Hi Finn,
On 13/09/21 19:20, Michael Schmitz wrote:
Incidentally - have you ever checked whether Al Viro's signal handling
fixes have an impact on these bugs?
I will try that patch series if you think it is related.
Initial tests look promising (but I've said that before).
So far the problem seems to be confined to one machine. Stress tests on
other mac models did not yet reproduce the problem.
Yes, that's suspicious. I'll keep you posted.
I've seen neither format errors in resume_userspace() nor bus errors in
setup_frame() after two days of tests (and would have expected multiple
errors in that amount of time). Al's patches appear to have fixed these
errors for me. And that makes a lot of sense, considering handling of
multiple signals was broken before, and would have been expected to
result in corrupt exception frames on return from exception, and just
the sort of format errors I'd seen.
I still have to trace what issues the incorrect exception stack handling
would have caused in setup_frame() to force a bus error on the first
__put_user() there.
My tests were run successfully without any disabling of interrupts games
BTW.
The issues on 040 would be very similar, and I don't think you'd seen
format errors on exception return or bus errors in setup_frame(), so
this might not matter to you.
Cheers,
Michael
