This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH/RFA] SH: DWARF2 exception handling
- From: Joern Rennecke <joern dot rennecke at superh dot com>
- To: kaz Kojima <kkojima at rr dot iij4u dot or dot jp>
- Cc: gcc-patches at gcc dot gnu dot org, aoliva at redhat dot com, stephen dot clarke at earthling dot net
- Date: Fri, 15 Nov 2002 13:41:25 +0000
- Subject: Re: [PATCH/RFA] SH: DWARF2 exception handling
- Organization: SuperH UK Ltd.
- References: <3DD3D528.8312CE0D@superh.com> <200211150105.gAF15eO10893@r-rr.iij4u.or.jp>
kaz Kojima wrote:
> Joern Rennecke <joern.rennecke@superh.com> wrote:
> > So, is sh_set_return_address, after calling calc_live_regs, you
Oops, that should have read 'in sh_set_return_address'.
> > can just check
> > if (live_regs_mask[pr_reg/32] & (1 << (pr_reg % 32)))
> > to decide if you should write directly into the pr register,
> > or if you need to store into a stack slot.
>
> If PR / PR_MEDIA_REG isn't life, then the epilogue returns with
> rts / blink trN, r63 without popping pr_reg. Right? Does it mean
> that we can set PR / (the result of sh_media_register_for_return)
> to RA directly in sh_set_return_address in such case, without
> using any stack slots?
Basically, yes. The epilogue for sh[1-4] / SHcompact uses rts,
as you said, and SHmedia uses ptabs r18,tr0; blink tr0,r63 .
[For the bystanders: PR_MEDIA_REG is r18.]
--
--------------------------
SuperH (UK) Ltd.
2410 Aztec West / Almondsbury / BRISTOL / BS32 4QX
T:+44 1454 465658