[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <c45dd329-9606-4103-60bd-904f9f29f01c@linux.microsoft.com>
Date: Tue, 11 Jan 2022 13:12:31 -0800
From: Vijay Balakrishna <vijayb@...ux.microsoft.com>
To: Heiner Kallweit <hkallweit1@...il.com>,
Michael Chan <michael.chan@...adcom.com>
Cc: Netdev <netdev@...r.kernel.org>
Subject: Re: [bnxt] Error: Unable to read VPD
On 1/11/2022 11:41 AM, Heiner Kallweit wrote:
> On 10.01.2022 22:15, Michael Chan wrote:
>> On Mon, Jan 10, 2022 at 1:02 PM Vijay Balakrishna
>> <vijayb@...ux.microsoft.com> wrote:
>>>
>>>
>>> Since moving to 5.10 from 5.4 we are seeing
>>>
>>>> Jan 01 00:00:01 localhost kernel: bnxt_en 0008:01:00.0 (unnamed net_device) (uninitialized): Unable to read VPD
>>>>
>>>> Jan 01 00:00:01 localhost kernel: bnxt_en 0008:01:00.1 (unnamed net_device) (uninitialized): Unable to read VPD
>>>
>>> these appear to be harmless and introduced by
>>>
>>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=a0d0fd70fed5cc4f1e2dd98b801be63b07b4d6ac
>>> Does "Unable to read VPD" need to be an error or can it be a warning
>>> (netdev_warn)?
>>>
>>
>> We can change these to warnings. Thanks.
>
> Since 5.15 it is a pci_warn() already. Supposedly "Unable to read VPD" here simply means
> that the device has no (valid) VPD. Does "lspci -vv" list any VPD info?
> If VPD is an optional feature, then maybe the warning should be changed to info level
> and the text should be less alarming.
See no VPD info.
> /sys/bus/pci/devices# ls
> 0000:00:00.0 0000:01:00.0 0000:01:00.1 0008:00:00.0 0008:01:00.0 0008:01:00.1
> /sys/bus/pci/devices# for i in `ls`; do cd $i;ls -lR | grep vpd;cd ..; done
> /sys/bus/pci/devices# lspci -vv | egrep -i '(vpd|vital)'
> /sys/bus/pci/devices
Powered by blists - more mailing lists