Commen=
t # 26
on bug 66932<=
/a>
from Alex Deucher
(In reply to comment #=
24)
> (In reply to comment #22)
> > (In reply to comment #18<=
/a>)
> > > Created attachment 82559 [detai=
ls] [review] [review] [review] [review]
> > > weird fix
> > >=20
> >=20
> > Hmmm, looks like a compiler bug. what compiler are you using?
>=20
> gcc version 4.8.1 (Ubuntu/Linaro 4.8.1-6ubuntu1)=20
>=20
> But I think the compiler is right in this case. struct ATOM_INIT_REG_B=
LOCK
> doesn't represent what radeon_atom_init_mc_reg_table() is doing.
>=20
> The struct in the header reads:
> ...
> ATOM_INIT_REG_INDEX_FORMAT asRegIndexBuf[1];
> ATOM_MEMORY_SETTING_DATA_BLOCK asRegDataBuf[1];
>=20
> So reg_block->asRegIndexBuf[i] for i>0 is not defined, its within
> asRegDataBuf or even pass the size of the struct.
>=20
> It looks like this gcc is killing the while loop because i can never be
> anything else than zero.
It works on older version of gcc. I'm using 4.6.2 and 4.7.2 on several boxes
and haven't run into any problems. Also, even if the loop gets skipped, the
"last =3D " lines should still get printed.