> On June 3, 2012, 2:58 p.m., Steve Reinhardt wrote:
> > src/arch/x86/tlb.cc, line 273
> > <http://reviews.gem5.org/r/1243/diff/5/?file=27193#file27193line273>
> >
> >     I see that altAddr and defAddr are defined by Gabe's bitfield package, 
> > which I assume makes them unsigned.  Would it make more sense to change 
> > logSize to unsigned than to cast those fields to signed?
> 
> Nilay Vaish wrote:
>     IIRC, gcc's error output did not say anything related to signed vs 
> unsigned values.
>     It was about these two bitfields representing different parts of entire 
> thing. But
>     I'll leave it to Jayneel to confirm whether my recollection is correct.
> 
> Jayneel Gandhi wrote:
>     Yes, Nilay. You are right on that. gcc was not complaining about signed 
> vs unsigned. It was complaining that the typecast was missing since we are 
> extracting parts of a structure.
> 
> Jayneel Gandhi wrote:
>     This was the error that I got from gcc.
>     error: operands to ?: have different types 
> 'BitfieldBackend::RegularBitfieldTypes<long unsigned int>::Bitfield<15, 14>' 
> and 'BitfieldBackend::RegularBitfieldTypes<long unsigned int>::Bitfield<13, 
> 12>'
>     
>     @Steve:
>     I don't mind them making all of them unsigned int since logSize should 
> never be negative. Is that fine.

Thanks for the details, Jayneel.  I misunderstood the issue before.  I hadn't 
seen that error message previously.  It would be nice if there was a way to 
avoid the casting, but it would have to involve changing the Bitfield type 
somehow, and I don't know of an easy fix.

It probably is better to make everything unsigned.


- Steve


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/1243/#review2889
-----------------------------------------------------------


On May 30, 2012, 9:57 p.m., Jayneel Gandhi wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviews.gem5.org/r/1243/
> -----------------------------------------------------------
> 
> (Updated May 30, 2012, 9:57 p.m.)
> 
> 
> Review request for Default.
> 
> 
> Description
> -------
> 
> Changeset 9035:62ffaa5ca519
> ---------------------------
> TLB: Fix for gcc 4.4.3
> Due to recent changes to TLB, gem5 stopped compiling on gcc version 4.4.3. 
> This
> patch provides the fix for that problem. The patch is tested on gcc 4.4.3. The
> change is not required for more recent versions of gcc (like on 4.6.3).
> 
> 
> Diffs
> -----
> 
>   src/arch/x86/tlb.cc 8b9f227b64d8 
> 
> Diff: http://reviews.gem5.org/r/1243/diff/
> 
> 
> Testing
> -------
> 
> Compiles with gcc 4.4.3
> 
> 
> Thanks,
> 
> Jayneel Gandhi
> 
>

_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to