[PATCH 1/2]: export-irq-functionality.patch

Andy Green andy at openmoko.com
Sat Nov 15 19:37:56 CET 2008


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:

|>> But it's a problem to solve if and when we concentrate on getting
|>> lis302dl upstream, difficult as that may anyway be, so I sent both of
|>> these on stable-tracking.
|> Thanks. I'm also slightly worried about the upstreaming. Although I'm
|> hardy an expert, I really see no other way than extending the SPI
|> interface with some "raw" interface that can be used from interrupt
|> context. That looks to me like a daunting task.

| I don't know it if this could be usefull but there's been a submission
| of an accelerometer driver (ST LIS3LV02Dx) for mainline inclusion, see
[1].

| [1] - http://lkml.org/lkml/2008/10/18/144

Thanks for the tip... I guess this explains why that driver doesn't have
the difficulty around SPI IO in interrupt context...

''* Andrew Morton asked to investigate using interrupt-based sensor
reading (instead of timer-based): it turns out not useful because the
sensor generate interrupts even if the data has not changed, so we end
up waking up at 40Hz''

I think it is useful alright to have a reliably constant period between
samples, it just makes trouble for IO subsystem that can't operate in
interrupt context; the threshold stuff can help with the absolute amount
of samples.

- -Andy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkkfFwQACgkQOjLpvpq7dMp6eQCeLjXtEZczK0B8W9KpTEdxvcID
Su8AniSQP9qvTyYdrq61f0zx03TnBxjw
=B/vn
-----END PGP SIGNATURE-----



More information about the openmoko-kernel mailing list