On Thu, Feb 26, 2015 at 2:04 PM, Bin Liu <binmlist@xxxxxxxxx> wrote: > On Thu, Feb 26, 2015 at 1:59 PM, Felipe Balbi <balbi@xxxxxx> wrote: >> On Thu, Feb 26, 2015 at 01:49:51PM -0600, Bin Liu wrote: >>> On Thu, Feb 26, 2015 at 1:48 PM, Felipe Balbi <balbi@xxxxxx> wrote: >>> > On Thu, Feb 26, 2015 at 01:30:21PM -0600, Bin Liu wrote: >>> >> Felipe, >>> >> >>> >> On Thu, Feb 26, 2015 at 12:27 PM, Felipe Balbi <balbi@xxxxxx> wrote: >>> >> > On Thu, Feb 26, 2015 at 12:25:16PM -0600, Felipe Balbi wrote: >>> >> >> FSDEV is set for both HIGH and FULL speeds, >>> >> >> the correct HIGHSPEED check is done through >>> >> >> power register's HSMODE bit. >>> >> >> >>> >> >> Signed-off-by: Felipe Balbi <balbi@xxxxxx> >>> >> > >>> >> > I'm still unsure if we should really ignore babble on FS/LS. It seems to >>> >> > me we should never ignore it, but I really don't have a way to prove >>> >> > this statement. For the sake of reducing impact, we will just fix HS >>> >> > check here. >>> >> > >>> >> >>> >> I believe we should drop speed check in here and not ignore babble >>> >> regardless. We have seen many cases that full-speed babble causes MUSB >>> >> stop working. >>> > >>> > I'll make that as a separate patch then, just to make sure we can revert >>> > it later if something goes wrong ;-) >>> >>> Agreed. >> >> I noticed something else. If we really don't need to reset musb in case >> of babble, then we can drop that recover_work completely which >> simplifies babble handling quite a bit. >> >> I'll fiddle with that, if you don't mind. > > That is fine with me. I am writing the comments for the dropping reset > patch right now ;) I meant your idea is better. I will drop my patch. > > We only need the following in musb_recover_work() for the recovery. > > 1852 /* > 1853 * When a babble condition occurs, the musb controller > 1854 * removes the session bit and the endpoint config is lost. > 1855 */ > 1856 if (musb->dyn_fifo) > 1857 status = ep_config_from_table(musb); > 1858 else > 1859 status = ep_config_from_hw(musb); > 1860 > 1861 /* start the session again */ > 1862 if (status == 0) > 1863 musb_start(musb); > >> >> -- >> balbi -- 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