> On 2011-11-06 01:00:17, Gabe Black wrote: > > It would be nice if we could specialize the flags for SE and FS modes since > > the behavior and requirements are different, but that probably wouldn't be > > worth the effort right now. Looks good to me. > > Ali Saidi wrote: > In theory it should be the same for both, with fs we were just getting > lucky that the fault to handle the syscall is serializing the execution. > >
I wouldn't call it lucky, I'd say the reason it's implemented (in real life) as a fault is partly to provide the serialization, and arguably the reason it doesn't work in se is because we don't treat it like a fault as we should. But this works too... - Steve ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/901/#review1634 ----------------------------------------------------------- On 2011-11-03 13:29:05, Ali Saidi wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.m5sim.org/r/901/ > ----------------------------------------------------------- > > (Updated 2011-11-03 13:29:05) > > > Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and > Nathan Binkert. > > > Summary > ------- > > ARM: Add IsSerializeAfter and IsNonSpeculative flag to the syscall > instruction . > > Squashes the subsequent instructions in O3 pipe after the service call, so > that > they see the effect of the system call when re-executed. This isn't really an > issue > with FS mode, but can show up in SE mode. > > > Diffs > ----- > > src/arch/arm/isa/insts/misc.isa 5fb918115c07 > > Diff: http://reviews.m5sim.org/r/901/diff > > > Testing > ------- > > > Thanks, > > Ali > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
