Hi Vladimir, first of all - I hope You will feel better soon. According to EHCI patch: There is real bug in QH handling. My solution is maybe not optimal but it looks to be working and solving the problem. Of course, You can improve it when You'll feel better :-) The main problem was: The original "allocation" algorithm can select QH which is linked in wrong list (Asynchronous/Synchronous). It led to problem described in my previous e-mail(s) (non-working keyboard on non-root port where was USB disk connected previously). But there were also some another minor bugs, e.g. in intialization of QH and in transfer canceling routine.
According to fuloong 2F: Mostly I agree with You. It looks for me like OHCI driver is not working at all in Javier case - because also USB disk is not working. It is very suspicious... - but I cannot imagine yet what pmon can do so wrong with USB controller(s)... I try at least to clarify what is going wrong - if Javier will be able to capture debug output and send to me. And then we will see... BR, Ales > > Hi, > > thanks for the test(s). > > > > It looks like there are two possible cases: > > - OHCI driver is not working on fuloong 2f > > - root hub events handling is still not correct on fuloong 2f > > > I doubt either of those. Everything works fine on my fuloong 2F which is > exactly the same machine. So either: > a) it's because reporter uses pmon and pmon may put USB in weird state. > b) his keyboard doesn't follow the spec. > > Fortunately fuloong 2F has more or less classic serial interface so we > > probably can capture debug output from whole USB subsystem. > > Do You have some another PC which can capture debug output from Fuloong 2f > > via serial cable? > > If yes, we will contiue with tests at the end of week, currently I am busy. > > > > BR, > > Ales > > > >> | Hm. > >> | Could You make following additional test? > >> | ... > >> | > >> | 4. > >> | Boot the PC. During boot check this: > >> | - If USB disk blinks when ls command is issued (i.e. if USB disk is > >> | accessed by GRUB). > >> > >> No, it doesn't. > >> > >> | - If ls command output contains somewhere at least "usb(0)" or "usb(0) > >> | usb(0,msdos1)" or whatever similar. > >> > >> Unfortunately no. > >> > >> | > >> | 5. > >> | Send the result to us. :-) > >> > >> Done, :-) > >> > >> Thanks, > >> > >> -- > >> Javier. > >> > >> _______________________________________________ > >> Grub-devel mailing list > >> Grub-devel@gnu.org > >> https://lists.gnu.org/mailman/listinfo/grub-devel > > > > > > > > _______________________________________________ > > Grub-devel mailing list > > Grub-devel@gnu.org > > https://lists.gnu.org/mailman/listinfo/grub-devel > > > > > _______________________________________________ > Grub-devel mailing list > Grub-devel@gnu.org > https://lists.gnu.org/mailman/listinfo/grub-devel _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel
