[PATCH] locomo.c: make irq register access typesafe
H Hartley Sweeten
hartleys at visionengravers.com
Mon Mar 8 18:22:49 EST 2010
On Wednesday, January 20, 2010 8:00 PM, H Hartley Sweeten wrote:
> On Wednesday, January 20, 2010 7:42 PM, Eric Miao wrote:
>> On Thu, Jan 21, 2010 at 10:31 AM, H Hartley Sweeten
>> <hartleys at visionengravers.com> wrote:
>>> On Wednesday, January 20, 2010 7:13 PM, Eric Miao wrote:
>>>> On Thu, Jan 21, 2010 at 2:10 AM, H Hartley Sweeten
>>>> <hartleys at visionengravers.com> wrote:
>>>>> In locomo_setup_irq a void __iomem * is stored by set_irq_chip_data
>>>>> to hold the base address for the irq registers. Unfortunately chip_data
>>>>> is a void *. This will cause a number of (different address space) sparse
>>>>> warnings.
>>>>>
>>>>
>>>> That's actually been fixed as a side effect in my previous proposed
>>>> irq cleanup patches. The way is to pass locomo chip as the data
>>>> instead of a raw pointer. Please let me know how you think, the
>>>> specific patch follows:
>>>
>>> There must be some other patches to locomo.c that remove the locomo_key_*,
>>> locomo_gpio_*, locomo_lt_*, and locomo_spi_* related stuff. Were all of
>>> them merged into common code? Where can I find a reference to the
>>> original patches?
>>>
>>
>> Cannot find the reference atm but they were posted weeks ago.
>
> Located them at:
>
> http://lists.infradead.org/pipermail/linux-arm-kernel/2009-December/thread.html
>
> I'll try to take a good look at them tomorrow and give you any feedback
> I have.
>
>>> Other than that it looks good to me.
>>
>> Thanks.
Eric,
Did your locomo irq cleanup patches get lost somewhere? I see three pending
patches in linux-next but they did not show up in Linus' tree during the
merge window.
commit da8065ac3ea29d1bfdb5a163be346fbe35066f8d
[ARM] locomo: avoid unnecessary cascaded keyboard IRQ
commit 00dd8027b913088ff9b656c5aaa6336c303b7f26
[ARM] locomo: remove unused IRQs and avoid unnecessary cascade
commit ac609d266e4af4ebf586d610bd76e04dddae0c4c
[ARM] locomo: allow cascaded IRQ base to be specified by platforms
Regards,
Hartley
More information about the linux-arm-kernel
mailing list