[etherlab-users] reg_read causes kernel panic in stable-1.5 in xenomai

[George Broz]

Hello,

If I issue ethercat reg_read at the command line several times a 
kernel oops results. The master is in "IDLE" phase during this time.

I am running Xenomai 2.5.6 with Linux 2.6.37.6, using the 
IgH EtherCAT master 1.5.0 ec8e1151b8a7 (2266) with the e1000e native driver.

I originally encountered this issue after applying Graeme's patch (see thread at
http://lists.etherlab.org/pipermail/etherlab-users/2012/001578.html) but I see
it happening without the patch, i.e. on the standard ec8e1151b8a7 (2266) fileset.

There is not a problem when I use the most recent "default" release 
(changeset 1798bcdaa8d0 (2271) on the "default" branch). Is the "default" 
branch intended for Xenomai users?


Here is a typical dump from from the kernel oops:

Feb 15 11:05:13 nxtgenhd kernel: [  157.752034] eth2: no IPv6 routers present
Feb 15 11:05:36 nxtgenhd kernel: [  180.580069] BUG: unable to handle kernel paging request at 84dd917f
Feb 15 11:05:36 nxtgenhd kernel: [  180.580217] IP: [<c01f98e7>] __kmalloc+0xc7/0x200
Feb 15 11:05:36 nxtgenhd kernel: [  180.580324] *pde = 00000000
Feb 15 11:05:36 nxtgenhd kernel: [  180.580390] Oops: 0000 [#1] SMP
Feb 15 11:05:36 nxtgenhd kernel: [  180.580470] last sysfs file: /sys/bus/pci/drivers/ec_e1000e/uevent
Feb 15 11:05:36 nxtgenhd kernel: [  180.580586] Modules linked in: ec_e1000e ec_master binfmt_misc i915 drm_kms_helper drm xeno_native xeno_posix ppdev xeno_rt
dm xeno_nucleus i2c_algo_bit intel_agp parport_pc i2c_core psmouse intel_gtt video serio_raw agpgart lp output parport usbhid hid ata_piix libata [last unloade
d: e1000e]
Feb 15 11:05:36 nxtgenhd kernel: [  180.581284]
Feb 15 11:05:36 nxtgenhd kernel: [  180.581321] Pid: 1977, comm: EtherCAT-IDLE Not tainted 2.6.37.6.010512-xen2.5.6 #7 Intel Corporation Lobster Rock/To be fil
led by O.E.M.
Feb 15 11:05:36 nxtgenhd kernel: [  180.581582] EIP: 0060:[<c01f98e7>] EFLAGS: 00010082 CPU: 2
Feb 15 11:05:36 nxtgenhd kernel: [  180.581687] EIP is at __kmalloc+0xc7/0x200
Feb 15 11:05:36 nxtgenhd kernel: [  180.581766] EAX: f6e103b0 EBX: 00000004 ECX: 84dd917f EDX: 00000000
Feb 15 11:05:36 nxtgenhd kernel: [  180.581882] ESI: f6002200 EDI: 000000d0 EBP: f2f4ff50 ESP: f2f4ff24
Feb 15 11:05:36 nxtgenhd kernel: [  180.581998]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
Feb 15 11:05:36 nxtgenhd kernel: [  180.582099] Process EtherCAT-IDLE (pid: 1977, ti=f2f4e000 task=f2ed1590 task.ti=f2f4e000)
Feb 15 11:05:36 nxtgenhd kernel: [  180.582246] I-pipe domain Linux
Feb 15 11:05:36 nxtgenhd kernel: [  180.582307] Stack:
Feb 15 11:05:36 nxtgenhd kernel: [  180.582349]  eef003c0 f60fd7b0 f2f4ff34 f80f02c1 f2f4ff50 c01f8ab7 84dd917f 00000200
Feb 15 11:05:36 nxtgenhd kernel: [  180.582558]  f25bfebc f26ae97c f26aebe0 f2f4ff74 f80f02c1 f2f4ff60 c01977cd f2f4ff68
Feb 15 11:05:36 nxtgenhd kernel: [  180.582766]  f26ae800 f26ae800 f26aee24 f26accb8 f2f4ff7c f80ee6df f2f4ffb8 f80fa70a
Feb 15 11:05:36 nxtgenhd kernel: [  180.582974] Call Trace:
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<f80f02c1>] ? ec_fsm_master_state_reg_request+0xb1/0x180 [ec_master]
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c01f8ab7>] ? kfree+0xb7/0x140
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<f80f02c1>] ? ec_fsm_master_state_reg_request+0xb1/0x180 [ec_master]
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c01977cd>] ? __ipipe_restore_root+0x2d/0x40
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<f80ee6df>] ? ec_fsm_master_exec+0x1f/0x30 [ec_master]
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<f80fa70a>] ? ec_master_idle_thread+0x9a/0x1b0 [ec_master]
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c01977cd>] ? __ipipe_restore_root+0x2d/0x40
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<f80fa670>] ? ec_master_idle_thread+0x0/0x1b0 [ec_master]
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c01592c4>] ? kthread+0x74/0x80
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c0159250>] ? kthread+0x0/0x80
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023]  [<c01034fe>] ? kernel_thread_helper+0x6/0x10
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023] Code: 80 69 c0 8d 80 a0 91 69 c0 f0 0f ba 28 00 9d 8b 06 64 03 05 9c 80 69 c0 8b 10 85 d2 89 55 ec 0f 84 1d 01
00 00 8b 56 10 8b 4d ec <8b> 14 11 89 10 8b 4d f0 85 c9 74 55 31 c0 e8 a6 de f9 ff 8b 55
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023] EIP: [<c01f98e7>] __kmalloc+0xc7/0x200 SS:ESP 0068:f2f4ff24
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023] CR2: 0000000084dd917f
Feb 15 11:05:36 nxtgenhd kernel: [  180.584023] ---[ end trace 82dbf39d2516cd7f ]---
root at nxtgenhd:~/src/igh/etherlabmaster-ec8e1151b8a7#



Is anyone out there using Xenomai with stable-1.5? Can anyone suggest something to try/fix? 
(I could port the differences from the "default" to the "stable-1.5" branch for the functions
in the trace but I would be doing that blindly!)


Thanks in advance!
--George Broz
Moog Inc. Industrial Group