[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <fc773151-4f80-4a85-8dc6-5c2aa4e0a8e6@intel.com>
Date: Tue, 13 Feb 2024 14:12:29 +0200
From: Sasha Neftin <sasha.neftin@...el.com>
To: Jakub Kicinski <kuba@...nel.org>
CC: Tony Nguyen <anthony.l.nguyen@...el.com>, <davem@...emloft.net>,
<pabeni@...hat.com>, <edumazet@...gle.com>, <netdev@...r.kernel.org>, "Paul
Menzel" <pmenzel@...gen.mpg.de>, Naama Meir <naamax.meir@...ux.intel.com>,
"Ruinskiy, Dima" <dima.ruinskiy@...el.com>, "Neftin, Sasha"
<sasha.neftin@...el.com>
Subject: Re: [PATCH net 2/2] igc: Remove temporary workaround
On 12/02/2024 19:08, Jakub Kicinski wrote:
> On Sun, 11 Feb 2024 08:53:36 +0200 Sasha Neftin wrote:
>>> Any more info on this one?
>>> What's the user impact?
>>> What changed (e.g. which FW version fixed it)?
>>
>> User impact: PHY could be powered down when the link is down (ip link
>> set down <device>)
>
> to make it a tiny bit clearer:
>
> s/could/can now/ ?
yes
> s/link is down/device is down/ ?
I meant to the Ethernet link. The device is enumerated and PCIe
configuration space is accessible. (ip link set down <device>)
>
>> Fix by PHY firmware and deployed via OEM updates (automatically, with
>> OEM SW/FW updates). We checked the IEEE behavior and removed w/a.
>>
>> The PHY vendor no longer works with Intel, but I can say this was fixed
>> on a very early silicon step (years ago).
>
> And the versions of the PHY components are not easily accessible so we
> can't point to the version that was buggy or at least the oldest you
> tested? If that's the case - it is what it is, please repost with the
> improved commit msg.
The version of the PHY FW is easily accessible (ethtool -i <device>).
The problem is that I can not point to the specific buggy version. It
was on the very early silicon step (it is what it is).
I will add the oldest version of the NVM we tested..
Powered by blists - more mailing lists