[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAG-rBihOr+aAZhO4D2VBwSx-EGg_gbgBYKN3fSBTPKCXdz9AqA@mail.gmail.com>
Date: Wed, 21 Feb 2024 10:52:52 -0500
From: Sven van Ashbrook <svenva@...omium.org>
To: Vlastimil Babka <vbabka@...e.cz>
Cc: Andrew Morton <akpm@...ux-foundation.org>, bgeffon@...gle.com,
cujomalainey@...omium.org, kramasub@...omium.org,
linux-kernel@...r.kernel.org, linux-mm@...ck.org, linux-sound@...r.kernel.org,
perex@...ex.cz, stable@...r.kernel.org, tiwai@...e.com, tiwai@...e.de,
Michal Hocko <mhocko@...nel.org>, Mel Gorman <mgorman@...hsingularity.net>
Subject: Re: [PATCH] mm, vmscan: prevent infinite loop for costly GFP_NOIO |
__GFP_RETRY_MAYFAIL allocations
Thanks so much ! We will stress test this on our side.
We do this by exhausting memory and triggering many suspend/resume
cycles. This reliably reproduces the problem (before this patch).
Of course, as we all know, absence of evidence (no more stalls in stress tests)
does not equal evidence of absence (stalls are gone in all code paths).
Powered by blists - more mailing lists