[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5af3e4a0-8013-4b50-af56-b8c0c6f71b5e@amd.com>
Date: Wed, 8 Oct 2025 13:42:32 -0500
From: Babu Moger <bmoger@....com>
To: Dave Martin <Dave.Martin@....com>,
Reinette Chatre <reinette.chatre@...el.com>
Cc: Babu Moger <babu.moger@....com>, tony.luck@...el.com,
james.morse@....com, tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
dave.hansen@...ux.intel.com, x86@...nel.org, hpa@...or.com,
linux-kernel@...r.kernel.org, peternewman@...gle.com, eranian@...gle.com,
gautham.shenoy@....com
Subject: Re: [PATCH] x86/resctrl: Fix buggy overflow when reactivating
previously Unavailable RMID
Hi Dave,
On 10/8/25 10:23, Dave Martin wrote:
> Hi,
>
> On Tue, Oct 07, 2025 at 01:23:36PM -0700, Reinette Chatre wrote:
>> Hi Babu,
>>
>> Thank you for catching and preparing a fix for this issue.
>>
>> On 10/6/25 4:13 PM, Babu Moger wrote:
>>> The issue was observed during testing on systems with multiple resctrl
>>> domains, where tasks were dynamically moved between domains.
>> (please let changelog stand on its own and not be a continuation of subject line)
> [...]
>
> [Quick, drive-by observation:]
>
> Can I add that the commit message also seems way too long?
I agree. But problem is bit complicated, so had to give all the steps.
>
> I think some of the description of the problem symptom could probably
> be after the tearoff -- there doesn't seem to be a clear statement of
> what is actually wrong in the code, or of why the change made in the
> patch fixes it (or if there is, I struggled to find it.)
>
> I puzzled over this for some minutes before I figured out this patch is
> fixing something that is not upstream, yet. A statement to that effect
> would have helped.
No. Code is already in upstream. Just that it is not resetting the
internal state when RMID reports "Unavailable".
Updated the changes log now with brief summary in the beginning.
thanks
Babu
Powered by blists - more mailing lists