On Wed, Oct 14, 2009 at 10:23:13AM -0400, Alan Stern wrote: > On Tue, 13 Oct 2009, Sarah Sharp wrote: > > > Ah, ok, I see the flaw there. I was perplexed because the fix worked on > > this platform 20 times, and failed 20 times without it. I can see from > > my logs that I stopped using cheese before the microframe index wrapped. > > Cheese? > > > I can turn this into a patch and test it on the box on Friday. At this > > point it's really your code, so would you like to make the patch? > > Okay. Post the patch you test; I'll review it and send it to Greg. The next email is the patch I successfully tested on the new chipset. I found out that scheduling high speed iTDs within the IST doesn't cause this bug, so I want to rework the patch to only do this new scheduling with full speed and low speed devices. I also found the same bug on an Intel ICH9 chipset (in my quad core desktop and my x200s Lenovo laptop). It seems like this chipset bug was in reused silicon IP, although I'm not sure how far back it goes. Sarah -- 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
