Badness in local_bh_enable at kernel/softirq.c:137 (linux_denx_2.6.19.2)

Sending ioctls to a socket I receive the following problem (only the
first time I send a ioctl):
 
Badness in local_bh_enable at kernel/softirq.c:137
Call Trace:
[C0593C90] [C000A6E8] show_stack+0x48/0x190 (unreliable)
[C0593CC0] [C0003734] check_bug_trap+0xa4/0xb8
[C0593CD0] [C0004354] program_check_exception+0x2e4/0x4fc
[C0593CF0] [C0003344] ret_from_except_full+0x0/0x4c
[C0593DB0] [C01A1BD4] klist_iter_exit+0x28/0x44
[C0593DC0] [C01198E0] phy_read+0x40/0x58
[C0593DE0] [C0119A94] phy_mii_ioctl+0x70/0x18c
[C0593E00] [C011BFCC] fs_ioctl+0x60/0xb4
[C0593E20] [C0146C30] dev_ifsioc+0x174/0x3dc
[C0593E40] [C0147784] dev_ioctl+0x468/0x4d0
[C0593EB0] [C013A1D0] sock_ioctl+0x38/0x210
[C0593ED0] [C009C8D0] do_ioctl+0x38/0x84
[C0593EE0] [C009C9A8] vfs_ioctl+0x8c/0x418
[C0593F10] [C009CD74] sys_ioctl+0x40/0x74
[C0593F40] [C0002CE4] ret_from_syscall+0x0/0x38
 
 
 

DI BACCO ANTONIO - technolabs wrote:

Sending ioctls to a socket I receive the following problem (only the
first time I send a ioctl):
 
Badness in local_bh_enable at kernel/softirq.c:137

You're not supposed to nest a spin_lock_bh inside a spin_lock_irq. 
fs_ioctl() does this anyway, hence the warning.

-Scott