netdev - Re: [PATCH net-next] eth: fbnic: fix csr boundary for RPM RAM section

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: <20241220130338.7c8aca32@kernel.org>
Date: Fri, 20 Dec 2024 13:03:38 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: Andrew Lunn <andrew@...n.ch>
Cc: Mohsin Bashir <mohsin.bashr@...il.com>, netdev@...r.kernel.org,
 alexanderduyck@...com, andrew+netdev@...n.ch, davem@...emloft.net,
 edumazet@...gle.com, pabeni@...hat.com, kernel-team@...a.com
Subject: Re: [PATCH net-next] eth: fbnic: fix csr boundary for RPM RAM
 section

On Thu, 19 Dec 2024 09:25:41 +0100 Andrew Lunn wrote:
> On Wed, Dec 18, 2024 at 03:25:58PM -0800, Mohsin Bashir wrote:
> > The CSR dump support leverages the FBNIC_BOUNDS macro, which pads the end
> > condition for each section by adding an offset of 1. However, the RPC RAM
> > section, which is dumped differently from other sections, does not rely
> > on this macro and instead directly uses end boundary address. Hence,
> > subtracting 1 from the end address results in skipping a register.  
> 
> Maybe it would be better to actually use FBNIC_BOUNDS macro, to make
> it the same as all the others, and so avoid errors like this because
> it is special?

That may require changing the format / contents of the dump a bit
to include unimplemented registers (which would read as ff).
Note that the subject is incorrect, the patch is for net, 
so the one-liner fix is more appealing at this stage.
But I may be biased...