Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <b140c37fe1aac315e07464dd6a2d7a8f463e6fe4.camel@redhat.com>
Subject: Re: [PATCH v6] net: atlantic: always deep reset on pm op, fixing up
 my null deref regression
From:   Paolo Abeni <pabeni@redhat.com>
To:     Manuel Ullmann <labre@posteo.de>,
        Igor Russkikh <irusskikh@marvell.com>
Cc:     linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
        regressions@lists.linux.dev, davem@davemloft.net,
        ndanilov@marvell.com, kuba@kernel.org, edumazet@google.com,
        Jordan Leppert <jordanleppert@protonmail.com>,
        Holger Hoffstaette <holger@applied-asynchrony.com>,
        koo5 <kolman.jindrich@gmail.com>
Date:   Tue, 10 May 2022 10:37:24 +0200
In-Reply-To: <87bkw8dfmp.fsf@posteo.de>
References: <87bkw8dfmp.fsf@posteo.de>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.42.4 (3.42.4-2.fc35) 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Sun, 2022-05-08 at 00:36 +0000, Manuel Ullmann wrote:
> From 18dc080d8d4a30d0fcb45f24fd15279cc87c47d5 Mon Sep 17 00:00:00 2001
> Date: Wed, 4 May 2022 21:30:44 +0200
> 
> The impact of this regression is the same for resume that I saw on
> thaw: the kernel hangs and nothing except SysRq rebooting can be done.
> 
> Fixes regression in commit cbe6c3a8f8f4 ("net: atlantic: invert deep
> par in pm functions, preventing null derefs"), where I disabled deep
> pm resets in suspend and resume, trying to make sense of the
> atl_resume_common() deep parameter in the first place.
> 
> It turns out, that atlantic always has to deep reset on pm
> operations. Even though I expected that and tested resume, I screwed
> up by kexec-rebooting into an unpatched kernel, thus missing the
> breakage.
> 
> This fixup obsoletes the deep parameter of atl_resume_common, but I
> leave the cleanup for the maintainers to post to mainline.
> 
> Suspend and hibernation were successfully tested by the reporters.
> 
> Fixes: cbe6c3a8f8f4 ("net: atlantic: invert deep par in pm functions, preventing null derefs")
> Link: https://lore.kernel.org/regressions/9-Ehc_xXSwdXcvZqKD5aSqsqeNj5Izco4MYEwnx5cySXVEc9-x_WC4C3kAoCqNTi-H38frroUK17iobNVnkLtW36V6VWGSQEOHXhmVMm5iQ=@protonmail.com/
> Reported-by: Jordan Leppert <jordanleppert@protonmail.com>
> Reported-by: Holger Hoffstaette <holger@applied-asynchrony.com>
> Tested-by: Jordan Leppert <jordanleppert@protonmail.com>
> Tested-by: Holger Hoffstaette <holger@applied-asynchrony.com>
> CC: <stable@vger.kernel.org> # 5.10+
> Signed-off-by: Manuel Ullmann <labre@posteo.de>
> ---
> I'm very sorry for this regression. It would be nice, if this could
> reach mainline before 5.18 release, if applicable. This restores the
> original suspend behaviour, while keeping the fix for hibernation. The
> fix for hibernation might not be the root cause, but still is the most
> simple fix for backporting to stable while the root cause is unknown
> to the maintainers.
> 
> Changes in v2:
> Patch formatting fixes
> ~ Fix Fixes tag
> ~ Simplify stable Cc tag
> ~ Fix Signed-off-by tag
> 
> Changes in v3:
> ~ Prefixed commit reference with "commit" aka I managed to use
>   checkpatch.pl.
> ~ Added Tested-by tags for the testing reporters.
> ~ People start to get annoyed by my patch revision spamming. Should be
>   the last one.
> 
> Changes in v4:
> ~ Moved patch changelog to comment section
> ~ Use unicode ndash for patch changelog list to avoid confusion with
>   diff in editors
> ~ Expanded comment
> ~ Targeting net-next by subject
> 
> Changes in v5:
> ~ Changed my MTA transfer encoding to 8 bit instead of
>   quoted-printable. Git should like this a bit more.
> 
> Changes in v6:
> ~ Reducing content to 7 bit chars, because nipa did not apply v4 and v5, while
>   git does against a fresh net-next HEAD. Maybe it chokes on the
>   additional bit.
> ~ Omitting target tree to resemble the last passing patch version the most.

For future submission, please always specify the target tree (which is
-net in this case). No need to resubmit: I'm applying it.

Thanks,

Paolo