[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] qemu-iotests: move check-block back to Makefiles

From: Lukáš Doktor
Subject: Re: [PATCH] qemu-iotests: move check-block back to Makefiles
Date: Thu, 3 Sep 2020 17:51:56 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.11.0

Dne 02. 09. 20 v 12:21 Thomas Huth napsal(a):
> On 02/09/2020 10.37, Paolo Bonzini wrote:
>> On 02/09/20 10:19, Daniel P. Berrangé wrote:
>>> On Wed, Sep 02, 2020 at 04:00:46AM -0400, Paolo Bonzini wrote:
>>>> check-block has its own test harness, unlike every other test.  If
>>>> we capture its output, as is in general nicer to do without V=1,
>>>> there will be no sign of progress.  So for lack of a better option
>>>> just move the invocation of the test back to Makefile rules.
>>> I expect the correct long term solution here is to stop using the
>>> check-block.sh script.  Instead have code which sets up each
>>> of the I/O tests as an explicit test target in meson. We could
>>> use meson's test grouping features too.
>> I'm not sure, "check-acceptance" will never be integrated in Meson, and
>> it may well be the same for "check-block".  Actually I wonder if Avocado
>> would be a better check-block.sh than check-block.sh.
> Wasn't there even some support for the iotests in avocado (or
> avocado-vt) at one point in time? ... not sure anymore, Cleber, Wainer,
> do you remember?
>  Thomas

Avocado supports so called "external runner", which allows to specify a command 
to be executed and arguments as different variants so it was just a matter of 
coming up with the list of "./check" invocation like "./check -qcow2 -nbd 001". 
All of these were executed as a separate test and reported PASS/FAIL. Actually 
I created a wrapper, which also checked the output of the "./check" and allowed 
the test to result in "WARN" which I then used for reporting skipped tests.

Cleber, is there a better way of running qemu-iotests? I know Drew had a plan 
on adding kvm-unit-test runner which would support kvm-unit-tests out of the 
box including the proper setup, but I'm not aware of any such initiative for 


PS: Now I remembered that I actually contributed the kvm-unit-test wrapper 
upstream, it's slightly outdated compare to the downstream version but could 
work as an example (if you're interested I can try to synchronize it with the 
current downstream implementation): 

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]