On Wed, 2009-06-24 at 13:51 +0100, Alan Jenkins wrote: > On 6/24/09, Michal Marek <mmarek@xxxxxxx> wrote: > > Alan Jenkins napsal(a): > >> Michal Marek wrote: > >>> Jon Masters napsal(a): > >>> > >>>> Ok, I tested the different trees I was sent and then redid the master > >>>> tree just now. But I'm not finished fixing the breakage introduced in > >>>> merging the latest bits from Michal Marek - so give it a few hours as > >>>> I'm going to sleep now before I finish it up. > >>>> > >>> > >>> Ah, I didn't know you already merged them. I would happily rebase my > >>> patches on top of Andreas' and Alan's changes if you asked me to :). > >>> > >>> Michal > >>> > >> > >> I'm not sure that's the problem. The patches are applied to Jons tree, > >> but the tests fail (tests/runtests -vv). I bisected it to "elfops: > >> cleanup load_symbols()". > > > > It fails to find the exported symbol in tests/data/src/export_dep.c, > > which does it the old, pre-2.5.59 way (symbol names in the kernel_symbol > > struct directly). My patch removes support for that, so the failure > > isn't surprising here :-). Do we still care about such old kernels, or > > can I update the test? > > Go ahead. Jon removed back-compat for the "stable" 2.4 kernels, I'm > sure he's happy to drop cruft that only existed in the "unstable" 2.5 > series. > > And feel free to re-run the test to check for any other breakages :-). This will teach me to not read an entire thread before replying. Yeah, I do have the "secret knowledge" for building the modules - but it's not so secret really (they're just very simple ELF files built with a special section for mit), and it's probably better to just update the test modules so they have recent 2.6 style export (non-kernel_symbol). Oh, and I should be more clear next time when just the testsuite is breaking - but I thought I had, and it turns out I didn't. Like I said before, I'll poke at this myself too - yesterday I wound up having to fix my POWER5 box (power cycler blew up randomly, gave in and ordered a masterswitch for reliability) so I could build newer module files on it. Jon. -- To unsubscribe from this list: send the line "unsubscribe linux-modules" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
