[uml-devel] Errors in schedule / _switch_to

[uml-devel] Errors in schedule / _switch_to

[Oscar Hellström]

Hi,

I'm running an older version of the kernel, 2.6.32.3 and I'm experiencing
segmentations faults in _switch_to, which is called from schedule. My
stacktrace looks like:
3 <signal handler called>
#14 _switch_to (prev=0x43ff1340, next=0xffffffd4, last=0x43ff1340) at
arch/um/kernel/process.c:82
#15 0x08205cd9 in schedule () at kernel/sched.c:2886
#16 0x08205dda in io_schedule () at kernel/sched.c:6784
#17 0x0809af72 in sync_page (word=0x96ba180) at mm/filemap.c:187
#18 0x08205fde in __wait_on_bit_lock (wq=0x980b7d8, q=0x43fd5ce4,
action=0x809af37 <sync_page>, mode=2) at kernel/wait.c:229
#19 0x0809af21 in __lock_page (page=0x96ba180) at mm/filemap.c:601
#20 0x0809afbd in find_lock_page (mapping=0x47608668, offset=2596) at
include/linux/pagemap.h:317
#21 0x0809c888 in filemap_fault (vma=0x43d70764, vmf=0x43fd5d8c) at
mm/filemap.c:1538
#22 0x080a88ad in __do_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, pmd=0x472c0914, pgoff=2596, flags=0, 
orig_pte={pte = 0}) at mm/memory.c:2719
#23 0x080a9fec in handle_mm_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, flags=0) at mm/memory.c:2917
#24 0x0805c5f1 in handle_page_fault (address=2437412980, ip=1076410045,
is_write=0, is_user=1, code_out=0x43fd5e40)
at arch/um/kernel/trap.c:71
#25 0x0805c729 in segv (fi={error_code = 4, cr2 = 2437412980, trap_no =
14}, ip=1076410045, is_user=1, regs=0x43ff1550)
at arch/um/kernel/trap.c:178
#26 0x0805c91e in segv_handler (sig=11, regs=0x43ff1550) at
arch/um/kernel/trap.c:150
#27 0x0806c00a in userspace (regs=0x43ff1550) at
arch/um/os-Linux/skas/process.c:421
#28 0x0805aa75 in fork_handler () at arch/um/kernel/process.c:179
#29 0x00000000 in ?? ()

I've compiled in a driver into the kernel which fakes interrupts by sending
signals to the UML process. I guess this could interfere with something,
but I do believe I've seen this without the signals being enabled.
Unfortunately I don't have any easy way to reproduce this, it usually
happens seemingly at random.

Any pointer to what this can be and where I can start looking would be
appreciated!

Thanks
/Oscar
-- 
Oscar Hellström, oscar@hellstrom.st
web: http://oscar.hellstrom.st
xmpp: oscar@hellstrom.st

------------------------------------------------------------------------------
Why Cloud-Based Security and Archiving Make Sense
Osterman Research conducted this study that outlines how and why cloud
computing security and archiving is rapidly being adopted across the IT 
space for its ease of implementation, lower cost, and increased 
reliability. Learn more. http://www.accelacomm.com/jaw/sfnl/114/51425301/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[uml-devel] Errors in schedule / _switch_to

[Oscar Hellström]

Hi,

I'm running an older version of the kernel, 2.6.32.3, and I'm experiencing
segmentations faults in _switch_to, which is called from schedule. My
stacktrace looks like:
...
#13 <signal handler called>
#14 _switch_to (prev=0x43ff1340, next=0xffffffd4, last=0x43ff1340) at
arch/um/kernel/process.c:82
#15 0x08205cd9 in schedule () at kernel/sched.c:2886
#16 0x08205dda in io_schedule () at kernel/sched.c:6784
#17 0x0809af72 in sync_page (word=0x96ba180) at mm/filemap.c:187
#18 0x08205fde in __wait_on_bit_lock (wq=0x980b7d8, q=0x43fd5ce4,
action=0x809af37 <sync_page>, mode=2) at kernel/wait.c:229
#19 0x0809af21 in __lock_page (page=0x96ba180) at mm/filemap.c:601
#20 0x0809afbd in find_lock_page (mapping=0x47608668, offset=2596) at
include/linux/pagemap.h:317
#21 0x0809c888 in filemap_fault (vma=0x43d70764, vmf=0x43fd5d8c) at
mm/filemap.c:1538
#22 0x080a88ad in __do_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, pmd=0x472c0914, pgoff=2596, flags=0, orig_pte={pte =
0}) at mm/memory.c:2719
#23 0x080a9fec in handle_mm_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, flags=0) at mm/memory.c:2917
#24 0x0805c5f1 in handle_page_fault (address=2437412980, ip=1076410045,
is_write=0, is_user=1, code_out=0x43fd5e40) at arch/um/kernel/trap.c:71
#25 0x0805c729 in segv (fi={error_code = 4, cr2 = 2437412980, trap_no =
14}, ip=1076410045, is_user=1, regs=0x43ff1550) at
arch/um/kernel/trap.c:178
#26 0x0805c91e in segv_handler (sig=11, regs=0x43ff1550) at
arch/um/kernel/trap.c:150
#27 0x0806c00a in userspace (regs=0x43ff1550) at
arch/um/os-Linux/skas/process.c:421
#28 0x0805aa75 in fork_handler () at arch/um/kernel/process.c:179
#29 0x00000000 in ?? ()

