BUG kmalloc-2048: Poison overwritten

Previous thread: [2.6.26-rc4] Problems with rt2x00 USB interface on powerpc by Rogério on Sunday, June 1, 2008 - 12:30 pm. (21 messages)

Next thread: [patch -v2 00/23] use simple_read_from_buffer and memory_read_from_buffer by akinobu.mita on Sunday, June 1, 2008 - 4:13 pm. (1 message)
From: Justin Mattock
Date: Sunday, June 1, 2008 - 4:13 pm

I received this after the second simultaneous suspend: (I will try and
see if this is re creatable or not)

[  232.270737] BUG: unable to handle kernel paging request at 6b6b6c6b
[  232.270743] IP: [<c013f96d>] module_put+0x12/0x2b
[  232.270750] *pde = 00000000
[  232.270753] Oops: 0002 [#1] SMP
[  232.270757] Modules linked in: hci_usb cpufreq_ondemand
cpufreq_performance cpufreq_powersave rfcomm hidp l2cap bluetooth fan
ipmi_watchdog ipmi_msghandler uinput wlan_tkip ieee80211_crypt_tkip
ieee80211_crypt arpt_mangle arptable_filter arp_tables
nf_conntrack_ipv4 nf_conntrack iptable_mangle iptable_filter ip_tables
x_tables intelfb i2c_i810 i2c_algo_bit coretemp eeprom acpi_cpufreq
fglrx(P) applesmc joydev uvcvideo appletouch wlan_scan_sta
ath_rate_sample firewire_ohci firewire_core snd_hda_intel ath_pci
snd_pcm wlan ohci1394 intel_agp evdev ath_hal(P) snd_timer ieee1394
pata_acpi thermal ehci_hcd uhci_hcd snd_page_alloc snd_hwdep agpgart
video button processor
[  232.270800]
[  232.270802] Pid: 2690, comm: btdelconn Tainted: P
(2.6.26-rc4-00033-g5463508 #4)
[  232.270805] EIP: 0060:[<c013f96d>] EFLAGS: 00010246 CPU: 0
[  232.270808] EIP is at module_put+0x12/0x2b
[  232.270810] EAX: 00000000 EBX: 00000000 ECX: c020f49b EDX: 6b6b6b6b
[  232.270812] ESI: f4858830 EDI: f572aab8 EBP: f4893f78 ESP: f4893f78
[  232.270815]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
[  232.270817] Process btdelconn (pid: 2690, ti=f4892000 task=f7fe6220
task.ti=f4892000)
[  232.270819] Stack: f4893f8c f8bcb705 f572aaac f5b8bcb0 f572aaa8
f4893fa8 c012f90d f5b8bcc0
[  232.270826]        f8bcb6a6 f5b8bcb0 f4893fb0 f5b8bcc8 f4893fd0
c0130119 00000000 f7fe6220
[  232.270832]        c01324cd f4893fbc f4893fbc f5b8bcb0 c0130063
00000000 f4893fe0 c013240d
[  232.270838] Call Trace:
[  232.270842]  [<f8bcb705>] ? del_conn+0x5f/0x64 [bluetooth]
[  232.270854]  [<c012f90d>] ? run_workqueue+0x78/0xfb
[  232.270859]  [<f8bcb6a6>] ? del_conn+0x0/0x64 [bluetooth]
[  232.270869]  [<c0130119>] ? ...
From: Rafael J. Wysocki
Date: Sunday, June 1, 2008 - 4:25 pm

From: Justin Mattock
Date: Sunday, June 1, 2008 - 6:02 pm

-- 
Justin P. Mattock
--

From: Pavel Machek
Date: Monday, June 2, 2008 - 3:21 pm

hci_usb... that beast is known to corrupt random data on suspend and

...and we get oops in bluetooth.

Use btusb, instead.

Actually, I guess we should mark hci_usb as broken. I don't see why we
should debug the same problem over and over again.
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--

From: Justin Mattock
Date: Monday, June 2, 2008 - 4:12 pm

Well; "shizzle my frizzle," The funny thing is, is I have not seen
this after performing more suspends,
maybe this is a every once in a while scenario.
regards;

-- 
Justin P. Mattock
--

From: Pavel Machek
Date: Monday, June 2, 2008 - 11:43 pm

It was "one in three" in my tests; but you need usb actively in use
while you are suspending for this to bite.
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--

From: Justin Mattock
Date: Tuesday, June 3, 2008 - 1:30 am

Aahhhh the beast has been discovered.
That's pretty cool!! I quess from my side, I'll try what you had
suggested "btusb".
("I just have to allocate enough time to do so, ("come on people", get
out of my way!!"))
I'm finding myself in one of those days where I'm away from my laptop
(contingency or not).
regards;

-- 
Justin P. Mattock
--

From: Arjan van de Ven
Date: Sunday, June 1, 2008 - 7:01 pm

On Sun, 1 Jun 2008 23:13:51 +0000

   ^^^^^

                                        ^^^^^^^^^^


actually it's very serious, something is scribbling all over your
memory, it's just a matter of time before that overwrites something
important on your filesystem, and then you have a high risk of losing
most of your data.

However, since  you're using 2 proprietary kernel modules, you really
ought to contact the vendors for those modules for support; they're a
very likely cause for problems
(for example, these two are the cause for the number 1, 2, 5 and 7 of


-- 
If you want to reach me at my work email, use arjan@linux.intel.com
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org
--

From: Justin Mattock
Date: Sunday, June 1, 2008 - 8:53 pm

great!!!  Well The problem I have with these modules is A) madwifi
"waiting until support is in the kernel for this one, for now I
just have to deal with risking loosing any data". B) graphics driver I
can unload the driver but then loose suspend to mem "argh!!"
aside from that, I just performed the test again of simultaneously
suspending to memory and didn't receive this message, so now I'm at a
lost for words as to what is happening. The positive side is there
isn't any important data on my system i.g. important doc's etc..., but
there
is data with the applications.(alsa,acpid,etc...)
regards;
-- 
Justin P. Mattock
--

From: Arjan van de Ven
Date: Sunday, June 1, 2008 - 9:15 pm

On Mon, 2 Jun 2008 03:53:43 +0000

you mean the ath5k driver which already is in the kernel?

A current ATI driver in X.org ought to be close to getting this one


good luck.


-- 
If you want to reach me at my work email, use arjan@linux.intel.com
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org
--

From: Justin Mattock
Date: Sunday, June 1, 2008 - 9:44 pm

Thanks for the help: if the new athk5 supports model AR5418 then I'm a
go for that,
As for the rest of the issues "nice, now you have me freaked out," "I
guess I'll have to use OSX," but then If I do
I'll have to listen to this voice in my head saying " use Linux" "go
ahead Justin use Linux" "come on what are you chicken",
kind of like putting you're full day's worth of work in; and driving
on the 405, Then all of a sudden having a voice tell you
to go grab a nice cold refreshment(B$%^r)
regards;

-- 
Justin P. Mattock
--

Previous thread: [2.6.26-rc4] Problems with rt2x00 USB interface on powerpc by Rogério on Sunday, June 1, 2008 - 12:30 pm. (21 messages)

Next thread: [patch -v2 00/23] use simple_read_from_buffer and memory_read_from_buffer by akinobu.mita on Sunday, June 1, 2008 - 4:13 pm. (1 message)