Re: [GIT PULL] pstore updates for v6.6-rc1

[Kees Cook]

On Mon, Aug 28, 2023 at 06:44:02PM -0700, Linus Torvalds wrote:
> The only thing that is new is the kernel pstore implementation. Why
> was this not a problem before? The warning existed back then too, but
> I never actually got it.

Right -- if the compression method from before was different, it'll fail
now. (i.e. we removed everything but zlib.)

> I get the feeling that you are overlooking that basic fact.

That's why I was wondering about the prior config; it could confirm the
default compression algo. But digging around it seems like zlib is the
default in the F37 kernel config. I'll keep looking; there is clearly
some combination I don't know.

I remain concerned about why there are 124. That's a LOT, and without
prior warnings, I don't know why systemd-pstore wasn't removing them.
Can you send me "ls -la /sys/fs/pstore" ? Maybe they aren't a dump type
that systemd knows about.

I will try to reproduce this with an F37 image...

-- 
Kees Cook