> On 2010-08-23 10:56:55, Gabe Black wrote: > > How about using the IsSerializing/IsSerializeAfter flags? I think that's a > > little different where code is fetched but stopped from going through > > decode, maybe, so those might not work. Otherwise the code looks fine. The > > first summary line is too long and gets cut off.
Tried that first, but it doesn't work. The serializing flags stall the processor until all instructions ahead have been committed, but that says nothing about the store buffer state. The stores (to the instruction stream), were sitting in the store buffer and hadn't actually made it on the "bus" yet. I think this does a pretty realistic thing when a complete flush needs to happen. - Ali ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/230/#review307 ----------------------------------------------------------- On 2010-08-23 09:35:07, Ali Saidi wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.m5sim.org/r/230/ > ----------------------------------------------------------- > > (Updated 2010-08-23 09:35:07) > > > Review request for Default. > > > Summary > ------- > > ARM: Adding a bogus fault that does nothing. This fault can used to flush the > pipe, > not including the faulting instruction. > > The particular case I needed this was for a self-modifying code. It needed to > drain the store queue and force the following instruction to refetch from > icache. DCCMVAC cp15 mcr instruction is modified to raise this fault. > > > Diffs > ----- > > src/arch/arm/faults.hh 47d9409b2b7f > src/arch/arm/faults.cc 47d9409b2b7f > src/arch/arm/isa/formats/misc.isa 47d9409b2b7f > src/arch/arm/isa/formats/unimp.isa 47d9409b2b7f > > Diff: http://reviews.m5sim.org/r/230/diff > > > Testing > ------- > > > Thanks, > > Ali > > _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
