I am using 3 different digi products and using their out of kernel GPL drivers.
I am trying to get them to work with the 2.6.31.4 kernel. The digi people are
not really doing much to keep them current so I have to do it myself if I want
to run the latest kernels. I'm going from a 2.6.26 series to 2.6.31 BTW.
I have them all working but all 3 cause the following WARNING when I open the
ports on the board. Then one of the 3 causes the BUG msg below when I close
ports at task exit time and the task can't exit. They do function correctly
while the task is running BTW.
I know there have been many changes to the serial/tty stuff recently. I think
this is not a driver but kernel issue so I thought I'd bring it up here in hope
to find out for sure.
I know my Tainted flag is set here as I do use the nvidia driver but I have it
unloaded and do have other machines where this happens not using it at all.
WARNING: at lib/kref.c:43 kref_get+0x18/0x20()
Hardware name: Unknow
Modules linked in: rtom snd_pcm_oss snd_mixer_oss binfmt_misc snd_seq
snd_seq_device iptable_filter ip_tables ip6table_filter ip6_tables x_tables ipv6
fuse loop dm_mod tnt4882 synclink_gt snd_hda_codec_realtek nec7210 snd_hda_intel
dgap hdlc 3c59x snd_hda_codec dgdm st mii gpib_common osst snd_pcm snd_timer snd
rtc_cmos shpchp rtc_core pci_hotplug i2c_piix4 soundcore wmi sr_mod ati_agp
snd_page_alloc cdrom sky2 rtc_lib agpgart i2c_core sg sd_mod edd ext3 mbcache
jbd aic79xx aic7xxx scsi_transport_spi ide_pci_generic atiixp ide_core
ata_generic pata_atiixp ahci libata scsi_mod thermal processor thermal_sys [last
unloaded: rfm2g]
Pid: 5364, comm: vrsx Tainted: P 2.6.31.4 #1
Call Trace:
[<c0125281>] ? warn_slowpath_common+0x5e/0x8a
[<c01252b7>] ? warn_slowpath_null+0xa/0xc
[<c0208108>] ? kref_get+0x18/0x20
[<c025a5ec>] ? tty_open+0x128/0x384
[<c0197553>] ? chrdev_open+0x15a/0x171
[<c0193a75>] ? __dentry_open+0x14f/0x23e
[<c0193bf5>] ? nameidata_to_filp+0x29/0x3c
[<c01973f9>] ? chrdev_open+0x0/0x171
[<c019ebd1>] ? do_filp_open+0x418/0x7fb
[<c0313a74>] ? _spin_unlock+0xc/0x1f
[<c018fe10>] ? __slab_free+0xa8/0x286
[<c01832cf>] ? remove_vma+0x3e/0x43
[<c0313a74>] ? _spin_unlock+0xc/0x1f
[<c01a606b>] ? alloc_fd+0xc3/0xcd
[<c019382f>] ? do_sys_open+0x4c/0xe4
[<c019390b>] ? sys_open+0x1e/0x23
[<c01027f4>] ? sysenter_do_call+0x12/0x26
---[ end trace 1963594fafbe52d2 ]---
After task starts it shows 24 opens
#lsmod | grep dgdm
dgdm 300712 24
kernel BUG at mm/slub.c:2929!
invalid opcode: 0000 [#1] PREEMPT SMP
last sysfs file: /sys/firmware/edd/int13_dev80/extensions
Modules linked in: af_packet n_hdlc rtom snd_pcm_oss snd_mixer_oss binfmt_misc
snd_seq snd_seq_device iptable_filter ip_tables ip6table_filter ip6_tables
x_tables ipv6 fuse loop dm_mod tnt4882 synclink_gt snd_hda_codec_realtek nec7210
snd_hda_intel dgap hdlc 3c59x snd_hda_codec dgdm st mii gpib_common osst snd_pcm
snd_timer snd rtc_cmos shpchp rtc_core pci_hotplug i2c_piix4 soundcore wmi
sr_mod ati_agp snd_page_alloc cdrom sky2 rtc_lib agpgart i2c_core sg sd_mod edd
ext3 mbcache jbd aic79xx aic7xxx scsi_transport_spi ide_pci_generic atiixp
ide_core ata_generic pata_atiixp ahci libata scsi_mod thermal processor
thermal_sys [last unloaded: rfm2g]
Pid: 5364, comm: vrsx Tainted: P W (2.6.31.4-lcrs #1) Unknow
EIP: 0060:[<c01910db>] EFLAGS: 00010246 CPU: 3
EIP is at kfree+0x79/0xd5
EAX: 80000000 EBX: c16d9620 ECX: f2318c60 EDX: 00000000
ESI: c02080e6 EDI: f6cb13cc EBP: 00000047 ESP: f120de60
DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
Process vrsx (pid: 5364, ti=f120c000 task=f1cf0000 task.ti=f120c000)
Stack:
f2291000 f6cb13d0 c0258646 f6cb13cc 00000047 c02080e6 f1bc4800 f1bc4804
<0> c025a339 f1bc4804 c025a303 00000000 c02080e6 f1bc4800 00000000 c025a2b4
<0> f2a3b080 f1bc491c f1bc4910 0000011c 00000110 00000000 00000001 00000000
Call Trace:
[<c0258646>] ? destruct_tty_driver+0x0/0x81
[<c02080e6>] ? kref_put+0x36/0x40
[<c025a339>] ? release_one_tty+0x36/0x7f
[<c025a303>] ? release_one_tty+0x0/0x7f
[<c02080e6>] ? kref_put+0x36/0x40
[<c025a2b4>] ? tty_release_dev+0x3b9/0x3d3
[<c019a822>] ? free_pipe_info+0xe/0x1a
[<c025a2dd>] ? tty_release+0xf/0x18
[<c0195f28>] ? __fput+0xc0/0x16a
[<c0193742>] ? filp_close+0x4e/0x54
[<c012675d>] ? put_files_struct+0x60/0xa6
[<c0127aef>] ? do_exit+0x1b2/0x59c
[<c03158a0>] ? do_page_fault+0x359/0x36f
[<c0127f34>] ? do_group_exit+0x5b/0x82
[<c0127f6c>] ? sys_exit_group+0x11/0x14
[<c01027f4>] ? sysenter_do_call+0x12/0x26
Code: ff 10 76 7d 8d 9f 00 00 00 40 c1 eb 0c c1 e3 05 03 1d 40 fa 4e c0 66 83 3b
00 79 03 8b 5b 0c 8b 03 84 c0 78 16 66 a9 00 c0 75 04 <0f> 0b eb fe 5a 89 d8 5b
5e 5f 5d e9 d9 3f fe ff 8b 44 24 14 89
EIP: [<c01910db>] kfree+0x79/0xd5 SS:ESP 0068:f120de60
---[ end trace 1963594fafbe52d3 ]---
Fixing recursive fault but reboot is needed!
After task attempts to exit there are still 2 opens left
# lsmod | grep dgdm
dgdm 300712 2
Thanks in advance
Mark
> I know there have been many changes to the serial/tty stuff recently. I think
> this is not a driver but kernel issue so I thought I'd bring it up here in hope
> to find out for sure.
Its probably a combination of the two. While digi persist in keeping
their stuff out of tree its always going to lag and not get fixed in
parallel or even considered during tty reworking.
First guess would be that the driver has paths that free a tty kref but
don't take it (ie unbalanced counts) second that may be it can't cope
with the changes caused by the switches to refcounting throughout the tty
code.
One for Digi to answer given they insist on keeping it out of tree.