This is the mail archive of the binutils@sourceware.cygnus.com mailing list for the binutils project.


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

Re: All those old patches.


> Date: Thu, 25 May 2000 10:09:16 +0200
> From: Momchil Velikov <velco@fadata.bg>
> X-Accept-Language: en
> 
> Momchil Velikov wrote:
> > 
> > Geoff Keating wrote:
> > >
> > > > Date: Tue, 23 May 2000 10:10:10 +0200
> > > > From: Momchil Velikov <velco@fadata.bg>
> > > > http://sourceware.cygnus.com/ml/binutils/2000-04/msg00003.html
> > > >   This patch fixes a linker fatal signal.
> > >
> > > I don't like this patch.  Why initialize the table there?  There
> > > should be no need to initialize the table in multiple places all
> > > through elf32-ppc.c.  Instead, find a place before the table could
> > > ever be used, initialise it there, and delete the other
> > > initialisations.
> > 
> > OK. I'll look at that.
> 
> OK, I have looked at that ...
> I'd propose getting rid of `ppc_elf_howto_raw' and
> `ppc_elf_howto_init()'
> and statically initializing `ppc_elf_howto_table'. 

How do you propose to initialise ppc_elf_howto_table statically,
without encoding knowledge about the numbers the relocs have?

Otherwise, this would be good.  Statically initialised data is better.

> Alternatively, in addition to applying the patch
> we can remove the howto table initialization in
> `ppc_elf_relocate_section'
> -- resulting in the same count of invocasions of
> `ppc_elf_howto_init' -- at least not making stuff any worse.  

Are you sure that it's never possible for the initialisation you are
adding to be skipped?

-- 
- Geoffrey Keating <geoffk@cygnus.com>

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