Re: hwcap supporting architectures?
GOTO Masanori <email@example.com> schrieb am 13.01.05 02:01:11:
> At Tue, 11 Jan 2005 11:27:28 +0100,
> Falk Hueffner wrote:
> > Marcelo E. Magallon <firstname.lastname@example.org> wrote:
> > > IIRC, alpha does not define any hwcaps.
> > There's a patch for this, which works fine, but wasn't committed yet:
> > http://lists.debian.org/debian-glibc/2004/03/msg00143.html
> > Sensible options are ev56 and ev67; ev5 is not particularly useful, since
> > it has the same instruction set as the baseline ev4, only different scheduling.
> > -mieee is default anyway on Debian's gcc.
> Yes, we don't support it currently, but I think it should be available
> after sarge.
> BTW, this patch does not enable HWCAP_IMPORTANT - so the answer of
> first question from Marcelo is: alpha does not support library HWCAP
> directory loading even if this patch is applied.
That's right. The reason is that the Alpha architecture doesn't have a set
of capabilities, from which each CPU version picks a random few, but rather
every CPU version is a proper superset of the preceding version. So I did
not use the orthogonal HWCAP model at all.
> If other libraries
> like mesa and libssl want to use /usr/lib/ev67 and so on, we may
> consider to add HWCAP_IMPORTANT.
This should not be needed, since the library loader also looks in a directory
corresponding to the architecture name. The only problem with this
occurs when you have for example an ev56 library in lib/ev56, and a ev67
CPU. Then the loader looks in lib/ev67 and then falls back to lib. Since
glibc is very carefully undocumented in this area , I didn't want to try to
change this, but rather assumed one could add a symlink.
 _dl_important_hwcaps is a great text book example on how not to do
comments, if anybody ever needs one.