Re: [PATCH] Revert "mt76: mt7921: enable aspm by default"

From: Kalle Valo <kvalo@kernel.org>
To: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: Philippe Schenker <dev@pschenker.ch>,
	linux-wireless@vger.kernel.org, Felix Fietkau <nbd@nbd.name>,
	"David S. Miller" <davem@davemloft.net>,
	Deren Wu <deren.wu@mediatek.com>,
	Jakub Kicinski <kuba@kernel.org>,
	Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	Paolo Abeni <pabeni@redhat.com>,
	Ryder Lee <ryder.lee@mediatek.com>,
	Sean Wang <sean.wang@mediatek.com>,
	Shayne Chen <shayne.chen@mediatek.com>,
	YN Chen <YN.Chen@mediatek.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-mediatek@lists.infradead.org,
	netdev@vger.kernel.org,
	"regressions\@lists.linux.dev" <regressions@lists.linux.dev>
Subject: Re: [PATCH] Revert "mt76: mt7921: enable aspm by default"
Date: Tue, 12 Apr 2022 13:33:52 +0300	[thread overview]
Message-ID: <87sfqioajz.fsf@kernel.org> (raw)
In-Reply-To: <7aa0bbd0-5498-ba74-ad6d-6dacbade8a3d@leemhuis.info> (Thorsten Leemhuis's message of "Tue, 12 Apr 2022 11:55:13 +0200")

Thorsten Leemhuis <regressions@leemhuis.info> writes:

> On 12.04.22 11:37, Kalle Valo wrote:
>> Philippe Schenker <dev@pschenker.ch> writes:
>> 
>>> This reverts commit bf3747ae2e25dda6a9e6c464a717c66118c588c8.
>>>
>>> This commit introduces a regression on some systems where the kernel is
>>> crashing in different locations after a reboot was issued.
>>>
>>> This issue was bisected on a Thinkpad P14s Gen2 (AMD) with latest firmware.
>>>
>>> Link:
>>> https://lore.kernel.org/linux-wireless/5077a953487275837e81bdf1808ded00b9676f9f.camel@pschenker.ch/
>>> Signed-off-by: Philippe Schenker <dev@pschenker.ch>
>> 
>> Can I take this to wireless tree? Felix, ack?
>> 
>> I'll also add:
>> 
>> Fixes: bf3747ae2e25 ("mt76: mt7921: enable aspm by default")
>
> Sorry, stupid questions from the regression tracker: wouldn't this cause
> a regression for users of kernel versions post-bf3747ae2e25, as the
> power consumption is likely to increase for them? Without having dug
> into the backstory much: would disabling ASPM for this particular
> machine using a quirk be the better approach? Or are we assuming a lot
> of machines are affected?

Kernel crashing is far more serious than increased power consumption. If
there's a better fix available in the next day or two of course that can
be considered. But if there's no such fix available, we have to revert
the commit.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek