>>> Paul, Ben, given that Rusty hasn't come back with any opinion on this patch, >>do you >>> feel comfortable merging it via the ppc tree? Currently the earlyinit >routine >>> is only compiled in and used for your arch, so I think its fairly benign. >> >>Sorry, I *did* track down the archives for linuxppc-dev, then find your post, >>then read your patch. But I didn't actually reply. >> >>Other than minor issues, there's one significant one: you shouldn't be trying >>to change rodata. It might work on PPC today, but it's poor form at least. >> >>How's this? Untested on ppc. >> >I'll try grab a ppc64 system and test this soon. > >>Other changes: >>1) I also changed reloc_start to an array; this is a good idea for any >> linker-defined symbols so the compiler can't make assumptions about size. >Ok, I'll certainly trust your linker skill over mine :) > >>2) local.h? How about module.h? >Seems good to me >>3) I don't think the extra ". = 0" is necessary. >Its not, I was just trying to be clear about where reloc_start was to be >placed. > >>4) ARCH_USES_RELOC_ENTRIES isn't clear enough for me; I prefer >> ARCH_RELOCATE_KCRCTAB. >Sounds good to me. > > >I'll test this as soon as I'm able >Thanks! >Neil >
Just finished testing your patch Rusty, it all works quite well, Thanks! Will this be going in via your tree, or the ppc tree? Acked-by: Neil Horman <nhor...@tuxdriver.com> Neil _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
