Re: [BUG] tpm_crb/Intel PTT: Late S3/S5/Reboot ACPI power management deadlock

[Jarkko Sakkinen <jarkko@kernel.org>]

On Tue, Sep 29, 2026 at 01:44:29PM -0400, Sean Meadows wrote:
> 
> On 9/25/26 11:05, Jarkko Sakkinen wrote:
> > 
> > Thanks for providing the information.
> > 
> > Bugzilla bug under discussion: https://bugzilla.kernel.org/
> > show_bug.cgi? id=217890
> > 
> > I also found fix attempt of which 3rd version somehow went past me:
> > 
> > https://lore.kernel.org/all/20240308145313.40932-1-adamoa@gmail.com/
> > 
> > Does this or disabling hwrng sort out the issue?
> > 
> > I got one other idea too with no empirical basis (just hypothesis): perhaps
> > release memory mapping on suspend could help firmware.
> > 
> > 
> > Br, Jarkko
> 
> Acknowledged.
> 
> Disabling HWRNG via kernel cmdline failed, and since my research implied
> that path would require a custom kernel, I shifted my efforts toward
> forward-porting Mr. Alves's 2024 patch to 7.2.7. It took some work, but
> I have successfully compiled it and got it booting on my machine.
> 
> The patch appears to sort out the issue so far, but I'm only 43 hours
> into the test. As we discussed previously, this bug can take days to
> manifest, so I will reply again once I cross a week of sleep cycles
> without a crash.
> 
> I also have a USB3 debug cable on the way in hope of capturing the
> hardware state during the power transition. That said, if Mr. Alves's
> patch continues to work as well as it has so far, there may not be any
> need for that.
> 
> Sean

There's a commit with title "tpm: Call cmd_ready/go_idle for each command
transmission", which is strongest contender to make a difference to this
long running issue.

I.e. I think it will improve firmware behavior given full release of
resources per TPM command.

I'm including it to v7.4 pull request.

Br, Jarkko