Hi! I''m using xen-2.0.6 with a 2.6.11.10 kernel in dom0 on a dual Athon machine. In dom0, I have problems accessing the floppy drive. Most time sectors can be read, but very slowly due to timeouts etc., and sometimes a sector fails alltogether. First I thought that the disk itself is damaged, but a few tries later (several disks in 3 floppy drives) I realized that the dom0 kernel must have some kind of problem... All the disks work in all drives except in the one in the Xen machine. I see the following kernel messages in my syslog: floppy0: unexpected interrupt floppy0: sensei repl[0]=80 floppy0: -- FDC reply errorfloppy0: unexpected interrupt floppy0: sensei repl[0]=80 floppy0: -- FDC reply errorfloppy0: unexpected interrupt floppy0: sensei repl[0]=80 floppy0: -- FDC reply errorfloppy0: unexpected interrupt floppy0: sensei repl[0]=80 floppy driver state ------------------- now=5458913 last interrupt=5458633 diff=280 last called handler=00000000 timeout_message=floppy start last output bytes: e6 80 5458595 0 90 5458595 1b 90 5458595 0 90 5458595 a 90 5458595 2 90 5458595 12 90 5458595 1b 90 5458595 ff 90 5458595 e6 80 5458613 0 90 5458613 1b 90 5458613 0 90 5458613 c 90 5458613 2 90 5458613 12 90 5458613 1b 90 5458613 ff 90 5458613 8 90 5458633 8 80 5458633 last result at 5458633 last redo_fd_request at 5458613 80 status=80 fdc_busy=1 cont=c02c6d44 current_req=cb917838 command_status=-1 floppy0: floppy timeout called end_request: I/O error, dev fd0, sector 983 (sector numbers vary, of course) Looks somehow like is has to do with interrupts, right? :) I also wonder that the floppy irq has a really high counter in /proc/interrupts. It increments by 17055 just by mounting the disk (ext2 filesystem). Isn''t that much too much? Anybody know this effect or have a solution? Roman (please Cc replies to me, I''m not subscribed to the list) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> I''m using xen-2.0.6 with a 2.6.11.10 kernel in dom0 on a dual > Athon machine. In dom0, I have problems accessing the floppy > drive. Most time sectors can be read, but very slowly due to > timeouts etc., and sometimes a sector fails alltogether.It would be interesting to know what happens if you try the unstable tree (after a make uninstall). The floppy driver in xen linux is PIO mode only -- DMA can''t be used as the memory pages may be above the 16MB ISA DMA zone. Ian> First I thought that the disk itself is damaged, but a few > tries later (several disks in 3 floppy drives) I realized > that the dom0 kernel must have some kind of problem... All > the disks work in all drives except in the one in the Xen > machine. I see the following kernel messages in my syslog: > > floppy0: unexpected interrupt > floppy0: sensei repl[0]=80 > floppy0: -- FDC reply errorfloppy0: unexpected interrupt > floppy0: sensei repl[0]=80 > floppy0: -- FDC reply errorfloppy0: unexpected interrupt > floppy0: sensei repl[0]=80 > floppy0: -- FDC reply errorfloppy0: unexpected interrupt > floppy0: sensei repl[0]=80 > > floppy driver state > ------------------- > now=5458913 last interrupt=5458633 diff=280 last called > handler=00000000 > timeout_message=floppy start > last output bytes: > e6 80 5458595 > 0 90 5458595 > 1b 90 5458595 > 0 90 5458595 > a 90 5458595 > 2 90 5458595 > 12 90 5458595 > 1b 90 5458595 > ff 90 5458595 > e6 80 5458613 > 0 90 5458613 > 1b 90 5458613 > 0 90 5458613 > c 90 5458613 > 2 90 5458613 > 12 90 5458613 > 1b 90 5458613 > ff 90 5458613 > 8 90 5458633 > 8 80 5458633 > last result at 5458633 > last redo_fd_request at 5458613 > 80 > status=80 > fdc_busy=1 > cont=c02c6d44 > current_req=cb917838 > command_status=-1 > > floppy0: floppy timeout called > end_request: I/O error, dev fd0, sector 983 > > (sector numbers vary, of course) > > Looks somehow like is has to do with interrupts, right? :) I > also wonder that the floppy irq has a really high counter in > /proc/interrupts. It increments by 17055 just by mounting the > disk (ext2 filesystem). Isn''t that much too much? > > Anybody know this effect or have a solution? > > Roman > (please Cc replies to me, I''m not subscribed to the list) > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users