On Wed, Jan 30, 2013 at 05:32:32PM -0500, Solomon Peachy wrote: > This is my current to-do list: Was unexpectedly able to spend some time on this today! > * One final Kconfig option to eliminate Done. > * A possible crash (reported not three hours ago) that needs confirming Turns out to not be a crash, but it is a real problem. The driver can get into an inconsistent internal state if a connection loss event occurs. Proper fix triggers deadlock due to interdependencies. Hackish workaround in place now, and I'm working on a proper fix. Or will be once my head stops throbbing. The code is needlessly complex and interdependent, and unravelling a bit at a time is difficult. This state problem actually explains a bunch of other odd quirks I've seen, so fixing it properly is pretty important. > * Hot unplug is still iffy (but better!) More progress here, but more work to be done. (Only a problem now if device isn't idle when yanked -- if the supplicant is attached, it deadlocks still) > * Reworked SPI interrupt delivery completely untested > * IBSS is still busted (haven't had time to tackle it properly) > * Rebase against 3.8 > * Investigate adding OF bindings that can be used instead of platform_data No progress here yet. I'll post another patch set this weekend. - Solomon -- Solomon Peachy pizza at shaftnet dot org Melbourne, FL ^^ (mail/jabber/gtalk) ^^ Quidquid latine dictum sit, altum viditur.
Attachment:
pgpWjXBam8XcP.pgp
Description: PGP signature
