Re: [PATCH 2/2] locking/rwsem: Disable optimistic spinning for PA-RISC

[James Bottomley]

On Fri, 2014-06-06 at 10:57 -0700, James Bottomley wrote:
> On Fri, 2014-06-06 at 19:53 +0200, Peter Zijlstra wrote:
> > On Fri, Jun 06, 2014 at 10:22:21AM -0700, Linus Torvalds wrote:
> > > On Fri, Jun 6, 2014 at 10:11 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> > > >
> > > > +config ARCH_NO_ATOMIC_RMW
> > > > +       def_bool y
> > > > +       depends on PARISC || SPARC32 || METAG_ATOMICITY_LOCK1 || (TILE && !TILEGX) || (ARC && !ARC_HAS_LLSC)
> > > 
> > > Ugh. We've had these kinds of things before, and they are broken and
> > > nasty to maintain.
> > > 
> > > Just make it
> > > 
> > >     config ARCH_SUPPORTS_ATOMIC_RMW
> > >         bool
> > > 
> > > which defaults to no. And then make MUTEX_SPIN_ON_OWNER depend on that.
> > 
> > ---
> > Subject: locking, mutex: Make optimistic spinning depend on ARCH_SUPPORTS_ATOMIC_RMW
> > 
> > The optimistic spin code assumes regular stores and cmpxchg() play nice;
> > this is found to not be true for at least: parisc, sparc32, tile32,
> > metag-lock1, arc-!llsc and hexagon.
> > 
> > There is further wreckage, but this in particular seemed easy to
> > trigger, so blacklist this.
> > 
> > Opt in for known good archs.
> > 
> > Reported-by: Mikulas Patocka <mpatocka@xxxxxxxxxx>
> > Signed-off-by: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> > ---
> >  arch/arm/Kconfig     | 1 +
> >  arch/arm64/Kconfig   | 1 +
> >  arch/powerpc/Kconfig | 1 +
> >  arch/sparc/Kconfig   | 1 +
> >  arch/x86/Kconfig     | 1 +
> >  kernel/Kconfig.locks | 5 ++++-
> >  6 files changed, 9 insertions(+), 1 deletion(-)
> > 
> > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> > index a2c1a18a7275..b8a6a5078eae 100644
> > --- a/arch/arm/Kconfig
> > +++ b/arch/arm/Kconfig
> > @@ -6,6 +6,7 @@ config ARM
> >  	select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST
> >  	select ARCH_HAVE_CUSTOM_GPIO_H
> >  	select ARCH_MIGHT_HAVE_PC_PARPORT
> > +	select ARCH_SUPPORTS_ATOMIC_RMW
> >  	select ARCH_USE_BUILTIN_BSWAP
> >  	select ARCH_USE_CMPXCHG_LOCKREF
> >  	select ARCH_WANT_IPC_PARSE_VERSION
> > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> > index 5c30b10c0f6c..339dc07f15e2 100644
> > --- a/arch/arm64/Kconfig
> > +++ b/arch/arm64/Kconfig
> > @@ -3,6 +3,7 @@ config ARM64
> >  	select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE
> >  	select ARCH_USE_CMPXCHG_LOCKREF
> >  	select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST
> > +	select ARCH_SUPPORTS_ATOMIC_RMW
> >  	select ARCH_WANT_OPTIONAL_GPIOLIB
> >  	select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
> >  	select ARCH_WANT_FRAME_POINTERS
> > diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
> > index e0998997943b..5bb96f282373 100644
> > --- a/arch/powerpc/Kconfig
> > +++ b/arch/powerpc/Kconfig
> > @@ -145,6 +145,7 @@ config PPC
> >  	select HAVE_IRQ_EXIT_ON_IRQ_STACK
> >  	select ARCH_USE_CMPXCHG_LOCKREF if PPC64
> >  	select HAVE_ARCH_AUDITSYSCALL
> > +	select ARCH_SUPPORTS_ATOMIC_RMW
> >  
> >  config GENERIC_CSUM
> >  	def_bool CPU_LITTLE_ENDIAN
> > diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
> > index 29f2e988c56a..407c87d9879a 100644
> > --- a/arch/sparc/Kconfig
> > +++ b/arch/sparc/Kconfig
> > @@ -78,6 +78,7 @@ config SPARC64
> >  	select HAVE_C_RECORDMCOUNT
> >  	select NO_BOOTMEM
> >  	select HAVE_ARCH_AUDITSYSCALL
> > +	select ARCH_SUPPORTS_ATOMIC_RMW
> 
> Sparc64 does, Sparc32 doesn't, so shouldn't that be
> 
> select ARCH_SUPPORTS_ATOMIC_RMW if 64BIT
> 
> ?
> 
> Other than this, looks good to me (you can add reviewed by or acked by
> from me to your taste).

Oh, wait, I see how sparc does it, never mind ...

James


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/