netdev - Re: [PATCH net-next v3 3/4] net/mlx5e: Add logic to read RS-FEC histogram bin ranges from PPHCR

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20250918084000.1b4fb4f4@kernel.org>
Date: Thu, 18 Sep 2025 08:40:00 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Carolina Jubran <cjubran@...dia.com>
Cc: Vadim Fedorenko <vadim.fedorenko@...ux.dev>, Andrew Lunn
 <andrew@...n.ch>, Michael Chan <michael.chan@...adcom.com>, Pavan Chebbi
 <pavan.chebbi@...adcom.com>, Tariq Toukan <tariqt@...dia.com>, Gal Pressman
 <gal@...dia.com>, intel-wired-lan@...ts.osuosl.org, Donald Hunter
 <donald.hunter@...il.com>, Paolo Abeni <pabeni@...hat.com>, Simon Horman
 <horms@...nel.org>, netdev@...r.kernel.org, Yael Chemla
 <ychemla@...dia.com>, Dragos Tatulea <dtatulea@...dia.com>
Subject: Re: [PATCH net-next v3 3/4] net/mlx5e: Add logic to read RS-FEC
 histogram bin ranges from PPHCR

On Thu, 18 Sep 2025 18:16:14 +0300 Carolina Jubran wrote:
> On 18/09/2025 17:35, Jakub Kicinski wrote:
> > On Thu, 18 Sep 2025 17:25:40 +0300 Carolina Jubran wrote:  
> >>> why does MLX5E_FEC_RS_HIST_MAX exist?
> >>> We care that bins_cnt <= ETHTOOL_FEC_HIST_MAX - 1
> >>> or is there something in the interface that hardcodes 16?  
> >> My intention was to keep mlx5 capped at 16 even if ethtool raises its max.
> >> We’ll only increase this once we know the device should expose more than 16.
> >> Since our HW has internal modes that can report more than 16 bins, this
> >> ensures we don’t accidentally expose them if ethtool increases its max.  
> > But why?  
> 
> Because currently those modes shouldn't be exposed for ethernet.

I understand that the modes should not be exposed.
I don't get why this has anything to do with the number of bins.
Does the FW hardcode that the non-Ethernet modes use bins >=16?
When you say "internal modes that can report more than 16 bins"
it sounds like it uses bins starting from 0, e.g. 0..31.