[Linux] Kernel crash.

[HJ]

Randomly crashes system with only solution is to hard-reset PC. Seemingly related to USB\XBox360 controller's vibration, borderlands 2 had similar issue.

Apr  2 02:05:39 Honnouji kernel: [  582.648323] ------------[ cut here ]------------
Apr  2 02:05:39 Honnouji kernel: [  582.648346] WARNING: CPU: 2 PID: 22545 at /build/linux-SAvLSw/linux-3.16.7-ckt7/drivers/usb/core/urb.c:339 usb_submit_urb+0x586/0x5a0 [usbcore]()
Apr  2 02:05:39 Honnouji kernel: [  582.648347] URB ffff88040c606640 submitted while active
Apr  2 02:05:39 Honnouji kernel: [  582.648348] Modules linked in: rfcomm cfg80211 snd_hrtimer snd_seq cpufreq_stats cpufreq_powersave cpufreq_userspace cpufreq_conservative binfmt_misc bnep 
pci_stub vboxpci(O) vboxnetadp(O) vboxnetflt(O) vboxdrv(O) nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc nls_utf8 nls_cp437 vfat fat xpad ff_memless uvcvideo videobuf2_vmallo
c ecb btusb videobuf2_memops videobuf2_core v4l2_common bluetooth snd_usb_audio videodev snd_usbmidi_lib 6lowpan_iphc snd_rawmidi snd_seq_device media joydev nvidia(PO) x86_pkg_temp_thermal i
ntel_powerclamp intel_rapl kvm_intel kvm crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw gf128mul eeepc_wmi asus_wmi sparse_keymap rfkill iTCO_wdt iTCO_vendor_support glue_helper 
snd_hda_codec_hdmi ablk_helper cryptd mxm_wmi evdev snd_hda_codec_realtek snd_hda_codec_generic snd_hda_intel snd_hda_controller psmouse snd_hda_codec pcspkr efi_pstore lpc_ich mei_me tpm_inf
ineon efivars serio_raw snd_hwdep snd_pcm snd_timer drm snd mfd_core wmi mei tpm_tis soundcore tpm battery button video shpchp i2c_i801 processor uinput nct6775 coretemp adt7475 hwmon_vid i2c
_core fuse autofs4 ext4 crc16 mbcache jbd2 hid_generic sg sd_mod crc_t10dif crct10dif_generic usbhid hid crct10dif_pclmul crct10dif_common crc32c_intel ahci ehci_pci xhci_hcd libahci ehci_hcd libata scsi_mod e1000e ptp usbcore pps_core usb_common thermal fan thermal_sys
Apr  2 02:05:39 Honnouji kernel: [  582.648380] CPU: 2 PID: 22545 Comm: ShadowWarrior.b Tainted: P           O  3.16.0-4-amd64 #1 Debian 3.16.7-ckt7-1
Apr  2 02:05:39 Honnouji kernel: [  582.648381] Hardware name: ASUS All Series/SABERTOOTH Z87, BIOS 1504 10/03/2013
Apr  2 02:05:39 Honnouji kernel: [  582.648382]  0000000000000009 ffffffff81509e7c ffff88041ed03d50 ffffffff81067727
Apr  2 02:05:39 Honnouji kernel: [  582.648383]  ffff88040c606640 ffff88041ed03da0 ffff88040aad8c00 ffff88041ed03e28
Apr  2 02:05:39 Honnouji kernel: [  582.648384]  ffff88040a68ced0 ffffffff8106778c ffffffffa007da08 ffffffff00000020
Apr  2 02:05:39 Honnouji kernel: [  582.648386] Call Trace:
Apr  2 02:05:39 Honnouji kernel: [  582.648386]  <IRQ>  [<ffffffff81509e7c>] ? dump_stack+0x41/0x51
Apr  2 02:05:39 Honnouji kernel: [  582.648393]  [<ffffffff81067727>] ? warn_slowpath_common+0x77/0x90
