[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250416144447.1fde7ada@kernel.org>
Date: Wed, 16 Apr 2025 14:44:47 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Michael Chan <michael.chan@...adcom.com>
Cc: davem@...emloft.net, netdev@...r.kernel.org, edumazet@...gle.com,
pabeni@...hat.com, andrew+netdev@...n.ch, pavan.chebbi@...adcom.com,
andrew.gospodarek@...adcom.com, Kalesh AP
<kalesh-anakkur.purayil@...adcom.com>
Subject: Re: [PATCH net-next 1/4] bnxt_en: Change FW message timeout warning
On Wed, 16 Apr 2025 14:41:10 -0700 Michael Chan wrote:
> On Tue, Apr 15, 2025 at 8:14 PM Jakub Kicinski <kuba@...nel.org> wrote:
>
> > sysctl_hung_task_timeout_secs is an exported symbol, and it defaults to 120.
> > Should you not use it in the warning (assuming I understand the intent
> > there)?
> Yes, we have considered that. This is only printed once at driver
> load time, but the sysctl value can be changed at any time after the
> driver is loaded. So we just want to use a reasonable value well
> below the default sysctl_hung_task_timeout_secs value as the
> threshold.
>
> But we can reference and compare with the
> sysctl_hung_task_timeout_secs value if that makes more sense.
I see your point. We could also check against
CONFIG_DEFAULT_HUNG_TASK_TIMEOUT ?
I noticed that some arches set this value really low (10 or 20 sec),
it may be worth warning the users in such cases.
Powered by blists - more mailing lists