iegd 10.3.1 "bad_page" error message on wrslinux ver3.0

[HWada1]

Hello everybody.

My board is KTUS15 by Kontron. (CPU: ATOM Z510, Memory:2GHz)

I implemented the iegd driver to windriver linux ver3.0(Kernel version is 2.6.27).

And I tried to run X server.

It seemed that X server was run well. But the following message was logged in /var/log/messages.

Is this message fatal error? Or can this error be disregarded ?

The contents of error message is following,

Jun 3 02:42:26 localhost kernel: Pid: 2370, comm: X Not tainted 2.6.27.39-WR3.0.2ax_standard-00073-g6471dad-dirty # 11Jun 3 02:42:26 localhost kernel: [] bad_page+0x73/0xc0Jun 3 02:42:26 localhost kernel: [] free_hot_cold_page+0x1ef/0x260Jun 3 02:42:26 localhost kernel: [] agp_generic_destroy_page+0x49/0x70Jun 3 02:42:26 localhost kernel: [] agp_free_memory+0x87/0x100Jun 3 02:42:26 localhost kernel: [] agpioc_deallocate_wrap+0x35/0x50Jun 3 02:42:26 localhost kernel: [] agp_ioctl+0x456/0x530Jun 3 02:42:26 localhost kernel: [] __slab_free+0x68/0x310Jun 3 02:42:26 localhost kernel: [] unix_release_sock+0x1a7/0x2e0Jun 3 02:42:26 localhost kernel: [] skb_dequeue+0x4e/0x80Jun 3 02:42:26 localhost kernel: [] d_kill+0x3d/0x60Jun 3 02:42:26 localhost last message repeated 2 timesJun 3 02:42:26 localhost kernel: [] agp_ioctl+0x0/0x530Jun 3 02:42:26 localhost kernel: [] vfs_ioctl+0x2b/0x90Jun 3 02:42:26 localhost kernel: [] do_vfs_ioctl+0x67/0x2f0Jun 3 02:42:26 localhost kernel: [] fget_light+0xba/0xf0Jun 3 02:42:26 localhost kernel: [] sys_ioctl+0xca/0xe0Jun 3 02:42:26 localhost kernel: [] system_call_done+0x0/0x4

(Please don't care about date. Because my evalution board don't have battery.)

[Kirk_B_Intel · Accepted solution]

There is some interesting stuff that goes on "under the hood" sometimes. This looks like it may be related to an oddity in the memory-mapped IO handling of the US15w. If there was an actual problem in these areas, it would be VERY obvious as it would crash the system. Because you have X running, I believe you can safely ignore the warnings and operate in safety.

As a side note, you might want to take a look at the EMGD 1.0 release that is also now available as that implements a more efficient architecture for US15w (and future related products) and *may* have implemented this area in a more elegant way to avoid the issue all together. http://edc.intel.com/Software/Downloads/EMGD/ http://edc.intel.com/Software/Downloads/EMGD/" rel="nofollow">http://edc.intel.com/Software/Downloads/EMGD/

Hope this helps.

[HWada1]

Hello Kirk-san,

Thanks for your response!

I am so glad to know the reason(root cause) of the problem.

EMGD driver doesn't support Wind RiverLinux. But I challenge to run it on Wind River Linux.