[PATCH] [POWERPC] Optimize counting distinct entries in therelocation sections

Rusty Russell rusty at rustcorp.com.au
Tue Nov 13 15:27:39 EST 2007


On Tuesday 13 November 2007 13:49:15 Paul Mackerras wrote:
> Medve Emilian writes:
> > Seems like there are R_PPC_REL24 with r_addend != 0. Within a set of 41
> > modules (featuring 5457 R_PPC_REL24 relocations) already included within
> > the kernel tree I found 37 such relocations (R_PPC_REL24) with r_addend
> > != 0. In my test case, from 35K relocations, 7K are R_PPC_REL24 and from
> > those only 8 have r_addend != 0.
>
> I did a quick scan and the ones with r_addend != 0 all seem to be
> references to .text from the .init.text, .exit.text or .fixup
> sections.  Assuming we can get those allocated near each other they
> shouldn't need trampolines.
>
> Rusty, do we manage to put .init.text and .fixup near .text?

Definitely not.  That's why we trampoline between them.  But since we discard 
the init sections and the tramps with them, I wouldn't bother uniquifing 
them: just alloc that many.

> Also, do you know what we see in r_info for a relocation that is
> relative to a section rather than a symbol?

Can't remember off the top of my head, sorry.

Rusty.



More information about the Linuxppc-dev mailing list