Port-xen archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: boot hangs at uhci1



Manuel Bouyer wrote:
On Sun, Apr 01, 2007 at 11:57:27AM -0500, david l goodrich wrote:
Manuel Bouyer wrote:
On Sun, Apr 01, 2007 at 10:54:16AM -0500, david l goodrich wrote:
Manuel Bouyer wrote:
On Sat, Mar 31, 2007 at 05:22:20PM -0500, david l goodrich wrote:
I decided it would be a good idea to move a 4.0_BETA2 dom0 from
under
my desk to over with my other servers.

when i powered the machine back on after the move, the XEN3_DOM0
kernel won't boot, it hangs at

uhci1 at pci0 dev 29 function 1: Intel 82801GB/GR USB UHCI
Controller
(rev. 0x01)

but a GENERIC kernel will boot through grub's 'chainloader' just fine.

i'd just say wipe the hard drives and start over, but this has domUs
and data on it that i'd like to keep. and the stupid thing /used/
to
boot.

my / drive is small, of course.

nialas# df /
Filesystem  1K-blocks      Used     Avail Capacity  Mounted on
/dev/raid0a    508143    188655    294081    39%    /
nialas#

the dmesg from GENERIC is below.  any ideas?  thanks.
If you add -c to the kernel's boot command line, and enter
disable uhci
quit

does it boot ?
for those reading this later on that are confused like i was, this is
just to say add "-c" to the end of the "module" line in the appropriate
/grub/menu.lst entry.

unfortunately, all disabling uhci did was push the freeze back:

piixide1 at pci0 dev 31 function 2
piixide1: Intel 82801G/GR Serial ATA/Raid Controller (ICH7) (rev. 0x01)
piixide1: bus-master DMA support present
piixide1: primary channel configured to native-PCI mode
[hang]

i should note that the way I have determined it is hung is that 1) it
obviously doesn't progress any further in the boot process and 2) the
keyboard interrupts are no longer detected - capslock and numlock no
longer cause the keyboard lights to change.
it's normal at this point of the boot, interrupts are not yet enbaled to
the keyboard won't react anyway

Maybe you have updated this kernel recently, without
rebooting ?

no, this is not something I would do.  well, i hope not.
Besides, I have also tried booting with a new XEN3_DOM0 kernel from
ftp.netbsd.org's most recent daily build, and had the same problem.  In
fact, this most recent hang at piixide1 is using

<ftp://ftp.netbsd.org/pub/NetBSD-daily/netbsd-4/200703280002Z/i386/binary/kernel/netbsd-XEN3_DOM0.gz>,
gunzipped, of course.
It seems to have troubles with establishing interrupts. Can you try
disabling
ACPI ?
disable acpi
in userconf.

and now, hung at

Starting xen domains.
Using config file "/usr/pkg/etc/xen/router-meus".
xvif1.0: Ethernet address 00:16:3e:6d:73:75
xvif1.1: Ethernet address 00:16:3e:4c:db:4c
xbd backend: attach device vnd0d (size 10485760) for domain 1
Started domain router-meus
Creating a.out runtime link editor directory cache.
xbd backend 0x1 for domain 1 using event channel 14

the lines matching /^x/ are green, the others are white.

just now it added the line "Checking quotas: done." There was a 5-10 minute gap between the "xbd" line and the "Checking" line.

i've seen slow boots, but this is just about the slowest I've seen. Not something i'd expect from a 3ghz computer.

Hum, you should probably add acpi=off on the kernel line in grub (it's
an argument for xen.gz) when disabling acpi. This may be the cause of
the slowness.

i assume this is not supposed to cause a kernel panic.

i added "acpi=off" to the kernel line in menu.lst, entered "disable acpi" and "quit" in userconf, and then got about fifteen lines further before it panicked

uc> quit
Continuing...
mainbus 0 (root)
mainbus0: scanning 0x9fc00 to 0x9fff0 for MP signature
mainbus0: scanning 0x9f800 to 0x9fbf0 for MP signature
mainbus0: scanning 0xf0000 to 0xffff0 for MP signature
mainbus0: MP floating pointer found in bios at 0xfe680
mainbus0: MP config table at 0xfe690, 64 bytes long
cpu0 at mainbus0: apid 0 (boot processor)
cpu1 at mainbus0: apid 1 (application processor)
ioapic0 at mainbus0 apid 2 (I/O APIC)
ioapic0: pa 0xfec00000PHYSDEVOP_APIC_READ ret -22
panic: PHYSDEVOP_APIC_READ
Stopped in pid 0.1 (swapper) at netbsd:cpu_Debugger+0x4     popl  %ebp
db>

i'll leave it at the db> prompt for now, let me know if there's any information that could be helpful, but 'ps' just listed swapper.


what's interesting is that it can't register some interrupts with APCI.
Did you change something, like adding a PCI adapter ?


scout's honor, all I did was shut the machine down, unplug it, move it 15 feet, and plug it in again. i haven't added hardware in a few weeks. The only PCI adapter in there is an fxp(4) network card.
  --david





Home | Main Index | Thread Index | Old Index