[PATCH] ARM: hw_breakpoint: Clear breakpoints before enabling monitor mode

[sboyd@codeaurora.org (Stephen Boyd)]

On 09/24/12 10:19, Will Deacon wrote:
> On Thu, Sep 20, 2012 at 06:35:56PM +0100, Will Deacon wrote:
>> On Thu, Sep 20, 2012 at 05:57:40PM +0100, Stephen Boyd wrote:
>>> +/* Determine if halting mode is enabled */
>>> +static int halting_mode_enabled(void)
>>> +{
>>> +	u32 dscr;
>>> +
>>> +	ARM_DBG_READ(c1, 0, dscr);
>>> +
>>> +	if (WARN_ONCE(dscr & ARM_DSCR_HDBGEN,
>>> +		"halting debug mode enabled. Unable to access hardware resources.\n")) {
>>> +		return -EPERM;
>>> +	}
>>> +	return 0;
>>> +}
>> ...it looks like debug arch 7.1 defines this bit as UNKNOWN when the OS lock
>> is clear, so we probably shouldn't be reading it at all. I'll pour myself a
>> stiff drink and start reading the debug arch docs to work out what on Earth
>> we should do.
>>
>> Stay tuned.
> Ok, I've pushed a bunch of patches to my hw-breakpoint branch (head commit
> 55cb726797c7). I'll post them to the list after the merge window, but please
> do take them for a spin if you get a chance.
>

Sure, I'll try them later today. I would say just send them out so
people can add tested and reviewed tags. Otherwise we should all start
planning week long vacations every 7 to 8 weeks to coincide with the
merge window.

Also, it would be nice if we could fix this in 3.7 or even 3.6. Booting
on an MSM8660 is very fragile right now and this patch fixes it for me.

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by The Linux Foundation