Apr  2 02:05:39 Honnouji kernel: [  582.648395]  [<ffffffff8106778c>] ? warn_slowpath_fmt+0x4c/0x50
Apr  2 02:05:39 Honnouji kernel: [  582.648401]  [<ffffffffa00694c6>] ? usb_submit_urb+0x586/0x5a0 [usbcore]
Apr  2 02:05:39 Honnouji kernel: [  582.648403]  [<ffffffff810abef2>] ? up+0x12/0x50
Apr  2 02:05:39 Honnouji kernel: [  582.648405]  [<ffffffffa058f705>] ? ml_play_effects+0x105/0x620 [ff_memless]
Apr  2 02:05:39 Honnouji kernel: [  582.648407]  [<ffffffffa058fd50>] ? ml_ff_playback+0xf0/0xf0 [ff_memless]
Apr  2 02:05:39 Honnouji kernel: [  582.648408]  [<ffffffffa058fd8a>] ? ml_effect_timer+0x3a/0x2b0 [ff_memless]
Apr  2 02:05:39 Honnouji kernel: [  582.648411]  [<ffffffff81072a01>] ? call_timer_fn+0x31/0x100
Apr  2 02:05:39 Honnouji kernel: [  582.648412]  [<ffffffffa058fd50>] ? ml_ff_playback+0xf0/0xf0 [ff_memless]
Apr  2 02:05:39 Honnouji kernel: [  582.648414]  [<ffffffff81074039>] ? run_timer_softirq+0x209/0x2f0
Apr  2 02:05:39 Honnouji kernel: [  582.648415]  [<ffffffff8106c5e1>] ? __do_softirq+0xf1/0x290
Apr  2 02:05:39 Honnouji kernel: [  582.648416]  [<ffffffff8106c9b5>] ? irq_exit+0x95/0xa0
Apr  2 02:05:39 Honnouji kernel: [  582.648418]  [<ffffffff81512e15>] ? smp_apic_timer_interrupt+0x45/0x60
Apr  2 02:05:39 Honnouji kernel: [  582.648420]  [<ffffffff81510efd>] ? apic_timer_interrupt+0x6d/0x80
Apr  2 02:05:39 Honnouji kernel: [  582.648421]  <EOI>  [<ffffffff811bae13>] ? do_select+0xb3/0x770
Apr  2 02:05:39 Honnouji kernel: [  582.648424]  [<ffffffff811bb42f>] ? do_select+0x6cf/0x770
Apr  2 02:05:39 Honnouji kernel: [  582.648426]  [<ffffffff814b1778>] ? unix_stream_recvmsg+0x388/0x7f0
Apr  2 02:05:39 Honnouji kernel: [  582.648429]  [<ffffffff81280b10>] ? blk_finish_plug+0x10/0x40
Apr  2 02:05:39 Honnouji kernel: [  582.648430]  [<ffffffff8109c060>] ? update_curr+0x90/0x120
Apr  2 02:05:39 Honnouji kernel: [  582.648433]  [<ffffffff8140340a>] ? sock_recvmsg+0x9a/0xd0
Apr  2 02:05:39 Honnouji kernel: [  582.648435]  [<ffffffff810d152b>] ? futex_wait_queue_me+0xdb/0x130
Apr  2 02:05:39 Honnouji kernel: [  582.648437]  [<ffffffff810d180e>] ? futex_wait+0x17e/0x260
Apr  2 02:05:39 Honnouji kernel: [  582.648438]  [<ffffffff81402b97>] ? ___sys_recvmsg+0x107/0x2a0
Apr  2 02:05:39 Honnouji kernel: [  582.648440]  [<ffffffff811f3cf5>] ? compat_core_sys_select+0x185/0x230
Apr  2 02:05:39 Honnouji kernel: [  582.648441]  [<ffffffff811f3e46>] ? compat_SyS_select+0xa6/0xf0
Apr  2 02:05:39 Honnouji kernel: [  582.648443]  [<ffffffff8151271a>] ? sysenter_dispatch+0x7/0x21
Apr  2 02:05:39 Honnouji kernel: [  582.648444] ---[ end trace df1b1d7fb51577b2 ]---

[blackout24]

Please use something like Github Gist, Pastebin or at least the CODE enviroment of the forums, the line breaks make it unreadable.

