remove implicit slab.h inclusion from percpu.h
in [Kernel]
|
From: Alexey Dobriyan on 11 Mar 2010 12:50 On Thu, Mar 11, 2010 at 11:56:22PM +0900, Tejun Heo wrote: > Lee Schermerhorn was trying to use percpu from slab.h and ran into a > dependency loop. percpu.h was using slab.h for UP inline > implementation which isn't a big deal in itself but it turns out that > percpu.h ends up being included everywhere via module.h and sched.h. > So, removing that implicit inclusion breaks a lot of files. > > The following git tree contains trial conversion on x86_64. > allmodconfig builds fine on it but a lot of other archs are likely to > break although fixing them up shouldn't be too hard. > > git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git review-remove-implied-slab-inclusion > > This can be solved much easier by moving some of the stuff that's > necessary for slab.h from percpu.h into percpu-defs.h which originally > got separated so that it can be used by asm/percpu.h but it's hackish > and for longer term, it would be better to have slab.h explicitly > included where necessary. > > So, what do you guys think? Probably-the-right-thing-to-do >544 file > patch or somewhat-ugly-but-let's-worry-about-it-tommorrow two file > patch? You can include slab.h only for UP case. Since everyone tests on allmodconfig which has SMP=y, configuration will be more strict wrt headers, and compile breakages amount negligible. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Tejun Heo on 11 Mar 2010 17:40 Hello, On 03/12/2010 02:48 AM, Alexey Dobriyan wrote: >> So, what do you guys think? Probably-the-right-thing-to-do >544 file >> patch or somewhat-ugly-but-let's-worry-about-it-tommorrow two file >> patch? > > You can include slab.h only for UP case. > > Since everyone tests on allmodconfig which has SMP=y, configuration > will be more strict wrt headers, and compile breakages amount > negligible. But that wouldn't change anything about having to do an oneshot huge change, right? And, if we're gonna do that anyway, I think it would be better to remove the implicit dependency for UP case too so that for example slab.h in this case doesn't have to do ifdef on CONFIG_SMP before using percpu accessors. Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Tejun Heo on 16 Mar 2010 00:30 On 03/11/2010 11:56 PM, Tejun Heo wrote: > So, what do you guys think? Probably-the-right-thing-to-do >544 file > patch or somewhat-ugly-but-let's-worry-about-it-tommorrow two file > patch? Ping, any opinions? Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Tejun Heo on 16 Mar 2010 03:00 Hello, Ingo. On 03/16/2010 03:17 PM, Ingo Molnar wrote: > ( /me mumbles something about not having a patch in the email to review and > pulling the tree. 200k patch is just fine for lkml - i've attached it below > for easier review. percpu.h and percpu.c has the meat of the changes. ) I wanted to keep the discussion high level while giving a general idea about the extent of necessary changes. I'll include the patch from now on. > i like the dependency reduction. Noticed one small detail: > > this new 2000-lines #ifdef block percpu.c: > > +#ifdef CONFIG_SMP > +#else /* CONFIG_SMP */ > +#endif /* CONFIG_SMP */ > > feels a bit lame. A separate percpu_up.c file would be nicer i suppose? Sure. > Also, why should we make this opt-in and expose a wide range of configs to > build breakages? A more gradual approach would be to write a simple script > that adds a slab.h include to all .c's that include percpu.h, directly or > indirectly. > > You can map the pattern experimentally: the insertion pattern could be built > from the x86 allmodconfig build you did [i.e. extend the pattern until you > make it build on allmodconfig] - that would cover most cases in practice (not > just allmodconfig) - and would cover most architectures as well. I don't really get the 'experimental' part but if I count all the files which ends up including percpu.h directly or indirectly on allmodconfig it ends up including much more .c files than necessasry - 11203 to be exact, ~20 times more than necessary. Inclusions from .c files definitely are much less troublesome so the situation would be better than now but we'll still end up with a LOT of bogus inclusions without any good way to eventually remove them. Maybe a better way is to grab for slab API usages in .c files which don't have slab.h inclusion. If breaking the dependency is the way to go, I can definitely write up some scripts and do test builds on some archs. There sure will be some fallouts but I think it won't be too bad. Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Pekka Enberg on 16 Mar 2010 03:00 On Tue, Mar 16, 2010 at 8:17 AM, Ingo Molnar <mingo(a)elte.hu> wrote: > Also, why should we make this opt-in and expose a wide range of configs to > build breakages? A more gradual approach would be to write a simple script > that adds a slab.h include to all .c's that include percpu.h, directly or > indirectly. I am all for untangling the #include mess in slab.h and friends and I think Ingo's suggestion is probably the best way forward. We should avoid creating tree-wide breakage for this kind of cleanups. Pekka -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
|
Next
|
Last
Pages: 1 2 3 4 5 Prev: GFS2: Pull request (fixes) Next: [PATCH] Drivers: pci: code style cleanup |