[gcc] Yet another gen_lowpart code confused by prt_mode != Pmode

[gcc] Yet another gen_lowpart code confused by prt_mode != Pmode

Test from David Meyer, 64-bit mode with -O1:

typedef unsigned int size_t;

extern "C" {
  size_t strlen (const char *);
};

class StringRef {
private:
  const char *Data;
  size_t Length;

public:
 StringRef(const char *Str)
    : Data(Str), Length(::strlen(Str)) {}
};

StringRef GetDLLSuffix() {
  return &(".so"[1]);
}

Reports internal error while trying to replace Pmode register with ptr_mode CONST.

The change fixes the origin of such a CONST in gen_lowpart_for_combine. This is kinda hacky and clearly not upstreamable...

BUG=http://code.google.com/p/nativeclient/issues/detail?id=1279
TEST=see above

Committed: http://git.chromium.org/gitweb/?p=nacl-toolchain.git;a=commit;h=ffc2ce5

[khim, 2010-12-24 14:38:17]

Well, this is clearly not a good fix, but let's commit it to unblock other
teams.

Please add BUG to the description, then commit.

[khim, 2010-12-24 15:17:20]

LGTM

[pasko-google - do not use, 2010-12-24 16:13:09]

http://codereview.chromium.org/5968010/diff/2001/gcc/gcc/combine.c
File gcc/gcc/combine.c (right):

http://codereview.chromium.org/5968010/diff/2001/gcc/gcc/combine.c#newcode9895
gcc/gcc/combine.c:9895: /* HACK?
"TODO(eaeltsin)" would be more stylish than "HACK?"