[UPSTREAM] Attempt at improving ADC access
balajirrao at openmoko.org
Thu Oct 16 11:25:23 CEST 2008
On Thu, Oct 16, 2008 at 09:36:27AM +0100, Andy Green wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> Somebody in the thread at some point said:
> | On Thu, Oct 16, 2008 at 12:18:50AM +0100, Andy Green wrote:
> |> -----BEGIN PGP SIGNED MESSAGE-----
> |> Hash: SHA1
> |> Somebody in the thread at some point said:
> |> | Also, don't the queue head/tail manipulation functions require
> locking ?
> |> When I wrote them I recall coming to the idea that no locking is needed
> |> because each of the "threads" only controls the head or the tail
> |> exclusively.
> | Hmm.. I don't understand. Since the queue update statement is not
> | atomic, it could be preempted anywhere 'in between' by an interrupt
> | handler which might also want to add to the queue. This can lead
> | to problems. Isn't it? Is there any reason why this can't happen ?
> | Please explain.
> The explanation is you are quite right... it should lock interrupts in
> the add to queue thing. Maybe you can take care of this?
Thank you for clarifying this. I'll take care of it.
More information about the openmoko-kernel