If this happens with other games, too you might want to file a bug against the Linux kernel instead.

[NoXPhasma]

That's a bug in the Unreal Engine 3, this happens when you have a gamepad connected to the PC. Happens to other UE3 ports (White Noise Online, Borderlands 2, Borderlands: TPS) too.

Would be great if that could be fixed by KnockoutGames, but I doubt that.

[HJ]

Sorry about not using pastebin, i'll keep in mind that next time.

Originally posted by NoXPhasma:

That's a bug in the Unreal Engine 3, this happens when you have a gamepad connected to the PC. Happens to other UE3 ports (White Noise Online, Borderlands 2, Borderlands: TPS) too.

Would be great if that could be fixed by KnockoutGames, but I doubt that.

Thanks for info! Is there more info about it?

[metalcaedes]

This is definitely a Kernel bug.

An unpriviliged userspace program must not be able to crash the Kernel.

[NoXPhasma]

Originally posted by HJ:

Thanks for info! Is there more info about it?

I have no further informations about this, except that this only happens to me when I have a gamepad plugged in and play a Unreal Engine 3 game. Never experianced this to other prorgams/games.

[metalcaedes]

I think http://www.spinics.net/lists/linux-input/msg31446.html might be related it's a kernel-patch for xpad that is supposed to fix crash that come from too much rumbling-requests.

For some ♥♥♥♥♥♥♥ reason no one even answered that post on the kernel mailing list, and of course didn't integrate that patch either.

Maybe from the game-side the problem could be worked around by doing less rumbling (or disable it completely).



EDIT: http://www.spinics.net/lists/linux-input/msg31450.html updated version of the patch, again no one answered...

[NoXPhasma]

What helps is to disable the rumble for the gamepad in the settings of Shadow Warrior. No more Kernel panics here.

Forget this, I get still Kernel Panics ...

[Skully]

Well this game doesn't use Ue3 anyhow. It uses the Road Hog engine.

Aside from that, due to the current mouse issues, I have only used my wireless 360 controller with this game and have not experienced any problems like this from the rumble. I am using the kernels xpad driver.

[NoXPhasma]

Originally posted by Skully:

Well this game doesn't use Ue3 anyhow. It uses the Road Hog engine.

Aside from that, due to the current mouse issues, I have only used my wireless 360 controller with this game and have not experienced any problems like this from the rumble. I am using the kernels xpad driver.

The question is, is it a bug in the xpad module with the wired xbox360 pad, or is it something else? It's really hard to say for me, I'm not a C programmer nor a Kernel hacker.

*update*
So, I get already a Kernel Panic while I'm in the menu, when I use the gamepad. No matter if Rumble is activated or not.

Just for completition:

