Re: [Patch 1/1] PPC64-HWBKPT: Implement hw-breakpoints for PPC64

["K.Prasad" <prasad@linux.vnet.ibm.com>]

On Wed, Apr 07, 2010 at 06:03:31PM +1000, Benjamin Herrenschmidt wrote:
> Ok so too many problems with your last patch, I didn't have time to fix
> them all, so it's not going into -next this week.
> 
> Please, test with a variety of defconfigs (iseries, cell, g5 for
> example), and especially with CONFIG_PERF_EVENTS not set. There are
> issues in the generic header for that (though I'm told some people are
> working on a fix).
> 
> Basically, we need something like CONFIG_HW_BREAKPOINTS that is set
> (maybe optionally, maybe not) with CONFIG_PERF_EVENTS and
> CONFIG_HAVE_HW_BREAKPOINT. Then you can use that to properly fix the
> ifdef'ing in include/linux/hw_breakpoints.h, and fix the various other
> cases in powerpc code where you are testing for the wrong thing.
> 
> Cheers,
> Ben.
>

Hi Ben,
	I've sent a new patch (linuxppc-dev message-id ref:
20100414034340.GA6571@in.ibm.com) that builds against the defconfigs for
various architectures pointed out by you (I did see quite a few errors
that aren't induced by the patch).

The source tree is buildable even without CONFIG_PERF_EVENTS, and is
limited to scope using CONFIG_HAVE_HW_BREAKPOINT. At this stage, I
didnot find a need for a seperate CONFIG_HW_BREAKPOINTS though.

Let me know what you think.

Thanks,
K.Prasad