On Thu, Dec 11, 2008 at 01:10:23PM +0300, Alexander Beregalov wrote: > BUG: sleeping function called from invalid context at mm/slub.c:1619 > in_atomic(): 1, irqs_disabled(): 0, pid: 8096, name: sh > INFO: lockdep is turned off. > Pid: 8096, comm: sh Tainted: G W 2.6.28-rc7-next-20081210 #2 > Call Trace: > [<ffffffff80256e5a>] ? __debug_show_held_locks+0x1b/0x24 > [<ffffffff8022fb55>] __might_sleep+0x120/0x122 > [<ffffffff8029c526>] __kmalloc+0x70/0xfc > [<ffffffff803534d2>] kmem_alloc+0x75/0xd7 > [<ffffffff803350ba>] xfs_fs_vcmn_err+0x46/0x94 > [<ffffffff80335297>] xfs_fs_cmn_err+0x59/0x5e xfs_fs_vcmn_err does a sleeping memory allocation under a spinlock, nasty. I have an idea that will get rid of the memory allocation completely, I'll come back to you once I have hacked something up. > Filesystem "sda2": xlog_space_left: head behind tail > tail_cycle = 1802201963, tail_bytes = -690563584 > GH cycle = 3078, GH bytes = 69183384 This btw is a bug, too.. -- To unsubscribe from this list: send the line "unsubscribe linux-next" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html