BUG: scheduling while atomic

BUG: scheduling while atomic

[Mushtaq Khan]

Hi,
 I see error message "BUG: scheduling while atomic"intermittently. As
per other post on similar  issue, this happens when schedule is called
from either interrupt
 context or some other atomic operation (Might be after acquiring
spinlock). But from the call trace am not getting any clue, please
help me guide  on this issue.
Using Linux 3.4rt.

 BUG: scheduling while atomic: dslmgmt/538/0x00000201
 Modules linked in: iptable_mangle init_addr(  (null) -   (null)),
core_addr(c036
 9000 - c0369210)
  iptable_filter init_addr(  (null) -   (null)), core_addr(c0362000 - c03620b0)
  ip_tables init_addr(  (null) -   (null)), core_addr(c0359000 - c035b618)
  xt_multiport init_addr(  (null) -   (null)), core_addr(c034f000 - c034f2d8)
  xt_mark init_addr(  (null) -   (null)), core_addr(c0348000 - c0348098)
  xt_mac init_addr(  (null) -   (null)), core_addr(c0342000 - c03420b4)
  xt_DSCP init_addr(  (null) -   (null)), core_addr(c033c000 - c033c388)
  xt_dscp init_addr(  (null) -   (null)), core_addr(c0335000 - c0335128)
  pwrmngtd(P) init_addr(  (null) -   (null)), core_addr(c032e000 - c032f54c)
  bcmvlan(P) init_addr(  (null) -   (null)), core_addr(c030a000 - c031a598)
  p8021ag(P) init_addr(  (null) -   (null)), core_addr(c02e0000 - c02e1120)
  bcmarl(P) init_addr(  (null) -   (null)), core_addr(c02d7000 - c02d7e74)
  nciTMSkmod(P) init_addr(  (null) -   (null)), core_addr(c0280000 - c029ffc8)
  bcm_enet init_addr(  (null) -   (null)), core_addr(c01e4000 - c02078f4)
  adsldd(P) init_addr(  (null) -   (null)), core_addr(c011b000 - c0150c78)
  bcmxtmcfg(P) init_addr(  (null) -   (null)), core_addr(c009a000 - c00a5e28)
  pktflow(P) init_addr(  (null) -   (null)), core_addr(c0067000 - c0074800)
  bcm_bpm(P) init_addr(  (null) -   (null)), core_addr(c0045000 - c0046470)
  bcm_ingqos(P) init_addr(  (null) -   (null)), core_addr(c003a000 - c003b070)
  chipinfo(P) init_addr(  (null) -   (null)), core_addr(c0031000 - c0031198)
 Call Trace:
 [<802a0d6c>] printk+0x24/0x30
 [<80074714>] print_modules+0x88/0x144
 [<802a43c8>] __schedule+0x54c/0x58c
 [<802a12d0>] __schedule_bug+0x38/0x54
 [<802a0b30>] dump_stack+0x8/0x34
 [<802a43c8>] __schedule+0x54c/0x58c
 [<8017d0f0>] do_output_char+0x1b4/0x220
 [<801876ec>] uart_write+0xd8/0x15c
 [<8017d19c>] process_output+0x40/0x6c
 [<802a4580>] schedule+0x58/0xb4
 [<800f1dc0>] fsnotify+0x1b0/0x2dc
 [<802a3514>] do_nanosleep+0xf8/0x13c
 [<80055c98>] hrtimer_nanosleep+0xb4/0x158
 [<80054dd8>] hrtimer_wakeup+0x0/0x2c
 [<80044f00>] sys_rt_sigaction+0x78/0xd4
 [<80055da8>] sys_nanosleep+0x6c/0x94
 [<8001b95c>] stack_done+0x20/0x40

 Call Trace:
 [<802a0b30>] dump_stack+0x8/0x34
 [<802a43c8>] __schedule+0x54c/0x58c
 [<802a4580>] schedule+0x58/0xb4
 [<802a3514>] do_nanosleep+0xf8/0x13c
 [<80055c98>] hrtimer_nanosleep+0xb4/0x158
 [<80055da8>] sys_nanosleep+0x6c/0x94
 [<8001b95c>] stack_done+0x20/0x40

 Regards,
 Mushtaq Khan

Re: BUG: scheduling while atomic

[Sebastian Andrzej Siewior]

* Mushtaq Khan | 2014-09-17 14:42:30 [+0530]:

>Hi,
> I see error message "BUG: scheduling while atomic"intermittently. As
>per other post on similar  issue, this happens when schedule is called
>from either interrupt
> context or some other atomic operation (Might be after acquiring
>spinlock). But from the call trace am not getting any clue, please
>help me guide  on this issue.
>Using Linux 3.4rt.

That P in your drivers suggests that you have non GPL code in your
kernel.
You could increase the debugging level and check who disabled preemption
and then working around it.

> BUG: scheduling while atomic: dslmgmt/538/0x00000201
>  pwrmngtd(P) init_addr(  (null) -   (null)), core_addr(c032e000 - c032f54c)
>  bcmvlan(P) init_addr(  (null) -   (null)), core_addr(c030a000 - c031a598)
>  p8021ag(P) init_addr(  (null) -   (null)), core_addr(c02e0000 - c02e1120)
>  bcmarl(P) init_addr(  (null) -   (null)), core_addr(c02d7000 - c02d7e74)
>  nciTMSkmod(P) init_addr(  (null) -   (null)), core_addr(c0280000 - c029ffc8)
>  bcm_enet init_addr(  (null) -   (null)), core_addr(c01e4000 - c02078f4)
>  adsldd(P) init_addr(  (null) -   (null)), core_addr(c011b000 - c0150c78)
>  bcmxtmcfg(P) init_addr(  (null) -   (null)), core_addr(c009a000 - c00a5e28)
>  pktflow(P) init_addr(  (null) -   (null)), core_addr(c0067000 - c0074800)
>  bcm_bpm(P) init_addr(  (null) -   (null)), core_addr(c0045000 - c0046470)
>  bcm_ingqos(P) init_addr(  (null) -   (null)), core_addr(c003a000 - c003b070)
>  chipinfo(P) init_addr(  (null) -   (null)), core_addr(c0031000 - c0031198)

Sebastian