Re: appchk/libc functions
On Wed, 11 Apr 2001, Dave Prosser wrote:
> It's one of the "new" sections added for IA-64. The s stands for
> "short". There's also .sdata, .sdata1, and so on. It's helpful for
> this architecture as there is a substantial cut off in register+offset
> addressing at around 22-24 bits, which is clearly much smaller than
> the full 64 bit general addressing.
I figured it was something like that, but he is seeing it on an IA32 system.
> But, why does appchk care what names are used for ELF sections?
They are listed in the specs, so it checks. What does the linker/loader do
with a section, if it doesn't recognize it?
> It's the attributes that are most important, since they are what's
> used to determine how allocated sections are merged into segments.
I'm still waiting for the coffee to kick in, but do the attributes alone
fully define the semantics of a section? IF I'm thinking of the right thing,
the attributes aren't unique across sections, so it would somewhat ambiguous
to depend on them alone.
Stuart R. Anderson email@example.com
Metro Link Incorporated South Carolina Office
5807 North Andrews Way 129 Secret Cove Drive
Fort Lauderdale, Florida 33309 Lexington, SC 29072
voice: 954.660.2500 voice: 803.951.3630
http://www.metrolink.com/ XFree86 Core Team
Creative Applications Lab Chair - SIGGRAPH 2001