Vince Weaver wrote:
> Hello
>
> I've tested this patch. It is broken on 32-bit x86 unless the below patch
> is applied. With that applied, it seems to work fine on the various test
> cases I have, including truncate64 and ftruncate64.
>
> Vince
>
> diff -r 62800dbea5a8 src/arch/x86/process.cc
> --- a/src/arch/x86/process.cc Tue Oct 27 15:44:35 2009 -0400
> +++ b/src/arch/x86/process.cc Tue Oct 27 16:13:49 2009 -0400
> @@ -715,7 +715,7 @@
> I386LiveProcess::getSyscallArg(ThreadContext *tc, int &i)
> {
> assert(i < NumArgumentRegs32);
> - return tc->readIntReg(ArgumentReg32[i]);
> + return tc->readIntReg(ArgumentReg32[i++]);
> }
>
> X86ISA::IntReg
> @@ -723,10 +723,10 @@
> {
> assert(width == 32 || width == 64);
> assert(i < NumArgumentRegs);
> - uint64_t retVal = ArgumentReg[i++] & mask(32);
> + uint64_t retVal = tc->readIntReg(ArgumentReg32[i++]) & mask(32);
> if (width == 64)
> - retVal |= ((uint64_t)ArgumentReg[i++] << 32);
> - return tc->readIntReg(retVal);
> + retVal |= ((uint64_t)tc->readIntReg(ArgumentReg[i++]) << 32);
> + return retVal;
> }
>
> void
> _______________________________________________
> m5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/m5-dev
>
Yes, thank you for fixing my fairly silly mistakes here. I should
refrain from working on M5 when I don't have the time to pay attention
properly.
Gabe
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