Apr  3 02:58:35 localhost kernel: [  683.999188] ------------[ cut here ]------------
Apr  3 02:58:35 localhost kernel: [  683.999194] WARNING: CPU: 3 PID: 17749 at /build/buildd/linux-lts-utopic-3.16.0/drivers/usb/core/urb.c:339 usb_submit_urb+0x64/0x80()
Apr  3 02:58:35 localhost kernel: [  683.999196] URB ffff88006066d6c0 submitted while active
Apr  3 02:58:35 localhost kernel: [  683.999197] Modules linked in: xpad(OE) ff_memless pci_stub vboxpci(OE) vboxnetadp(OE) vboxnetflt(OE) vboxdrv(OE) ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp bridge stp llc ip6table_filter ip6_tables iptable_filter ip_tables x_tables binfmt_misc snd_usb_audio snd_usbmidi_lib snd_hwdep joydev nvidia(POE) eeepc_wmi asus_wmi sparse_keymap video snd_ctxfi snd_pcm snd_seq_midi snd_seq_midi_event snd_rawmidi dm_multipath scsi_dh kvm_amd kvm crct10dif_pclmul snd_seq crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw gf128mul glue_helper ablk_helper cryptd snd_seq_device serio_raw snd_timer fam15h_power edac_core edac_mce_amd k10temp snd sp5100_tco soundcore drm i2c_piix4 shpchp parport_pc ppdev tpm_infineon mac_hid it87 hwmon_vid lp parport dm_mirror dm_region_hash dm_log hid_generic usbhid hid mxm_wmi psmouse ahci r8169 libahci mii wmi
Apr  3 02:58:35 localhost kernel: [  683.999236] CPU: 3 PID: 17749 Comm: ShadowWarrior.b Tainted: P           OE 3.16.0-33-generic #44~14.04.1-Ubuntu
Apr  3 02:58:35 localhost kernel: [  683.999238] Hardware name: To be filled by O.E.M. To be filled by O.E.M./M5A97 R2.0, BIOS 2201 12/10/2013
Apr  3 02:58:35 localhost kernel: [  683.999239]  0000000000000009 ffff88022fcc3d00 ffffffff81763720 ffff88022fcc3d48
Apr  3 02:58:35 localhost kernel: [  683.999241]  ffff88022fcc3d38 ffffffff8106dd7d 0000000000000010 ffff88018bd2e308
Apr  3 02:58:35 localhost kernel: [  683.999243]  ffff88018bd2e000 ffff88022fcc3df0 000000000000ffff ffff88022fcc3d98
Apr  3 02:58:35 localhost kernel: [  683.999245] Call Trace:
Apr  3 02:58:35 localhost kernel: [  683.999246]  <IRQ>  [<ffffffff81763720>] dump_stack+0x45/0x56
Apr  3 02:58:35 localhost kernel: [  683.999252]  [<ffffffff8106dd7d>] warn_slowpath_common+0x7d/0xa0
Apr  3 02:58:35 localhost kernel: [  683.999255]  [<ffffffff8106ddec>] warn_slowpath_fmt+0x4c/0x50
Apr  3 02:58:35 localhost kernel: [  683.999257]  [<ffffffff810aa720>] ? update_curr+0x80/0x180
Apr  3 02:58:35 localhost kernel: [  683.999259]  [<ffffffff815692c4>] usb_submit_urb+0x64/0x80
Apr  3 02:58:35 localhost kernel: [  683.999262]  [<ffffffffc0d663b2>] xpad_play_effect+0x1b2/0x280 [xpad]
Apr  3 02:58:35 localhost kernel: [  683.999264]  [<ffffffffc0d567c2>] ml_play_effects+0x102/0x600 [ff_memless]
Apr  3 02:58:35 localhost kernel: [  683.999267]  [<ffffffffc0d56e00>] ? ml_ff_playback+0x100/0x100 [ff_memless]
Apr  3 02:58:35 localhost kernel: [  683.999269]  [<ffffffffc0d56e3f>] ml_effect_timer+0x3f/0x200 [ff_memless]
Apr  3 02:58:35 localhost kernel: [  683.999271]  [<ffffffff8107a496>] call_timer_fn+0x36/0x100
Apr  3 02:58:35 localhost kernel: [  683.999273]  [<ffffffffc0d56e00>] ? ml_ff_playback+0x100/0x100 [ff_memless]
Apr  3 02:58:35 localhost kernel: [  683.999275]  [<ffffffff8107bc4f>] run_timer_softirq+0x20f/0x310
Apr  3 02:58:35 localhost kernel: [  683.999277]  [<ffffffff810730c5>] __do_softirq+0xf5/0x2e0
Apr  3 02:58:35 localhost kernel: [  683.999279]  [<ffffffff81073585>] irq_exit+0x105/0x110
Apr  3 02:58:35 localhost kernel: [  683.999282]  [<ffffffff8176ed55>] smp_apic_timer_interrupt+0x45/0x60
Apr  3 02:58:35 localhost kernel: [  683.999284]  [<ffffffff8176cdfd>] apic_timer_interrupt+0x6d/0x80
Apr  3 02:58:35 localhost kernel: [  683.999285]  <EOI>  [<ffffffff8176e88c>] ? cstar_dispatch+0x7/0x21
Apr  3 02:58:35 localhost kernel: [  683.999288] ---[ end trace 68005acc2dcc3e17 ]---

[HJ]

Originally posted by NoXPhasma:

Originally posted by Skully:

Well this game doesn't use Ue3 anyhow. It uses the Road Hog engine.

