On Tue, Jul 07, 2026 at 11:35:17AM -0700, Borislav Petkov wrote:

On Tue, Jul 07, 2026 at 09:53:34AM -0700, Breno Leitao wrote:

quoted

The current vmcoreinfo implementation defaults to enabled, so I wanted to
preserve that behavior to avoid silently removing symbols that existing
tools may depend on. Would you prefer a different default?

You do know Linus' stance on enabling things by default which are not
ubiquitous, right?

Oh yea, I am familiar with it. :-)

Logging hw errors in vmcore is really necessary to be enabled everywhere?

I am happy to make it no, if we transform it to a KCONFIG. Right now it
is not a Kconfig, so, it comes with VMCORE set of exported fields.

quoted

I agree the sysfs exposure duplicates existing infrastructure, points
taken. However, tracking fatal hardware errors provides value at crash
analysis time—it lets us quickly determine whether a fatal hardware
error occurred during the kernel's lifetime, which is useful for
root-cause attribution.

Yes, that's why you put it in vmcore. You can't read sysfs if you encounter
a fatal hw error.

Ack. Right now it is not tracking fatal error (just recoverable error),
although I think it is a good idea to also track Fatal error (of course
that not on sysfs).

quoted

For this, would you like to keep it in vmcore info (as of today), or
move to RAS subsystem?

You mean, would I like to pay attention to more patches than now?

Not really - I can barely manage as it is.

Come one Borislav, we need your insights/review/opinions here as well.
:-)