linux-kernel - Re: [REGRESSION] suspend to ram fails in 6.2-rc1 due to tpm errors

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: <Y7btcIGI0LpiqN68@zx2c4.com>
Date:   Thu, 5 Jan 2023 16:32:00 +0100
From:   "Jason A. Donenfeld" <Jason@...c4.com>
To:     James Bottomley <James.Bottomley@...senPartnership.com>
Cc:     Thorsten Leemhuis <regressions@...mhuis.info>,
        Vlastimil Babka <vbabka@...e.cz>,
        Peter Huewe <peterhuewe@....de>,
        Jarkko Sakkinen <jarkko@...nel.org>,
        Jason Gunthorpe <jgg@...pe.ca>, Jan Dabros <jsd@...ihalf.com>,
        regressions@...ts.linux.dev, LKML <linux-kernel@...r.kernel.org>,
        linux-integrity@...r.kernel.org,
        Dominik Brodowski <linux@...inikbrodowski.net>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Johannes Altmanninger <aclopte@...il.com>
Subject: Re: [REGRESSION] suspend to ram fails in 6.2-rc1 due to tpm errors

On Thu, Jan 05, 2023 at 04:27:02PM +0100, Jason A. Donenfeld wrote:
> On Thu, Jan 05, 2023 at 10:17:57AM -0500, James Bottomley wrote:
> > On Thu, 2023-01-05 at 14:59 +0100, Thorsten Leemhuis wrote:
> > > On 29.12.22 05:03, Jason A. Donenfeld wrote:
> > > > On Wed, Dec 28, 2022 at 06:07:25PM -0500, James Bottomley wrote:
> > > > > On Wed, 2022-12-28 at 21:22 +0100, Vlastimil Babka wrote:
> > > > > > Ugh, while the problem [1] was fixed in 6.1, it's now happening
> > > > > > again on the T460 with 6.2-rc1. Except I didn't see any oops
> > > > > > message or "tpm_try_transmit" error this time. The first
> > > > > > indication of a problem is this during a resume from suspend to
> > > > > > ram:
> > > > > > 
> > > > > > tpm tpm0: A TPM error (28) occurred continue selftest
> > > > > > 
> > > > > > and then periodically 
> > > > > > 
> > > > > > tpm tpm0: A TPM error (28) occurred attempting get random
> > > > > 
> > > > > That's a TPM 1.2 error which means the TPM failed the selftest. 
> > > > > The original problem was reported against TPM 2.0  because of a
> > > > > missing try_get_ops().
> > > > 
> > > > No, I'm pretty sure the original bug, which was fixed by "char:
> > > > tpm: Protect tpm_pm_suspend with locks" regards 1.2 as well,
> > > > especially considering it's the same hardware from Vlastimil
> > > > causing this. I also recall seeing this in 1.2 when I ran this with
> > > > the TPM emulator. So that's not correct.
> > > 
> > > James, are you or some other TPM developer looking into this? Or is
> > > this deadlocked now?
> > 
> > Not really: TPM 1.2 way predates my interest in the TPM subsystem, and
> > I've only ever done patches to 2.0.  I can look at the paths
> > theoretically, but I don't have any hardware.  Self Test failures tend
> > to be hardware specific, so even poking around in the emulator is
> > unlikely to give what might be the cause.
> > 
> > >  And if so: how can we get this unstuck to get this regression
> > > solved?
> > 
> > One of the TPM maintainers with hardware (possibly the specific TPM ...
> > what is it, by the way?) needs to get involved.
> 
> I already wrote in my last email [1] that this simply isn't the case.
> The issue reproduces in QEMU + the emulator. It's far more likely to be
> a locking/race situation [2] than some kind of obscure hardware bug.
> 
> Jason
> 
> [1] https://lore.kernel.org/lkml/Y60RoP77HnwaukEA@zx2c4.com/
> [2] https://lore.kernel.org/lkml/Y60Uu8HcGyXasnOO@zx2c4.com/

I should add, though, that I don't fault you for lack of interest in TPM
1.2. Folks work on what they care about. If the rest of the TPM team has
a similar feeling, maybe it's best to just merge [3], rather than let
the bug linger forever? And possibly talk about sunsetting TPM 1 support
entirely if there's nobody to work on it and fix issues that inevitably
will come up? Just trying to look at the big picture here, you know...

[3] https://lore.kernel.org/lkml/20230105144742.3219571-1-Jason@zx2c4.com/