[PATCH v3 0/6] ARM: Remove the __ARCH_WANT_INTERRUPTS_ON_CTXSW definition

[Will Deacon]

On Mon, Jan 23, 2012 at 09:53:38AM +0000, Catalin Marinas wrote:
> On Mon, Jan 23, 2012 at 09:47:59AM +0000, Russell King - ARM Linux wrote:
> > On Mon, Jan 23, 2012 at 09:26:24AM +0000, Catalin Marinas wrote:
> > > On Mon, Jan 23, 2012 at 09:15:31AM +0000, Peter Zijlstra wrote:
> > > > On Fri, 2012-01-20 at 17:42 +0000, Catalin Marinas wrote:
> > > > > This is version 3 of the set of patches removing
> > > > > __ARCH_WANT_INTERRUPTS_ON_CTXSW on ARM. The series was rebased on top of
> > > > > 3.3-rc1 and fixed the conflicts with the kernel/sched/ changes and the
> > > > > ARM LPAE patches. There are no functional changes from v2. I plan to
> > > > > push this to -next and get it ready for 3.4-rc1.
> > > > > 
> > > > > Question for Peter/Ingo - how do we merge the first patch that
> > > > > introduces finish_arch_post_lock_switch? Do you pick it up or I can
> > > > > merge it via rmk (with your ack)?
> > > > 
> > > > I'm fine either way, I'll probably ask Ingo to pull your tree so that I
> > > > can stack some other patches on top.
> > > 
> > > In which case I would need Russell's acked-by.
> > 
> > That depends on knowing what CPU architectures this has been tested on,
> > and whether anyone external has tested it.  It's definitely a change
> > which needs some tested-by tags on it.
> 
> I agree. On my side, I tested it on:
> 
> Versatile Express + Cortex-A9 (SMP configuration, ASIDs)
> Versatile PB926 (UP configuration, no ASIDs)

For what it's worth, I've also tested this on:

Realview PB1176 (UP, ASIDs)
Cortex-A5 (SMP, ASIDs)
Cortex-A7 (SMP, ASIDs)

and I haven't seen any problems with native, parallel kernel builds. I don't
have anything prior to ARMv5 available, but at least we seem to have covered
v5-v7.

Will