BUG: sleeping function called from illegal context

i saw that one after modprobe snd-sb16...
kernel 2.5.67-bk

rgds
-daniel

Debug: sleeping function called from illegal context at mm/slab.c:1664

Call Trace:
 [<c0120777>] __might_sleep+0x5f/0x68
 [<c0148f35>] kmalloc+0x61/0x174
 [<c01292f7>] __request_region+0x1b/0xb4
 [<c01293aa>] __check_region+0x1a/0x40
 [<c01dc533>] pnp_check_port+0x5f/0x13c
 [<c01de28b>] pnp_manual_config_dev+0x12b/0x2fc
 [<c88c01f7>] snd_card_sb16_pnp+0x123/0x1bc [snd_sb16]
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c889906f>] +0x6f/0x59c [snd_sb16]
 [<c88c249c>] snd_sb16_pnpids+0x3dc/0x474 [snd_sb16]
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c88c249c>] snd_sb16_pnpids+0x3dc/0x474 [snd_sb16]
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c88c02c5>] snd_sb16_pnp_detect+0x35/0x64 [snd_sb16]
 [<c88c249c>] snd_sb16_pnpids+0x3dc/0x474 [snd_sb16]
 [<c01dae4b>] card_probe+0x3f/0x68
 [<c88c249c>] snd_sb16_pnpids+0x3dc/0x474 [snd_sb16]
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c01db8ff>] pnp_register_card_driver+0x17f/0x19c
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c88c1fc0>] port+0x0/0x20 [snd_sb16]
 [<c88996a7>] alsa_card_sb16_init+0x83/0xbc [snd_sb16]
 [<c88c2580>] sb16_pnpc_driver+0x0/0xb4 [snd_sb16]
 [<c88c2634>] possible_ports.1451+0x0/0x2c [snd_sb16]
 [<c889959c>] snd_sb16_probe_legacy_port+0x0/0x88 [snd_sb16]
 [<c88c2660>] +0x0/0xe0 [snd_sb16]
 [<c013f366>] sys_init_module+0x222/0x2d8
 [<c010b65f>] syscall_call+0x7/0xb

pnp: res: The resources requested do not match those set for the PnP device '01:01.00'.
sb16: no OPL device at 0x388-0x38a

Smatch found this bug also.
http://kbugs.org/cgi-bin/index.py?page=bug_list&script=SpinSleepLazy&skernel=2.5.67&sfile=sound

regards,
dan carpenter

Adam, could you consider to use mutexes rather than spinlocks in your code? I think that the pnp configuration is not time critical.

Jaroslav

-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs