Ognyan Kulev wrote:
So hanging on
cat /proc/bus/usb/devices happens only in 2nd stage and not in the next
reboot, and it depends somewhat on selected tasks. Cyrillic environment
asks some questions before xfree86, so it's possible that it's the
source of the problem. I will check that.
I've just made an install with only tasks X and cyrillic selected, and
there was no problem, so it's not cyrillic environment alone.
I can think of only one more thing: In both my unstable+experimental and
d-i testing systems, after heavy network activity for a long time (like
downloading 600M from Debian mirror), the following kernel message is
printed: "Disabling IRQ #10". In my unstable+experimental system with
2.6.6-686-smp, additionally in kern.log there is the following information:
May 23 18:03:47 roller kernel: irq 10: nobody cared!
May 23 18:03:47 roller kernel: Call Trace:
May 23 18:03:47 roller kernel: [__report_bad_irq+42/144]
__report_bad_irq+0x2a/0x90
May 23 18:03:47 roller kernel: [note_interrupt+112/176]
note_interrupt+0x70/0xb0
May 23 18:03:47 roller kernel: [do_IRQ+352/416] do_IRQ+0x160/0x1a0
May 23 18:03:47 roller kernel: [common_interrupt+24/32]
common_interrupt+0x18/0x20
May 23 18:03:47 roller kernel: [default_idle+0/64] default_idle+0x0/0x40
May 23 18:03:47 roller kernel: [default_idle+44/64] default_idle+0x2c/0x40
May 23 18:03:47 roller kernel: [cpu_idle+59/80] cpu_idle+0x3b/0x50
May 23 18:03:47 roller kernel: [unknown_bootoption+0/352]
unknown_bootoption+0x0/0x160
May 23 18:03:47 roller kernel: [start_kernel+443/528]
start_kernel+0x1bb/0x210
May 23 18:03:47 roller kernel: [unknown_bootoption+0/352]
unknown_bootoption+0x0/0x160
May 23 18:03:47 roller kernel:
May 23 18:03:47 roller kernel: handlers:
May 23 18:03:47 roller kernel: [__crc_d_alloc_anon+7442156/10888228]
(usb_hcd_irq+0x0/0x70 [usbcore])
May 23 18:03:47 roller kernel: [__crc_d_alloc_anon+7442156/10888228]
(usb_hcd_irq+0x0/0x70 [usbcore])
May 23 18:03:47 roller kernel: Disabling IRQ #10
So there is one more task for me: to see if after making IRQ #10 to be
disabled, the problem appears.
Anyway, please someone knowledgable clone this bug and reassign the
clone to some appropriate place, because 251020 is assigned to xfree86
team because of the other problem in the installation report (horiz/vert
refresh rates are not taken from DDC).
Regards,
ogi