linux-kernel - Re: [PATCH AUTOSEL 5.17 089/135] kunit: bail out of test filtering logic quicker if OOM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 31 May 2022 09:15:11 -0700
From:   Daniel Latypov <dlatypov@...gle.com>
To:     Sasha Levin <sashal@...nel.org>
Cc:     linux-kernel@...r.kernel.org, stable@...r.kernel.org,
        Zeal Robot <zealci@....com.cn>, Lv Ruyi <lv.ruyi@....com.cn>,
        Brendan Higgins <brendanhiggins@...gle.com>,
        Shuah Khan <skhan@...uxfoundation.org>,
        linux-kselftest@...r.kernel.org, kunit-dev@...glegroups.com
Subject: Re: [PATCH AUTOSEL 5.17 089/135] kunit: bail out of test filtering
 logic quicker if OOM

On Mon, May 30, 2022 at 6:36 AM Sasha Levin <sashal@...nel.org> wrote:
>
> From: Daniel Latypov <dlatypov@...gle.com>
>
> [ Upstream commit a02353f491622e49c7ddedc6a6dc4f1d6ed2150a ]
>

Is it possible to make sure the fix for this commit gets picked up as well [1]?
I was waiting a bit to see if it was going to get picked up
automatically, but I don't see such an email yet:
https://lore.kernel.org/stable/?q=kunit+OOM
(Perhaps the automation just hasn't gotten around to it yet?)

Alternatively, reverting just this commit would also work, if that's easier.

[1] commit 1b11063d32d7e11366e48be64215ff517ce32217
Author: Daniel Latypov <dlatypov@...gle.com>
Date:   Fri May 13 11:37:07 2022 -0700

    kunit: fix executor OOM error handling logic on non-UML

    The existing logic happens to work fine on UML, but is not correct when
    running on other arches.

    1. We didn't initialize `int err`, and kunit_filter_suites() doesn't
       explicitly set it to 0 on success. So we had false "failures".
       Note: it doesn't happen on UML, causing this to get overlooked.
    2. If we error out, we do not call kunit_handle_shutdown().
       This makes kunit.py timeout when using a non-UML arch, since the QEMU
       process doesn't ever exit.

    Fixes: a02353f49162 ("kunit: bail out of test filtering logic
quicker if OOM")
    Signed-off-by: Daniel Latypov <dlatypov@...gle.com>
    Reviewed-by: Brendan Higgins <brendanhiggins@...gle.com>
    Signed-off-by: Shuah Khan <skhan@...uxfoundation.org>

Without 1b11063d32d7 above, this "fix" breaks more cases than it fixes
due to my sloppiness.

Thanks,
Daniel