Messages in this thread | | | From | Alexander Stein <> | Subject | Re: [PATCH v3] pch_gpio: Support interrupt function | Date | Wed, 22 Jun 2011 11:15:31 +0200 |
| |
On Wednesday 22 June 2011 10:14:10 Tomoya MORINAGA wrote: > (2011/06/22 15:28), Alexander Stein wrote: > > On Wednesday 22 June 2011 02:10:28 Tomoya MORINAGA wrote: > >> (2011/06/21 18:10), Alexander Stein wrote: > >>> I wrote a small program which use poll(2) on value, after edge hase > >>> been configured from shell. I cannot test the edge detection > >>> appropriately due to debouncing using jumpers. > >> > >> Really ? > >> In out test environment, my test program could detect edge interrupt > >> correctly. > > > > My test program used poll/select on edge successfully. But I could not > > test the raising or falling feature actually, as my input is bouncing. > > So, it doesn't matter if I set both/raising/falling in edge, due to my > > inputs the effect is the same. Any input event will end in several > > raising and falling interrupts events. > > I hope I could make this more clear. > > Our test program also uses poll (We don't use select().)
I noticed poll/select work nearly as the same. Both worked fine here.
> BTW, is the above your saying related to your patch "[PATCH] pch_gpio: > transform mutex into spinlock" ?
Should be unrelated. I did the GPIO interrupt test before stumbling on the mutex/spinlock problem. The latter one was caused by setting a GPIO from interrupt context (in a different driver).
Alexander
| |