Isn't Wifi supposed to be off by default?

Werner Almesberger werner at openmoko.org
Wed Jan 14 13:49:54 CET 2009


Andy Green wrote:
> Does the 2442 peripheral device support recognition of SDIO interrupt
> signalling action?  Where does that problem begin?

The S3C2442B54 does nominally support SDIO interrupts, yes. The S3C
MMC driver doesn't, but there is a patch by Christer Weinigel that
adds this functionality. Unfortunately, it only works on some chips
of the S3C family, while a few, including ours, act completely
erratically.

The S3C2442B54 has some glitches that are unique in the S3C family,
e.g., the off-by-one issues in SPI. Samsung say that the S3C MMC
controller is fine.

It gets better, though: Samuel's HIF for the Atheros SDIO stack
uses SDIO interrupts but does not exhibit those problems. None of
the differences I found between the two drivers seem to explain why
they shouldn't behave the same. Maybe it's a silicon malfunction
that is affected by supposedly unrelated elements, and we're just
unlucky enough to have hit that magic bit or whatever it is.

I plan to return to figuring out what's going on there once the
more pressing WLAN matters are handled. Sometimes it helps to take
a fresh look after a break.

By the way, here is some previous discussion of the issues:
Subject: Re: PATCH: s3cmci, add card interrupt support
on linux-arm-kernel at lists.arm.linux.org.uk and
Subject: Re: [PATCH 0/3] RFC: fix bug http://docs.openmoko.org/trac/ticket/1884
on openmoko-kernel.

- Werner



More information about the openmoko-kernel mailing list