This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [PATCH] ia64: unwind directive handling
- From: "Jan Beulich" <JBeulich at novell dot com>
- To: <cary at cup dot hp dot com>,<wilson at specifixinc dot com>
- Cc: <david_mosberger at hp dot com>,<binutils at sources dot redhat dot com>
- Date: Mon, 13 Jun 2005 01:06:52 -0600
- Subject: Re: [PATCH] ia64: unwind directive handling
>>> Cary Coutant <cary@cup.hp.com> 10.06.05 23:52:01 >>>
>> I checked the SCRA. Appendix B constains tables listing all of the
>> valid
>> unwind encodings. This does not include psp_psprel. I see that it
>> comes
>> from the .vframepsp directive. Curiously, the asm language manual has
>> two
>> typos in the vframepsp docs. It emits two unwind records, and both of
>> them
>> are mispelled. Maybe this is a late change that was never properly
>> documented.
>> How do you know that P8 r=0 is correct?
>
>Jim is right -- there is no psp_psprel unwind directive, and Format P8,
>r=0 is unassigned.
As I said, ias uses this, so it can't be completely invalid.
>If you think about it, storing the previous stack pointer at a location
>relative to psp is pointless at best, since you would need to know the
>psp in order to recover the psp!
Hmm, indeed. This, however, still doesn't mean no change is needed; .vframepsp itself is pointless, then, and support for it should be removed.
Jan