On Tue, 2025-02-11 at 09:42 +0800, Yan Zhao wrote:
> > Is this a fix for the intermittent failure we saw on the v6.13-rc3 based kvm
> > branch? Funnily, I can't seem to reproduce it anymore, with or without this
> > fix.
> Hmm, it can be reproduced in my SPR (non TDX) almost every time.
> It depends on the timing when mprotect(PROT_READ) is completed done.
>
> Attached the detailed error log in my machine at the bottom.
I must be getting lucky on timing. BTW, in the above I actually meant on either
the new or old *kernel*.
>
> > On the fix though, doesn't this remove the coverage of writing to a region
> > that
> > is in the process of being made RO? I'm thinking about warnings, etc that
> > may
> > trigger intermittently based on bugs with a race component. I don't know if
> > we
> > could fix the test and still leave the write while the "mprotect(PROT_READ)
> > is
> > underway". It seems to be deliberate.
> Write before "mprotect(PROT_READ)" has been tested in stage 0.
> Not sure it's deliberate to test write in the process of being made RO.
> If it is, maybe we could make the fix by writing to RO memory a second time
> after mprotect_ro_done is true:
That could work if it's desirable to maintain the testing. I would mention the
reduced scope in the log at least. Maybe Sean will chime in.
Also, I think it needs:
Fixes: b6c304aec648 ("KVM: selftests: Verify KVM correctly handles
mprotect(PROT_READ)")
