This defect is considered MUST-FIX for Florence Release-Candidate #1

This is a common symptom of APM bios bugs.

Try booting with apm=no and see if that helps.

If this is ACPI-related, it will probably just go away because
we have removed ACPI from the kernel builds for now.

Please test with wolverine

This is still a problem.  We have tested with RC2 and QA0309.

Has your testing included "apm=no"?  If so, let us know.  If not,
please test it.

In either case, please test "apm=realmode" as well and let us know
what you see.

Passing those parameters to the kernel has no effect. Problem happens only with 
the SMP kernel (APM is disabled). Reboot works correctly with UP kernel, where 
APM is turned on. Both kernels reboot probperly on other systems.

Can you try adding one of the following items to lilo.conf as append= ?

reboot=c
reboot=h
reboot=b
reboot=w

None of the suggestions have worked

Created attachment 14304 [details]
ksymoops decoded

I've attached newt-oops.txt, a decoded oops, from qa0328, kernel
2.4.2-0.1.40BOOT.
-Matt

Adding MKJ who has commented on this before.

Pasting oops into here directly:

Oops: 0000
CPU:    0
EIP:    0010:[<6d6f7264>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010286
eax: 00000000   ebx: f8a76924   ecx: 2560f680   edx: 00000000
esi: 00000000   edi: 00000001   ebp: bffffeb8   esp: f7ff5e90
ds: 0018   es: 0018   ss: 0018
Process linuxrc (pid: 8, stackpage=f7ff5000)
Stack: c011b3f1 f8a76924 00000001 00000000 f7ff4000 00000000 00000072 c011b6f6 
       c0274888 00000001 00000000 c011095d f7ff5eec f78d5090 c0242000 c0172b3c 
       f7ff4000 00000000 c0242000 c0261520 f7ff5ef8 00000286 f7ff5ef8 c01106ce 
Call Trace: [<c011b3f1>] [<f8a76924>] [<c011b6f6>] [<c011095d>] [<c0172b3c>]
[<c01106ce>] [<c0110600>] 
       [<c013c5d1>] [<c013cb25>] [<c012e9a5>] [<c0108ebf>] [<fee1dead>] 
Code:  Bad EIP value.

>>EIP; 6d6f7264 Before first symbol   <=====
Trace; c011b3f1 <notifier_chain_register+35/38>
Trace; f8a76924 <[cdrom].rodata.start+224/1b3f>
Trace; c011b6f6 <sys_reboot+8e/1c0>
Trace; c011095d <schedule+239/398>
Trace; c0172b3c <clear_buffer_attributes+c/78>
Trace; c01106ce <schedule_timeout+56/a0>
Trace; c0110600 <reschedule_idle+68/98>
Trace; c013c5d1 <sys_poll+61/2f8>
Trace; c013cb25 <locks_alloc_lock+35/48>
Trace; c012e9a5 <init_special_inode+25/b4>
Trace; c0108ebf <system_call+33/38>
Trace; fee1dead <END_OF_CODE+6381ab6/????>

Kernel panic: VFS: Unable to mount root fs on 08:41
Oops: 0000
CPU:    0
EIP:    0010:[<6d6f7264>]
EFLAGS: 00010286
eax: 00000000   ebx: f8a76924   ecx: 07772c7c   edx: 00000000
esi: 00000000   edi: 00000001   ebp: bffffeb8   esp: f7ff7e90
ds: 0018   es: 0018   ss: 0018
Process linuxrc (pid: 8, stackpage=f7ff7000)
Stack: c011b43d f8a76924 00000001 00000000 f7ff6000 00000000 00000072 c011b742 
       c0272a28 00000001 00000000 c011098d f7ff7eec f7956090 c0240000 c017280c 
       f7ff6000 00000000 c0240000 c025f520 f7ff7ef8 00000286 f7ff7ef8 c01106fe 
Call Trace: [<c011b43d>] [<f8a76924>] [<c011b742>] [<c011098d>] [<c017280c>]
[<c01106fe>] [<c0110630>] 
       [<c013bea9>] [<c013c3fd>] [<c012e285>] [<c0108ebf>] [<fee1dead>] 
Code:  Bad EIP value.

>>EIP; 6d6f7264 Before first symbol   <=====
Trace; c011b43d <notifier_call_chain+1d/34>
Trace; f8a76924 <[cdrom].rodata.start+224/1b3f>
Trace; c011b742 <sys_reboot+da/1c0>
Trace; c011098d <schedule+269/398>
Trace; c017280c <con_put_char+1c/20>
Trace; c01106fe <schedule_timeout+86/a0>
Trace; c0110630 <process_timeout+0/48>
Trace; c013bea9 <do_select+1fd/214>
Trace; c013c3fd <sys_select+521/530>
Trace; c012e285 <sys_write+c1/cc>
Trace; c0108ebf <system_call+33/38>
Trace; fee1dead <END_OF_CODE+6381ab6/????>

Kernel panic: VFS: Unable to mount root fs on 08:41

*** Bug 32578 has been marked as a duplicate of this bug. ***

My mistake, 32578 is *not* a duplicate.  32578 causes an oops on reboot, while 
this one only hangs on reboot.  The oops messages above were incorrectly put 
into this bugzilla.  Sorry for the confusion.

putting append= "reboot=b" in lilo.conf fixes the issue.

I'll mail you a program we can use to detect the exact bios (as this is a bios
problem) this machine has, and then make "reboot=b" the default for this bios.
Please scream if that is unacceptable to you.

Created attachment 15924 [details]
Output of dmidecode for PE 1300

Thanks. This seems to be the second PE1300 model that has the problem...
I'll add both to the workaround list.

Should be fixed as of 2.4.3-2.12.1 (available from the beta-site)