Workqueue and serial link

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Use Vladistan GNU.IO.RXTX
2. Use RS232
3. Send and receive data

What is the expected output? What do you see instead?
The board freezes after a few seconds when sending/receiving adta through RS232:

WARNING: at kernel/workqueue.c:485 flush_cpu_workqueue+0x34/0x88()

Modules linked in: omaplfb libertas pvrsrvkm sdmak lpm_omap3530 dsplinkk cmemk

[<c003b9d8>] (unwind_backtrace+0x0/0xd8) from [<c0061374>] 
(warn_slowpath_common                                                           

+0x48/0x60)

[<c0061374>] (warn_slowpath_common+0x48/0x60) from [<c0070208>] 
(flush_cpu_workq                                                                

ueue+0x34/0x88)

[<c0070208>] (flush_cpu_workqueue+0x34/0x88) from [<c01be2e4>] 
(tty_ldisc_releas                                                               
                                                                   e+0x20/0x68)

[<c01be2e4>] (tty_ldisc_release+0x20/0x68) from [<c01b9af4>] 
(tty_release+0x3c0/                                                             
                                                                     0x420)

[<c01b9af4>] (tty_release+0x3c0/0x420) from [<c00b5b00>] (__fput+0x104/0x1d8)

[<c00b5b00>] (__fput+0x104/0x1d8) from [<c00b2c10>] (filp_close+0x6c/0x78)

[<c00b2c10>] (filp_close+0x6c/0x78) from [<c0062db8>] 
(put_files_struct+0x88/0xd                                                      
                                                                            c)

[<c0062db8>] (put_files_struct+0x88/0xdc) from [<c028e100>] 
(binder_deferred_fun                                                            

c+0x4e8/0x52c)

[<c028e100>] (binder_deferred_func+0x4e8/0x52c) from [<c0070724>] 
(worker_thread                                                                  
                                                                +0x150/0x1c4)

[<c0070724>] (worker_thread+0x150/0x1c4) from [<c0073794>] (kthread+0x78/0x80)

[<c0073794>] (kthread+0x78/0x80) from [<c0036ecc>] (kernel_thread_exit+0x0/0x8)

---[ end trace 4796369df1c2dc82 ]---

binder: 1282: binder_alloc_buf, no vma

binder: 941:968 transaction failed 29201, size 96-0

What version of the product are you using? On what operating system?
linux.2.6.32 kernel.

IGEP enhanced kernel.

Please provide any additional information below.

Original issue reported on code.google.com by ferre.ad...@gmail.com on 2 May 2011 at 9:17

[GoogleCodeExporter]

Hello,

I'm also experiencing this issue. Fresh Froyo source with a 2.6.32 kernel will 
crash after about 30 minutes.

Has anyone found a fix for this? Terrible the device cannot even stay on for 30 
minutes without locking up.

Original comment by kowalews...@gmail.com on 8 May 2011 at 4:11

[GoogleCodeExporter]

Hi,

you should patch the binder.c file using this patch: 
http://markmail.org/message/b5hwsqypdtpxf5mq

This worked for me.

Original comment by ferre.ad...@gmail.com on 8 May 2011 at 8:46

[GoogleCodeExporter]

Hello,

I tried that patch and unfortunately it did not work. The work queue messages 
do disappear, but I continuously get no vma errors.

Original comment by kowalews...@gmail.com on 8 May 2011 at 8:51