I've compiled in a driver into the kernel which fakes interrupts by sending
signals to the UML process. I guess this could interfere with something,
but I do believe I've seen this without the signals being enabled.
Unfortunately I don't have any easy way to reproduce this, it happens
seemingly at random.

Any pointer to what this can be and where I can start looking would be
appreciated!

Thanks
/Oscar

------------------------------------------------------------------------------
Why Cloud-Based Security and Archiving Make Sense
Osterman Research conducted this study that outlines how and why cloud
computing security and archiving is rapidly being adopted across the IT 
space for its ease of implementation, lower cost, and increased 
reliability. Learn more. http://www.accelacomm.com/jaw/sfnl/114/51425301/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Errors in schedule / _switch_to

[richard -rw- weinberger]

2011/9/9 Oscar Hellström <oscar@hellstrom.st>:
>
> Any pointer to what this can be and where I can start looking would be
> appreciated!
>

Does with also happen with a vanilla kernel?
Do you have a full trace? The value of EIP would be helpful.

-- 
Thanks,
//richard

------------------------------------------------------------------------------
Malware Security Report: Protecting Your Business, Customers, and the 
Bottom Line. Protect your business and customers by understanding the 
threat from malware and how it can impact your online business. 
http://www.accelacomm.com/jaw/sfnl/114/51427462/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Errors in schedule / _switch_to

[Oscar Hellström]

Hi,

I don't think I can reproduce this on a vanilla kernel, but I can't run my
application without my driver (and some other changes).

