lkml.org 
[lkml]   [2012]   [Mar]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCH] hugetlbfs: lockdep annotate root inode properly
    On 2012-03-08 13:02:56, Andrew Morton wrote:
    > On Thu, 8 Mar 2012 14:45:16 +0530
    > "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com> wrote:
    >
    > > From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
    > >
    > > This fix the below lockdep warning
    >
    > OK, what's going on here.
    >
    > > ======================================================
    > > [ INFO: possible circular locking dependency detected ]
    > > 3.3.0-rc4+ #190 Not tainted
    > > -------------------------------------------------------
    > > shared/1568 is trying to acquire lock:
    > > (&sb->s_type->i_mutex_key#12){+.+.+.}, at: [<ffffffff811efa0f>] hugetlbfs_file_mmap+0x7d/0x108
    > >
    > > but task is already holding lock:
    > > (&mm->mmap_sem){++++++}, at: [<ffffffff810f5589>] sys_mmap_pgoff+0xd4/0x12f
    > >
    > > which lock already depends on the new lock.
    > >
    > >
    > > the existing dependency chain (in reverse order) is:
    > >
    > > -> #1 (&mm->mmap_sem){++++++}:
    > > [<ffffffff8109fb8f>] lock_acquire+0xd5/0xfa
    > > [<ffffffff810ee439>] might_fault+0x6d/0x90
    > > [<ffffffff8111bc12>] filldir+0x6a/0xc2
    > > [<ffffffff81129942>] dcache_readdir+0x5c/0x222
    > > [<ffffffff8111be58>] vfs_readdir+0x76/0xac
    > > [<ffffffff8111bf6a>] sys_getdents+0x79/0xc9
    > > [<ffffffff816940a2>] system_call_fastpath+0x16/0x1b
    > >
    > > -> #0 (&sb->s_type->i_mutex_key#12){+.+.+.}:
    > > [<ffffffff8109f40a>] __lock_acquire+0xa6c/0xd60
    > > [<ffffffff8109fb8f>] lock_acquire+0xd5/0xfa
    > > [<ffffffff816916be>] __mutex_lock_common+0x48/0x350
    > > [<ffffffff81691a85>] mutex_lock_nested+0x2a/0x31
    > > [<ffffffff811efa0f>] hugetlbfs_file_mmap+0x7d/0x108
    > > [<ffffffff810f4fd0>] mmap_region+0x26f/0x466
    > > [<ffffffff810f545b>] do_mmap_pgoff+0x294/0x2ee
    > > [<ffffffff810f55a9>] sys_mmap_pgoff+0xf4/0x12f
    > > [<ffffffff8103d1f2>] sys_mmap+0x1d/0x1f
    > > [<ffffffff816940a2>] system_call_fastpath+0x16/0x1b
    > >
    > > other info that might help us debug this:
    > >
    > > Possible unsafe locking scenario:
    > >
    > > CPU0 CPU1
    > > ---- ----
    > > lock(&mm->mmap_sem);
    > > lock(&sb->s_type->i_mutex_key#12);
    > > lock(&mm->mmap_sem);
    > > lock(&sb->s_type->i_mutex_key#12);
    > >
    > > *** DEADLOCK ***
    > >
    > > 1 lock held by shared/1568:
    > > #0: (&mm->mmap_sem){++++++}, at: [<ffffffff810f5589>] sys_mmap_pgoff+0xd4/0x12f
    > >
    > > stack backtrace:
    > > Pid: 1568, comm: shared Not tainted 3.3.0-rc4+ #190
    > > Call Trace:
    > > [<ffffffff81688bf9>] print_circular_bug+0x1f8/0x209
    > > [<ffffffff8109f40a>] __lock_acquire+0xa6c/0xd60
    > > [<ffffffff8110e7b6>] ? files_lglock_local_lock_cpu+0x61/0x61
    > > [<ffffffff811efa0f>] ? hugetlbfs_file_mmap+0x7d/0x108
    > > [<ffffffff8109fb8f>] lock_acquire+0xd5/0xfa
    > > [<ffffffff811efa0f>] ? hugetlbfs_file_mmap+0x7d/0x108
    > >
    >
    > Why have these lockdep warnings started coming out now - was the VFS
    > changed to newly take i_mutex somewhere in the directory handling?

    I'm not sure that they're new warnings. My patch (linked to below) may
    have just gave folks a false hope that their nagging lockdep problems
    are over.

    >
    >
    > Sigh. Was lockdep_annotate_inode_mutex_key() sufficiently
    > self-explanatory to justify leaving it undocumented?
    >
    > <goes off and reads e096d0c7e2e>
    >
    > OK, the patch looks correct given the explanation in e096d0c7e2e, but
    > I'd like to understand why it becomes necessary only now.
    >
    > > NOTE: This patch also require
    > > http://thread.gmane.org/gmane.linux.file-systems/58795/focus=59565
    > > to remove the lockdep warning
    >
    > And that patch has been basically ignored.

    Al commented on it here:

    https://lkml.org/lkml/2012/2/16/518

    He said that while my patch is correct, taking i_mutex inside mmap_sem
    is still wrong.

    Tyler

    >
    > Sigh. I guess I'll grab both patches, but I'm not confident in doing
    > so without an overall explanation of what is happening here.
    >
    >
    [unhandled content-type:application/pgp-signature]
    \
     
     \ /
      Last update: 2012-03-08 22:21    [W:0.029 / U:1.876 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site