Re: bnxt_en: BCM57414: 0000:c4:00.0 (unnamed net_device) (uninitialized): Device requests max timeout of 100 seconds, may trigger hung task watchdog

From: Paul Menzel

Date: Thu Feb 19 2026 - 03:44:48 EST


Dear Michael,


Thank you for your reply.

Am 12.02.26 um 20:10 schrieb Michael Chan:
On Thu, Feb 12, 2026 at 10:30 AM Paul Menzel <pmenzel@xxxxxxxxxxxxx> wrote:

There seem to be two problems:

1. It’s logged before the name is known.

Yes, this information is received from the first FW response when
probing the device. So the netdev name is not known yet and the PCI
device number is used instead.

Hmm, it’s still ugly. ;-) Why not log it later, or don’t use that logging macro and use something else, that does not expect the netdev name?

2. The timeout of 100 s seems higher than the one you expect, which is
60 s [1].

I believe the latest FW should use 60 seconds, but there are older FW
in the field that use 100 seconds.

We just bought the Dell server, and the Broadcom BCM57414 has the firmware version below:

$ ethtool -i net02
[…]
firmware-version: 233.0.195.0/pkg 23.31.18.10
[…]

It’d be great, if you told me the current firmware version from Broadcom, so that I can point Dell to it and ask to release an update?

(By the way, why not log this version too, so it shows up in `dmesg`?)

> But 60 seconds is the timeout on most production FW and cannot be
> reduced further.

It looks like you addressed some of this in commit 0fcad44a86bd
(bnxt_en: Change FW message timeout warning) and commit 8ff617513996
(bnxt_en: hide CONFIG_DETECT_HUNG_TASK specific code), merged in Linux
v6.16-rc1. Both were not backported to the 6.12 LTS series. Also, it
looks like they do not address the first issue.

As we do have `CONFIG_DETECT_HUNG_TASK` unset, we would not see the
warning with the patches.

$ grep HUNG_TASK /boot/config-6.12.49.mx64.487
# CONFIG_DETECT_HUNG_TASK is not set

I’d recommend to have both commits backported. Will you do this, or
should I contact the stable folks?

If AUTOSEL has not selected these patches for stable, I can request to
have these backported. I agree that the warning is annoying. I will
do it later today.

Thank you for sending in the request. I found it in the archives, and it’s queued for 6.12.74 [2].


Kind regards,

Paul


[1]: https://lore.kernel.org/all/CACKFLik3qQ46i3_wRpm0OnaNnMcMc-hVNrFt_0U9yzzjNrm4VA@xxxxxxxxxxxxxx/t/#u
[2]: https://lore.kernel.org/all/20260217200005.998240758@xxxxxxxxxxxxxxxxxxx/