This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [patch] new sh relocs
- From: Alexandre Oliva <aoliva at redhat dot com>
- To: DJ Delorie <dj at redhat dot com>
- Cc: amodra at bigpond dot net dot au, ian at airs dot com, binutils at sources dot redhat dot com
- Date: 10 Sep 2003 04:56:34 -0300
- Subject: Re: [patch] new sh relocs
- Organization: GCC Team, Red Hat
- References: <200308152051.h7FKp5b14273@greed.delorie.com><m3k79e41t2.fsf@gossamer.airs.com><200308161338.h7GDcj618581@greed.delorie.com><20030816135520.GL27145@bubble.sa.bigpond.net.au><or3cg139xw.fsf@free.redhat.lsd.ic.unicamp.br><200309060007.h86077V14026@greed.delorie.com>
On Sep 5, 2003, DJ Delorie <dj@redhat.com> wrote:
>> Unfortunately we're the ones that got the ABI wrong, and Renesas tools
>> don't emit any EI_* markers we could use to tell, so we should just
>> take the risk and change the meaning of relocs we've been generating
>> with the wrong meaning.
> So where are we on this patch? Can I commit it, or what?
Since Renesas pretty much told us to decide which numbers to use
within the 0xa-0x20 reserved range, here are the changes I'd like us
to make:
33 (0x21) R_SH_SWITCH8 => move to 0x18, just before R_SH_SWITCH16
34 (0x22) R_SH_GNU_VTINHERIT => remove
35 (0x23) R_SH_GNU_VTENTRY => remove
The C++ vtable GCing feature has been removed from GCC, there's no
point in allocating official relocs (or using using relocs from our
reserved range) for this purpose.
36 (0x24) R_SH_LOOP_START => move to 0x16 or 0xa, your call
37 (0x25) R_SH_LOOP_END => move to 0x17 or 0xb, your call
I used to like 0x16 and 0x17 better, but now I'm thinking using 0xa
and 0xb for GNU extensions would more clearly mark the region reserved
for us.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist Professional serial bug killer