linux-kernel - Re: [PATCH] tpm: tis: Increase the default for timeouts B and C

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: <Z-53dR25MT8OUDhW@kitsune.suse.cz>
Date: Thu, 3 Apr 2025 13:56:37 +0200
From: Michal Suchánek <msuchanek@...e.de>
To: Jonathan McDowell <noodles@...th.li>
Cc: Peter Huewe <peterhuewe@....de>, Jarkko Sakkinen <jarkko@...nel.org>,
	Jason Gunthorpe <jgg@...pe.ca>, linux-integrity@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] tpm: tis: Increase the default for timeouts B and C

On Thu, Apr 03, 2025 at 12:00:36PM +0100, Jonathan McDowell wrote:
> On Wed, Apr 02, 2025 at 10:07:39PM +0200, Michal Suchánek wrote:
> > On Wed, Apr 02, 2025 at 06:45:40PM +0100, Jonathan McDowell wrote:
> > > On Wed, Apr 02, 2025 at 07:21:30PM +0200, Michal Suchanek wrote:
> > > > With some Infineon chips the timeouts in tpm_tis_send_data (both B and
> > > > C) can reach up to about 2250 ms.
> > > >
> > > > Extend the timeout duration to accommodate this.
> > > 
> > > The problem here is the bump of timeout_c is going to interact poorly with
> > > the Infineon errata workaround, as now we'll wait 4s instead of 200ms to
> > > detect the stuck status change.
> > 
> > Yes, that's problematic. Is it possible to detect the errata by anything
> > other than waiting for the timeout to expire?
> 
> Not that I'm aware of, nor have seen in my experimentation. It's a "stuck"
> status, so the timeout is how it's detected.
> 
> OOI, have you tried back porting the fixes that are in mainline for 6.15 to
> your frankenkernel? I _think_ the errata fix might end up resolving at least
> the timeout for valid for you, as a side effect? We're currently rolling
> them out across our fleet, but I don't have enough runtime yet to be sure
> they've sorted all the timeout instances we see.

When was that merged?

The change I see is that sometimes EAGAIN is returned instead of ETIME
but based on the previous discussion this is unlikely to help.

Thanks

Michal

> 
> J.
> 
> -- 
> /-\                             | He's weird? It's ok, I'm fluent in
> |@/  Debian GNU/Linux Developer |               weird.
> \-                              |