Power: Change bitfield name to avoid conflicts with range_map

Information
Submitter:	Andreas Hansson
Repository:	gem5
Branch:	default
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

Power: Change bitfield name to avoid conflicts with range_map

This patch changes the name of a bitfield from W to W_FIELD to avoid
clashes with W being used as a class (typename) in the templatized
range_map. It also changes L to L_FIELD to avoid future problems. The
problem manifestes itself when the CPU includes a header that in turn
includes range_map.hh. The relevant parts of the decoder are updated.

Testing Done

util/regress all passing (disregarding t1000 and eio)

looks fine to me, maybe change L at the same time?
Ali

Andreas Hansson March 21, 2012, 3:37 a.m. (March 21, 2012, 3:37 a.m.)
```
That did indeed cross my mind :)
```

Diff:

Revision 2 (+8 -5)

Show changes

	src/arch/power/isa/bitfields.isa
	src/arch/power/isa/decoder.isa

I'm curious how this issue arises in practice... wouldn't the scope of the templated range_map methods be disjoint from where this code would appear?  I guess if it's a real issue with g++ 4.6 we need to fix it, but I'd like to understand better whether it's a true C++ error or just a g++ idiosyncrasy.

Andreas Hansson March 21, 2012, 6:53 p.m. (March 21, 2012, 6:53 p.m.)

It turns out I should not have blamed gcc, and particularly not 4.6. Sorry gcc.

As clang rightfully points out, the power o3 CPU includes decoder.hh that defines:
#define W       bits(machInst, 16, 16)

The CPU also includes system.hh, which as part of the distributed memory patch includes physical.hh, which includes range_map.hh.

In file included from build/POWER/mem/physical.hh:43:
build/POWER/base/range_map.hh:264:37: error: C++ requires a type specifier for
      all declarations
    insert(const Range<U> &r, const W d)
                                    ^
In file included from build/POWER/arch/power/o3_cpu_exec.cc:8:
build/POWER/arch/power/decoder.hh:78:16: note: instantiated from:
#define W       bits(machInst, 16, 16)
                     ^~~~~~~~

build/POWER/base/range_map.hh:264:37: error: declaration of 'machInst' shadows
      template parameter
build/POWER/base/range_map.hh:262:29: note: template parameter is declared here
    template <class U,class W>
                            ^

Description:

		Power: Change bitfield name to avoid conflicts with range_map

		This patch changes the name of a bitfield from W to W_FIELD to avoid
		clashes with W being used as a class (typename) in the templatized
~		range_map. The problem only manifested itself with gcc 4.6 after
~		seemingly unrelated changes, but is likely to cause similar issues in
~		the future and thus worth fixing. The relevant parts of the decoder
	~	range_map. It also changes L to L_FIELD to avoid future problems. The
	~	problem manifestes itself when the CPU includes a header that in turn
	~	includes range_map.hh. The relevant parts of the decoder are updated.
-		are updated.

Diff:

Revision 3 (+8 -5)

Show changes

	src/arch/power/isa/bitfields.isa
	src/arch/power/isa/decoder.isa

Thanks for digging in to the explanation... this makes sense.  Surprised we didn't run into it sooner.

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Diff:

Description:

Diff:

Status: Closed (submitted)