HAVE_CMPXCHG_LOCAL arch support.

HAVE_CMPXCHG_LOCAL arch support.

[Ralf Baechle]

On Wed, Jun 10, 2015 at 04:58:04PM +0200, Ralf Baechle wrote:

> On Mon, Jun 08, 2015 at 03:14:43PM +0200, Xose Vazquez Perez wrote:
> 
> > If there is anything wrong, please report it in this thread:
> > https://marc.info/?t=143332955700003
> 
> 
> >    locking/ cmpxchg-local        : TODO |                  HAVE_CMPXCHG_LOCAL #  arch supports the this_cpu_cmpxchg() API
> 
> This one was easy - we have the functions in the code just no "select
> HAVE_CMPXCHG_LOCAL" Kconfig.

Something's wrong there.  The new file
Documentation/features/locking/cmpxchg-local/arch-support.txt in linux-next
claims correctly that only s390 and x86 define HAVE_CMPXCHG_LOCAL.  And a
git grep -w cmpxchg_local finds that in addition to these alpha, arm, arm64,
avr32, blackfin, c6x, frv, ia64, m32r, m68k, mips, parisc, powerpc, sparc,
unicore32 and xtensa define cmpxchg_local.

These architectures seem to not define cmpxchg_local in their arch/ dir:

  arc cris hexagon metag microblaze mn10300 nios2 openrisc score sh tile um

Microblaze and nios2 include <asm-generic/cmpxchg.h> into their arch
cmpxchg.h so they get a definition of these functions but don't define
HAVE_CMPXCHG_LOCAL.  Peter Zijlstra said it the local versions are ~ 20
cycles faster on x86 than the "global" version.  But I've found one user
of cmpxchg_local, mm/vmstat.c and one user of cmpxchg64_local,
drivers/iommu/intel-iommu.c.  Sure, fixing the issue was trivial for me
on MIPS but is having cmpxchg{,64}_local actually worth it?  

Cheers,

  Ralf

Re: HAVE_CMPXCHG_LOCAL arch support.

[Peter Zijlstra]

On Wed, 2015-06-10 at 17:57 +0200, Ralf Baechle wrote:
> On Wed, Jun 10, 2015 at 04:58:04PM +0200, Ralf Baechle wrote:
> 
> > On Mon, Jun 08, 2015 at 03:14:43PM +0200, Xose Vazquez Perez wrote:
> > 
> > > If there is anything wrong, please report it in this thread:
> > > https://marc.info/?t=143332955700003
> > 
> > 
> > >    locking/ cmpxchg-local        : TODO |                  HAVE_CMPXCHG_LOCAL #  arch supports the this_cpu_cmpxchg() API
> > 
> > This one was easy - we have the functions in the code just no "select
> > HAVE_CMPXCHG_LOCAL" Kconfig.
> 
> Something's wrong there.  The new file
> Documentation/features/locking/cmpxchg-local/arch-support.txt in linux-next
> claims correctly that only s390 and x86 define HAVE_CMPXCHG_LOCAL.  And a
> git grep -w cmpxchg_local finds that in addition to these alpha, arm, arm64,
> avr32, blackfin, c6x, frv, ia64, m32r, m68k, mips, parisc, powerpc, sparc,
> unicore32 and xtensa define cmpxchg_local.
> 
> These architectures seem to not define cmpxchg_local in their arch/ dir:
> 
>   arc cris hexagon metag microblaze mn10300 nios2 openrisc score sh tile um
> 
> Microblaze and nios2 include <asm-generic/cmpxchg.h> into their arch
> cmpxchg.h so they get a definition of these functions but don't define
> HAVE_CMPXCHG_LOCAL.  Peter Zijlstra said it the local versions are ~ 20
> cycles faster on x86 than the "global" version.  But I've found one user
> of cmpxchg_local, mm/vmstat.c and one user of cmpxchg64_local,
> drivers/iommu/intel-iommu.c.  Sure, fixing the issue was trivial for me
> on MIPS but is having cmpxchg{,64}_local actually worth it?  

If you traverse the obfuscation chain a little, you'll find that
local_cmpxchg() is implemented using cmpxchg_local(), and there are a
few more users of that.

Re: Generic kernel features that need architecture(mips) support

[Xose Vazquez Perez]

On 06/10/2015 04:58 PM, Ralf Baechle wrote:

> How are the documentation files in Documentation/features/ maintained?
> They were automatically generated so I wonder if I have to take care
> of anything.

CC: Ingo and related ml.

Re: Generic kernel features that need architecture(mips) support

[Ingo Molnar]

(Jon Cc:-ed)

* Xose Vazquez Perez <xose.vazquez@gmail.com> wrote:

> On 06/10/2015 04:58 PM, Ralf Baechle wrote:
> 
> > How are the documentation files in Documentation/features/ maintained?
> > They were automatically generated so I wonder if I have to take care
> > of anything.
> 
> CC: Ingo and related ml.

So changes to Documentation/features/ should come as simple patches done from hand 
editing, there's no need to preserve any initial (half-)automated generation. 
Formatting should be preserved so that Documentation/features/list-arch.sh still 
works as before.

Jon: would you like to receive all patches to Documentation/features/ so you can 
collect them in the documentation tree, or can maintainers patch it as part of any 
feature work that affects the tables? I think it's all finegrained enough to not 
create conflicts. That way they would become partly self-maintaining. (Or at least 
one can always hope! ;-)

Thanks,

	Ingo

Re: Generic kernel features that need architecture(mips) support

[Jonathan Corbet]

On Thu, 11 Jun 2015 08:24:53 +0200
Ingo Molnar <mingo@kernel.org> wrote:

> Jon: would you like to receive all patches to Documentation/features/ so you can 
> collect them in the documentation tree, or can maintainers patch it as part of any 
> feature work that affects the tables? I think it's all finegrained enough to not 
> create conflicts. That way they would become partly self-maintaining. (Or at least 
> one can always hope! ;-)

I sort of figured I'd handle it the way I do the rest of Documentation/ —
I'm happy to herd the patches, but I also have no problem staying out of
the way when other maintainers reach into it.

jon