Gadiyar, Anand wrote: > Felipe Balbi wrote: > > On Wed, Dec 09, 2009 at 02:30:35PM +0100, ext Gadiyar, Anand wrote: > > >Another issue with the whole series: > > > > > >testusb with g_zero fails on some OUT transfers: > > >I ran this: > > >sudo ./testusb -a -t 1 -c 1 -s 4096 > > > > > >On the OMAP, I see only 512 bytes were received, > > >and DMA is still waiting for the remaining 3584 > > >bytes. Maybe a bug in the Autoclear handling? > > > > > >IN transfers went through okay. > > > > > >Will debug some more tomorrow. Any chance you can give this a > > >shot on your boards as well? > > > > I can ran it tomorrow. I'll have time tomorrow for looking at this > > also. > > > > This issue occurs with "[PATCH 21/22] usb: musb: enabling dma > tx for gadget" > > Logs below (at debug level 5). > > Not sure when I'll next get a chance to look at this, so sent > this out now. Will revisit sometime later, unless someone beats > me to it. > Got it! This behavior seems to be due to the fallback interrupt mode transfer not working. With patch 8/22 (the workaround for simultaneous DMA usage on OMAP3430), if we cannot carry out a transfer in DMA mode, we return false in dma_channel_program. While this used to work before the mode1 patches, we dropped another patch from me fixing the aborted RX DMA path. I'm still not sure what change is required with the new mode1 series to fix this. I'll take a stab at it next week, unless someone can find a fix before then. - Anand -- 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