* Ingo Molnar <mingo@xxxxxxxxxx> wrote: > Assuming this does not fix the regression, could you apply the > minimal patch below - which reverts the old_rsp handling change. Assuming this solves the regression (it really should, it's now equivalent to a full revert minus comments): > @@ -395,6 +398,8 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) > /* > * Switch the PDA and FPU contexts. > */ > + prev->usersp = this_cpu_read(old_rsp); > + this_cpu_write(old_rsp, next->usersp); > this_cpu_write(current_task, next_p); > > /* can you confirm that your guest (sometimes) uses SYSENTER to do syscalls? If yes then my theory is that we broke SYSENTER (or SYSEXIT) support - and that this would not be visible in our normal tests of KVM because SYSCALL is used most of the time. Thanks, Ingo -- To unsubscribe from this list: send the line "unsubscribe linux-tip-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
|