<p dir="ltr"></p>
<p dir="ltr">On 14 Sep. 2016 10:12 pm, "Nicholas Piggin" <<a href="mailto:npiggin@gmail.com">npiggin@gmail.com</a>> wrote:<br>
><br>
> On Wed, 14 Sep 2016 21:46:39 +1000<br>
> Michael Neuling <<a href="mailto:mikey@neuling.org">mikey@neuling.org</a>> wrote:<br>
><br>
> > On Wed, 2016-09-14 at 21:28 +1000, Nicholas Piggin wrote:<br>
> > > Cc'ing Carlos<br>
> > ><br>
> > > On Wed, 14 Sep 2016 18:02:14 +1000</p>
<p dir="ltr">> > I think we might be able to detect this case in the kernel. If it's a tabort<br>
> > that's trapped on, we can't have been transactional. Hence we can safely PC+=4<br>
> > and leave off TM off. <br>
> ><br>
> > It would cost us a get_user(inst, regs->nip); but it might be worth it for this<br>
> > special but common case.<br>
><br>
> That would take an extra trap for every syscall, I think.</p>
<p dir="ltr">You're right. That wouldn't work.</p>
<p dir="ltr">Mikey<br></p>