tests: Run qtest cases in parallel

[Gonglei (Arei)]

> -----Original Message-----
> From: Fam Zheng [mailto:address@hidden
> Sent: Wednesday, September 28, 2016 9:45 AM
> To: Gonglei (Arei)
> Cc: Daniel P. Berrange; address@hidden; John Snow;
> address@hidden; Wubin (H)
> Subject: Re: [Qemu-devel] [PATCH RFC] tests: Run qtest cases in parallel
> 
> On Wed, 09/28 01:31, Gonglei (Arei) wrote:
> > # ./tests/virtio-net-test
> > **
> > ERROR:tests/libqtest.c:561:qtest_get_arch: assertion failed: (qemu != NULL)
> > Aborted (core dumped)
> >
> > # ./tests/virtio-blk-test
> > **
> > ERROR:tests/libqtest.c:561:qtest_get_arch: assertion failed: (qemu != NULL)
> > Aborted (core dumped)
> >
> > But they work after I set the environment variable to specify architecture:
> >
> > #
> QTEST_QEMU_BINARY=x86_64-softmmu/qemu-system-x86_64 ./tests/virtio-n
> et-test
> > /x86_64/virtio/net/pci/basic: OK
> > /x86_64/virtio/net/pci/rx_stop_cont: OK
> > /x86_64/virtio/net/pci/hotplug: OK
> >
> > #
> QTEST_QEMU_BINARY=x86_64-softmmu/qemu-system-x86_64 ./tests/virtio-b
> lk-test
> > /x86_64/virtio/blk/pci/basic: OK
> > /x86_64/virtio/blk/pci/indirect: OK
> > /x86_64/virtio/blk/pci/config: OK
> > /x86_64/virtio/blk/pci/msix: OK
> > /x86_64/virtio/blk/pci/idx: OK
> > /x86_64/virtio/blk/pci/hotplug: OK
> >
> > So, Maybe we should add check if the environment relied on is set
> > before executing specific operations in this kind of tests. Right?
> 
> Or make a guess based on $(realpath $0]) (in this case, print the found path 
> to
> avoid testing against wrong binary by mistake).
> 
What do you mean about the realpath? Can we use it to set the environment 
variable?

Regards,
-Gonglei