On 8/9/2023 4:33 PM, Jason A. Donenfeld wrote:
Hey Greg,
There was recently a bit of a snafoo with a maintainer taking the wrong version of a patch and sending that up to Linus. That patch had incorrect stable@ annotations and had a bug in it. That bug was fixed with a follow up patch. But of course the metadata couldn't be changed easily retroactively.
So I'm emailing to ask you to backport these two patches back to 5.5:
- 554b841d4703 ("tpm: Disable RNG for all AMD fTPMs")
- cacc6e22932f ("tpm: Add a helper for checking hwrng enabled")
I know the stable@ tag says 6.1+, but the actual right tags from the newer versioned patch that didn't get picked are:
Cc: stable@vger.kernel.org # 5.5+ Fixes: b006c439d58d ("hwrng: core - start hwrng kthread also for untrusted sources") Fixes: f1324bbc4011 ("tpm: disable hwrng for fTPM on some AMD designs") Fixes: 3ef193822b25 ("tpm_crb: fix fTPM on AMD Zen+ CPUs") Reported-by: daniil.stas@posteo.net Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217719 Reported-by: bitlord0xff@gmail.com Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217212 Reviewed-by: Jason A. Donenfeld Jason@zx2c4.com Signed-off-by: Mario Limonciello mario.limonciello@amd.com
Let me know if you need any more info.
Thanks, Jason
So I had a quick try with the backports to see what happens. 6.1.y and 6.4.y apply cleanly no problem.
However 5.15.y (and presumably 5.5.y) have a variety of issues that I think no longer make it a stable candidate. I started going down the rabbit hole of dependencies and it's massive unless hand modifications are done.
Realistically the problem is most severe in 6.1.y because of b006c439d58d. I don't know it's worth going back any further.