Description of problem: [ 1773.547262] ======================================================= [ 1773.555610] [ INFO: possible circular locking dependency detected ] [ 1773.562162] 3.1.0-0.rc3.git0.0.fc16.x86_64 #1 [ 1773.566802] ------------------------------------------------------- [ 1773.573365] disktest/6751 is trying to acquire lock: [ 1773.578613] (&sb->s_type->i_mutex_key#15){+.+.+.}, at: [<ffffffff811e59d0>] hugetlbfs_file_mmap+0x82/0x10a [ 1773.589216] [ 1773.589218] but task is already holding lock: [ 1773.595650] (&mm->mmap_sem){++++++}, at: [<ffffffff811181d1>] sys_mmap_pgoff+0xf8/0x16a [ 1773.604497] [ 1773.604499] which lock already depends on the new lock. [ 1773.604500] [ 1773.613572] [ 1773.613575] the existing dependency chain (in reverse order) is: [ 1773.621631] [ 1773.621632] -> #1 (&mm->mmap_sem){++++++}: [ 1773.628004] [<ffffffff8108eff1>] lock_acquire+0xf3/0x13e [ 1773.634282] [<ffffffff8110fbd7>] might_fault+0x89/0xac [ 1773.640393] [<ffffffff81151cd2>] filldir64+0x7f/0xcd [ 1773.646391] [<ffffffff81160635>] dcache_readdir+0x64/0x1fc [ 1773.652970] [<ffffffff81151e50>] vfs_readdir+0x7b/0xb4 [ 1773.659193] [<ffffffff81152040>] sys_getdents64+0x7e/0xca [ 1773.665691] [<ffffffff8150b082>] system_call_fastpath+0x16/0x1b [ 1773.672720] [ 1773.672721] -> #0 (&sb->s_type->i_mutex_key#15){+.+.+.}: [ 1773.680531] [<ffffffff8108e81e>] __lock_acquire+0xa1a/0xcf7 [ 1773.687076] [<ffffffff8108eff1>] lock_acquire+0xf3/0x13e [ 1773.693368] [<ffffffff815025db>] __mutex_lock_common+0x5d/0x39a [ 1773.700256] [<ffffffff81502a27>] mutex_lock_nested+0x40/0x45 [ 1773.706887] [<ffffffff811e59d0>] hugetlbfs_file_mmap+0x82/0x10a [ 1773.713783] [<ffffffff81117bdc>] mmap_region+0x274/0x46b [ 1773.720069] [<ffffffff8111807f>] do_mmap_pgoff+0x2ac/0x306 [ 1773.726524] [<ffffffff811181f1>] sys_mmap_pgoff+0x118/0x16a [ 1773.733085] [<ffffffff81012880>] sys_mmap+0x22/0x24 [ 1773.738937] [<ffffffff8150b082>] system_call_fastpath+0x16/0x1b [ 1773.745835] [ 1773.745837] other info that might help us debug this: [ 1773.745838] [ 1773.754711] Possible unsafe locking scenario: [ 1773.754713] [ 1773.761227] CPU0 CPU1 [ 1773.766051] ---- ---- [ 1773.770868] lock(&mm->mmap_sem); [ 1773.774701] lock(&sb->s_type->i_mutex_key); [ 1773.782013] lock(&mm->mmap_sem); [ 1773.788360] lock(&sb->s_type->i_mutex_key); [ 1773.793156] [ 1773.793158] *** DEADLOCK *** [ 1773.793159] [ 1773.805829] 1 lock held by disktest/6751: [ 1773.810124] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff811181d1>] sys_mmap_pgoff+0xf8/0x16a [ 1773.819450] [ 1773.819452] stack backtrace: [ 1773.824414] Pid: 6751, comm: disktest Tainted: G W 3.1.0-0.rc3.git0.0.fc16.x86_64 #1 [ 1773.833445] Call Trace: [ 1773.836195] [<ffffffff814f9b74>] print_circular_bug+0x1f8/0x209 [ 1773.842503] [<ffffffff8108e81e>] __lock_acquire+0xa1a/0xcf7 [ 1773.848453] [<ffffffff81131091>] ? deactivate_slab+0x28f/0x2b5 [ 1773.854668] [<ffffffff811e59d0>] ? hugetlbfs_file_mmap+0x82/0x10a [ 1773.861138] [<ffffffff8108eff1>] lock_acquire+0xf3/0x13e [ 1773.866841] [<ffffffff811e59d0>] ? hugetlbfs_file_mmap+0x82/0x10a [ 1773.873312] [<ffffffff811e59d0>] ? hugetlbfs_file_mmap+0x82/0x10a [ 1773.879787] [<ffffffff815025db>] __mutex_lock_common+0x5d/0x39a [ 1773.886085] [<ffffffff811e59d0>] ? hugetlbfs_file_mmap+0x82/0x10a [ 1773.892566] [<ffffffff81117b11>] ? mmap_region+0x1a9/0x46b [ 1773.898427] [<ffffffff81502a27>] mutex_lock_nested+0x40/0x45 [ 1773.904461] [<ffffffff811e59d0>] hugetlbfs_file_mmap+0x82/0x10a [ 1773.910763] [<ffffffff81117bdc>] mmap_region+0x274/0x46b [ 1773.916447] [<ffffffff8111807f>] do_mmap_pgoff+0x2ac/0x306 [ 1773.922319] [<ffffffff811181f1>] sys_mmap_pgoff+0x118/0x16a [ 1773.928262] [<ffffffff8108f439>] ? trace_hardirqs_on_caller+0x121/0x158 [ 1773.935256] [<ffffffff812536fe>] ? trace_hardirqs_on_thunk+0x3a/0x3f [ 1773.941980] [<ffffffff81012880>] sys_mmap+0x22/0x24 [ 1773.947252] [<ffffffff8150b082>] system_call_fastpath+0x16/0x1b Version-Release number of selected component (if applicable): kernel-3.1.0-0.rc3.git0.0.fc16.x86_64 How reproducible: Always Steps to Reproduce: 1. Lots of disk I/O to and from blocks mmaped in /dev/hugepages. Additional info:
*** This bug has been marked as a duplicate of bug 730998 ***