Re: [REGRESSION] x86/debug: After PTRACE_SINGLESTEP DR_STEP is no longer reported in dr6

[Peter Zijlstra <peterz@infradead.org>]

On Mon, Oct 26, 2020 at 04:30:32PM -0700, Andy Lutomirski wrote:
> Is there any compelling reason not to just drop the condition and do:
> 
> current->thread.virtual_dr6 |= (dr6 & DR_STEP);
> 
> unconditionally?

Because why should it?

> This DR6 cause, along with ICEBP, have the
> regrettable distinctions of being the only causes that a user program
> can trigger all on its own without informing the kernel first.  This
> means that we can't fully separate the concept of "user mode is
> single-stepping itself" from "ptrace or something else is causing the
> kernel to single step a program."

TIF_SINGLESTEP does that. If the kernel is single-stepping userspace it
has TIF_SINGLESTEP (and possibly TIF_FORCED_TF) to distinguish these
cases.

> I bet that, without making this tweak, the virtual_dr6 change will
> regress some horrific Wine use case.

Then we should make sure the Wine people are aware and test this. Do you
know who to poke?

If there are regressions, we'll fix them, but I'd prefer to not create a
mess just because. This whole #DB thing was a giant trainwreck, we'll
obviously have to be bug compatible, but only when people actually
notice.