Re: [PATCH 01/13] OMAP: Introduce a user list for each voltage domain instance in the voltage driver.

[Nishanth Menon <nm@ti.com>]

Felipe Balbi had written, on 09/02/2010 05:28 AM, the following:
> Hi,
> 
> On Thu, 2 Sep 2010 05:17:01 -0500, Nishanth Menon <nm@ti.com> wrote:
>> note - if we allow unlock of irqs at this point, we cannot predictably 
>> progress down the logic.
> 
> spin_unlock() would not re-enable IRQs, would it ? Isn't it so that
> spin_unlock_irq() would be the one re-enabling IRQ ?
> 
oopss.. my bad.. if we were to do regulator based implementation of 
voltage framework, looking closer at the code, driver/regulator/core.c 
-> rdev->mutex is held for set_voltage, set_mode and all entry functions 
for regulator operations -> this would be the only concern i have.. I 
may be barking up the wrong tree here, but i think if i read 
Documentation/mutex-design.txt right, "contexts such as tasklets and 
timers" and "mutexes may not be used in hardware or software interrupt" 
means to me dont do this in irq locked context such as the sitn in 
omap_sram_idle?

-- 
Regards,
Nishanth Menon