[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250415175359.3c6117c9@kernel.org>
Date: Tue, 15 Apr 2025 17:53:59 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Przemek Kitszel <przemyslaw.kitszel@...el.com>
Cc: Jaroslav Pulchart <jaroslav.pulchart@...ddata.com>,
<jdamato@...tly.com>, <intel-wired-lan@...ts.osuosl.org>,
<netdev@...r.kernel.org>, Tony Nguyen <anthony.l.nguyen@...el.com>, "Igor
Raits" <igor@...ddata.com>, Daniel Secik <daniel.secik@...ddata.com>,
"Zdenek Pesek" <zdenek.pesek@...ddata.com>, "Eric Dumazet"
<edumazet@...gle.com>, Martin Karsten <mkarsten@...terloo.ca>, "Ahmed Zaki"
<ahmed.zaki@...el.com>, "Czapnik, Lukasz" <lukasz.czapnik@...el.com>,
Michal Swiatkowski <michal.swiatkowski@...ux.intel.com>
Subject: Re: Increased memory usage on NUMA nodes with ICE driver after
upgrade to 6.13.y (regression in commit 492a044508ad)
On Tue, 15 Apr 2025 16:38:40 +0200 Przemek Kitszel wrote:
> > We traced the issue to commit 492a044508ad13a490a24c66f311339bf891cb5f
> > "ice: Add support for persistent NAPI config".
>
> thank you for the report and bisection,
> this commit is ice's opt-in into using persistent napi_config
>
> I have checked the code, and there is nothing obvious to inflate memory
> consumption in the driver/core in the touched parts. I have not yet
> looked into how much memory is eaten by the hash array of now-kept
> configs.
+1 also unclear to me how that commit makes any difference.
Jaroslav, when you say "traced" what do you mean?
CONFIG_MEM_ALLOC_PROFILING ?
The napi_config struct is just 24B. The queue struct (we allocate
napi_config for each queue) is 320B...
Powered by blists - more mailing lists