bt full gives me:
....
#13 <signal handler called>
No symbol table info available.
#14 _switch_to (prev=0x43ff1340, next=0xffffffd4, last=0x43ff1340) at
arch/um/kernel/process.c:82
from = (struct task_struct *) 0x43ff1340
to = (struct task_struct *) 0xffffffd4
#15 0x08205cd9 in schedule () at kernel/sched.c:2886
prev = (struct task_struct *) 0x43ff1340
next = (struct task_struct *) 0xffffffd4
switch_count = (long unsigned int *) 0x43ff14c8
#16 0x08205dda in io_schedule () at kernel/sched.c:6784
No locals.
#17 0x0809af72 in sync_page (word=0x96ba180) at mm/filemap.c:187
mapping = (struct address_space *) 0x0
#18 0x08205fde in __wait_on_bit_lock (wq=0x980b7d8, q=0x43fd5ce4,
action=0x809af37 <sync_page>, mode=2) at kernel/wait.c:229
ret = 158048640
#19 0x0809af21 in __lock_page (page=0x96ba180) at mm/filemap.c:601
wait = {key = {flags = 0x96ba180, bit_nr = 0}, wait = {flags = 1, private =
0x43ff1340, 
func = 0x8086ef9 <wake_bit_function>, task_list = {next = 0x980b7d8, prev =
0x980b7d8}}}
#20 0x0809afbd in find_lock_page (mapping=0x47608668, offset=2596) at
include/linux/pagemap.h:317
page = (struct page *) 0x96ba180
#21 0x0809c888 in filemap_fault (vma=0x43d70764, vmf=0x43fd5d8c) at
mm/filemap.c:1538
error = 0
file = (struct file *) 0x473b8ba0
mapping = (struct address_space *) 0x47608668
ra = (struct file_ra_state *) 0x473b8be0
inode = (struct inode *) 0x476085cc
offset = 2596
page = (struct page *) 0x0
ret = 4
#22 0x080a88ad in __do_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, pmd=0x472c0914, pgoff=2596, flags=0, 
orig_pte={pte = 0}) at mm/memory.c:2719
page = <value optimized out>
anon = <value optimized out>
dirty_page = <value optimized out>
vmf = {flags = 0, pgoff = 2596, virtual_address = 0x9147f000, page = 0x0}
ret = 115
page_mkwrite = <value optimized out>
#23 0x080a9fec in handle_mm_fault (mm=0x478855e0, vma=0x43d70764,
address=2437412980, flags=0) at mm/memory.c:2917
pgd = (pgd_t *) 0x472c0914
pte = (pte_t *) 0x404551fc
#24 0x0805c5f1 in handle_page_fault (address=2437412980, ip=1076410045,
is_write=0, is_user=1, code_out=0x43fd5e40)
at arch/um/kernel/trap.c:71
fault = 0
mm = (struct mm_struct *) 0x478855e0
vma = (struct vm_area_struct *) 0x43d70764
err = <value optimized out>
#25 0x0805c729 in segv (fi={error_code = 4, cr2 = 2437412980, trap_no =
14}, ip=1076410045, is_user=1, regs=0x43ff1550)
at arch/um/kernel/trap.c:178
si = {si_signo = 0, si_errno = 0, si_code = 196610, _sifields = {_pad = {0
<repeats 27 times>, 14, 16157}, _kill = {
_pid = 0, _uid = 0}, _timer = {_tid = 0, _overrun = 0, _pad = 0x43fd5e4c
"", _sigval = {sival_int = 0, 
sival_ptr = 0x0}, _sys_private = 0}, _rt = {_pid = 0, _uid = 0, _sigval =
{sival_int = 0, sival_ptr = 0x0}}, 
_sigchld = {_pid = 0, _uid = 0, _status = 0, _utime = 0, _stime = 0},
_sigfault = {_addr = 0x0, _addr_lsb = 0}, 
_sigpoll = {_band = 0, _fd = 0}}}
catcher = <value optimized out>
err = <value optimized out>
address = <value optimized out>
#26 0x0805c91e in segv_handler (sig=11, regs=0x43ff1550) at
arch/um/kernel/trap.c:150
fi = <value optimized out>
#27 0x0806c00a in userspace (regs=0x43ff1550) at
arch/um/os-Linux/skas/process.c:421
sig = <value optimized out>
timer = {it_interval = {tv_sec = 0, tv_usec = 0}, it_value = {tv_sec = 0,
tv_usec = 16001}}
nsecs = 1315569399711559000
err = <value optimized out>
status = 2943
op = <value optimized out>
pid = 16157
local_using_sysemu = 2
#28 0x0805aa75 in fork_handler () at arch/um/kernel/process.c:179
No locals.
#29 0x00000000 in ?? ()
No symbol table info available.


Stack level 14, frame at 0x43fd5c74:
eip = 0x805a900 in _switch_to (arch/um/kernel/process.c:82); saved eip
0x8205cd9
called by frame at 0x43fd5ca4, caller of frame at 0x43fd5c60
source language c.
Arglist at 0x43fd5c6c, args: prev=0x43ff1340, next=0xffffffd4,
last=0x43ff1340
Locals at 0x43fd5c6c, Previous frame's sp is 0x43fd5c74
Saved registers:
ebx at 0x43fd5c64, ebp at 0x43fd5c6c, esi at 0x43fd5c68, eip at 0x43fd5c70

Thanks
/Oscar 

On Sat, 10 Sep 2011 13:10:09 +0200, richard -rw- weinberger
<richard.weinberger@gmail.com> wrote:
> 2011/9/9 Oscar Hellström <oscar@hellstrom.st>:
>>
>> Any pointer to what this can be and where I can start looking would be
>> appreciated!
>>
> 
> Does with also happen with a vanilla kernel?
> Do you have a full trace? The value of EIP would be helpful.

-- 
Oscar Hellström, oscar@hellstrom.st
web: http://oscar.hellstrom.st
xmpp: oscar@hellstrom.st

------------------------------------------------------------------------------
Doing More with Less: The Next Generation Virtual Desktop 
What are the key obstacles that have prevented many mid-market businesses
from deploying virtual desktops?   How do next-generation virtual desktops
provide companies an easier-to-deploy, easier-to-manage and more affordable
virtual desktop model.http://www.accelacomm.com/jaw/sfnl/114/51426474/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel