Hello.
Wessel, Jason wrote:
> Interesting...
> I would have thought the generic fix I put in recently with respect to
> restoring the preempt count after a bad memory read would have solved
> this unless there is a another case that is not obvious with the rt
> patch.
No. I'm getting the last mentioned trace with just KGDB patchset atop of
2.6.21, no -rt patch.
> Jason.
>>-----Original Message-----
>>From: Sergei Shtylyov [mailto:[EMAIL PROTECTED]
>>Sent: Tuesday, June 19, 2007 1:53 PM
>>To: Wessel, Jason
>>Cc: [email protected]
>>Subject: Re: [Kgdb-bugreport] KGDB/RT integration woes
>>
>>Hello, I wrote:
>>
>>
>>>>>This patch fixes some corner cases where KGDB will
>>
>>silently hang or
>>
>>>>>kill the system, if a user accidentally tries to source step into a
>>>>>spin_unlock() call or source step in on a macro containing
>>>>>smp_processor_id(). The use of raw_smp_processor_id is desired in
>>>>>kernel/kgdb.c to fix this particular problem.
>>
>>>>>To fix issues with accidental source step in on spin_unlock(), the
>>>>>idea is to check for the existence of a break point on the second
>>>>>entry to kgdb and try to remove it. Next an attempt will
>>
>>be made to
>>
>>>>>continue normal operations. A third entry will generate a
>>
>>panic(),
>>
>>>>>so as to stop infinite loops.
>>
>>>>>Testing has shown that kgdb is much more robust with these changes
>>>>>and "random accidental run control".
>>
>>>> It seems that something has been broken WRT backtracing...
>>
>>> Here's the more relevant trace -- got it at the initial
>>
>>KGDB breakpoint:
>>
>> Well, it seems to be caused by KGDB using the infamous
>>unwinder which is a part of the -rt patch now... :-/
>> With the plain kernel, backtracing works well, however,
>>that's what I'm getting after hitting breaklpint set to
>>do_mount and continuing:
>>
>>BUG: scheduling while atomic: swapper/0x0000000a/1
>>
>>Call Trace:
>> [<ffffffff805724e0>] __sched_text_start+0x60/0x7ee
>> [<ffffffff8027d7fb>] ____cache_alloc_node+0x111/0x120
>> [<ffffffff8022c4f9>] default_wake_function+0xd/0xf
>> [<ffffffff80574a06>] __lock_text_start+0x2e/0x30
>> [<ffffffff80242734>] flush_cpu_workqueue+0x90/0xb8
>> [<ffffffff80246192>] autoremove_wake_function+0x0/0x37
>> [<ffffffff80242200>] __queue_work+0x7b/0x83
>> [<ffffffff80246192>] autoremove_wake_function+0x0/0x37
>> [<ffffffff8024229a>] queue_work+0x92/0x9e
>> [<ffffffff802427ba>] flush_workqueue+0x5e/0x7f
>> [<ffffffff80242c15>] flush_scheduled_work+0x10/0x12
>> [<ffffffff8056463b>] xs_connect+0xd3/0xd8
>> [<ffffffff80561f01>] xprt_connect+0x11a/0x123
>> [<ffffffff8056055c>] call_connect+0x79/0x7e
>> [<ffffffff80565a43>] __rpc_execute+0x84/0x248
>> [<ffffffff80565c27>] rpc_execute+0x20/0x24
>> [<ffffffff80560095>] call_start+0x0/0x72
>> [<ffffffff8055ffc5>] rpc_call_sync+0x84/0xab
>> [<ffffffff8056cca0>] rpc_getport_external+0xe1/0x104
>> [<ffffffff807abe8d>] root_nfs_getport+0x6e/0x79
>> [<ffffffff807ac04f>] nfs_root_data+0x1b7/0x32c
>> [<ffffffff802983a1>] sys_mount+0xc1/0xd3
>> [<ffffffff8028b9b2>] sys_rmdir+0x11/0x13
>> [<ffffffff80796f30>] mount_root+0x1d/0x138
>> [<ffffffff80797152>] prepare_namespace+0x107/0x134
>> [<ffffffff80796a60>] init+0x25e/0x270
>> [<ffffffff80574e9b>] _spin_unlock_irq+0x15/0x2f
>> [<ffffffff8022ac65>] schedule_tail+0x36/0x9d
>> [<ffffffff8020a3b8>] child_rip+0xa/0x12
>> [<ffffffff803711b8>] acpi_ds_init_one_object+0x0/0x88
>> [<ffffffff80796802>] init+0x0/0x270
>> [<ffffffff8020a3ae>] child_rip+0x0/0x12
WBR, Sergei
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Kgdb-bugreport mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/kgdb-bugreport
