On Tue, Dec 22, 2009 at 17:13, Felipe Balbi wrote: > On Tue, Dec 22, 2009 at 11:03:00PM +0100, ext Mike Frysinger wrote: >> and that leads into another patch we have ... the Blackfin MUSB has >> some anomalies where mode1 DMA is screwy, so we need this correctness >> for things to be transferred properly :) > > musb as a whole is screwy in lots of places :-p > > What version of the RTL are you guys using ? Do you have some piece of > errata you could share ? (even if it's off list) afaik, the current silicon out there (BF52x/BF54x) is based on MUSB 1.9. every anomaly/issue we've found so far is due to bugs in 1.9 and are fixed in 2.0, or due to how we hooked it up. unfortunately, the risk of upgrading the silicon in the current series of parts too great and so we're going to have to live with most of them. all future Blackfin parts will have the latest MUSB IP and the other connection fixes, so things should be peachy there. as for anomaly lists, you can find a concise public mirror here: http://blackfin.uclinux.org/gf/project/bfin-docs/frs if you're interested in even more back story on a specific anomaly, i can facilitate private information leakage in your direction ;). >> but i'll start another thread on these topics since we cant seem to >> find a way to push this crap into our driver ... we have to tweak a >> few common pieces of code with Blackfin ifdefs :( > > please don't use ifdefs. It took a lot of time to remove them from musb. this is why you havent seen me post patches for these issues :P only musb changes i've been pushing are the generalized ones > can provide you with patches for the mode1 support that we wrote and are > trying to get it tested. If you can base on top of those patches I'd be > really glad. Tomorrow I'll be stress testing the mode1 patches again, and if > those are fine I'm (actually was) willing to push those to Greg, but it'll > probably break blackfin. So let's study a way to have blackfin using mode0 > and everyone else using mode1. i can start a thread of "here's our crappy patches that work for us, please 2 help out" ... obvious understanding is that we dont get blasted on every one with "this is not how you do things!". we know that much already :P. -mike -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