Aside from that, due to the current mouse issues, I have only used my wireless 360 controller with this game and have not experienced any problems like this from the rumble. I am using the kernels xpad driver.

The question is, is it a bug in the xpad module with the wired xbox360 pad, or is it something else? It's really hard to say for me, I'm not a C programmer nor a Kernel hacker.

I remember on WindowsⓇ original wireless 360 controller is different story (and drivers) than wired. So it could be different in linux.

Problem is - I don't even use joypad, it is just connected to PC.

[NoXPhasma]

Yeah, probably the game tries to do something to the gamepad which let the xpad kernel module crash. It is really weird, because this only happens to a few games. Most games work fine for me. No matter if Gamepad only connected or used. As I said I saw this almost only on UE3 games before and now with this game.

I've also tried the SteamOS xpad Kernel module, without success, still Kernel Panics.

[wintermute]

I came across this problem too. I had a USB XBox controller plugged in and the kernel crashed when I attempted to change any item in the settings. Everything froze, couldn't even SSH in. This happened twice in the Video settings menu, had to reset the computer.

After reading this I unplugged the controller and tried again, no crashes.

[metalcaedes]

That patch I mentioned before indeed works (at least for me and NoXPhasma)

I've updated it so it applies to the latest kernel source.

How to use it:
In a Terminal do (the lines starting with "#" are comments and must not be executed in the shell):

mkdir xpad-impr
cd xpad-impr
# get the latest xpad.c from the official kernel source
wget https://github.com/torvalds/linux/raw/f20c86cd75f1c8c728dafd0218645ff3c5e8545d/drivers/input/joystick/xpad.c
# get the patch from Sarah Bessmer (updated by me)
wget -O sarah.patch http://pastebin.com/raw.php?i=JbyxBZeM
# get Makefile
wget -O Makefile http://pastebin.com/raw.php?i=HRqeQZ92

# for some stupid reason pastebin removes trailing newlines, add them again
echo -e "\n\n" >> sarah.patch

# apply Sarah's patch
patch -p1 -i sarah.patch

# build the kernel module
make 

# make sure it worked:
ls xpad.ko

# unload the old xpad module
sudo rmmod xpad

# load the new one
sudo insmod xpad.ko

Then start the game and play and see if it still crashes or still produces those nasty warnings in dmesg.

Note that after a reboot the module will be gone again and you'll need to rmmod the old one and insmod the new one again.

This is mostly for testing, maybe someone else can cook something nice with dkms for a more permanent solution.

And if this works for everyone, this patch should *really* get into the kernel.

[NoXPhasma]

Yes, I can confirm that this patch works fine here on Linux Mint 17.1 with Kernel 3.16.

*edit*
For automatically building this module for future kernel updates and permanent install at the curent one, follow this guide: http://wiki.centos.org/HowTos/BuildingKernelModules#head-d313bd351f90d4f25a2143b7bbcff73f927731f0

1. Create a new dkms.conf file in the folder which contains the patched xpad.c and the make file, with this content:

PACKAGE_NAME="xpad"
PACKAGE_VERSION="0.3"
MAKE[0]="make KVERSION=$kernelver"
CLEAN="make clean"
DEST_MODULE_LOCATION[0]="/extra"
BUILT_MODULE_NAME[0]="xpad"
AUTOINSTALL="yes"

The folder should only contain this 3 files.
2. Copy the folder to /usr/src.
3. execute this 3 commands as root to build and install the module:

dkms add -m xpad -v 0.3
dkms build -m xpad -v 0.3
dkms install -m xpad -v 0.3

Now this module is installed and will be build on each kernel update automatically.

[metalcaedes]

Can anyone else test this?


If it works for everyone, I'll ask the original author of the patch to try (again) to get it into the kernel.