This is the mail archive of the libc-alpha@sources.redhat.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: The usage of _dl_hwcap and dl_hwcap_mask is confusing


On Tue, Jun 24, 2003 at 10:12:21PM -0700, Roland McGrath wrote:
> > But dl_hwcap_mask is NOT set based on the hardware the binary is running
> > on. It is a compile-time constant. It can only be overridden via an env
> > variable. It seems to me that checking _dl_hwcap is more appropriate here.
> > Did I miss something?
> 
> Oh, I see.  My eyes fixed the wrong variable name before it got to my brain.
> That is definitely wrong and should be fixed.  It should in fact be 
> (GL(dl_hwcap) & GL(dl_hwcap_mask) & HWCAP_I386_XMM) that's tested, I think.
> 

Why shouldn't glibc support SSE when it is the part of the hardware?
Should a user application be allowed to use SSE when it is available?
I think it should be (GL(dl_hwcap) & HWCAP_I386_XMM).


H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]