From: Omar Sandoval <osandov@osandov.com>
To: Chris Murphy <lists@colorremedies.com>, Christoph Hellwig <hch@lst.de>
Cc: Btrfs BTRFS <linux-btrfs@vger.kernel.org>,
David Sterba <dsterba@suse.cz>
Subject: Re: status of swapfiles on Btrfs
Date: Thu, 8 Jun 2017 17:06:49 -0700 [thread overview]
Message-ID: <20170609000649.GA32051@vader> (raw)
In-Reply-To: <CAJCQCtTJYsYXChwochYocJN-uo-OG-BFfogMwkhQuN9X__1z6A@mail.gmail.com>
On Thu, Jun 08, 2017 at 03:35:10PM -0600, Chris Murphy wrote:
> What's the status?
I rebased the patches on v4.9 back in November and ran into a circular
locking issue between mmap_sem and i_rwsem. I never figured out how to
resolve that. Christoph was the last one that I talked to about this,
maybe he has some ideas. Latest code rebased on v4.12-rc4 is here:
https://github.com/osandov/linux/tree/btrfs-swap.
[ 991.245632] ======================================================
[ 991.246493] WARNING: possible circular locking dependency detected
[ 991.246590] 4.12.0-rc4-00006-g0e2e3e2ba974 #3 Not tainted
[ 991.246590] ------------------------------------------------------
[ 991.246590] swapme/626 is trying to acquire lock:
[ 991.246590] (&sb->s_type->i_mutex_key#16){++++++}, at: [<ffffffffa02804ee>] nfs_start_io_direct+0x1e/0x70 [nfs]
[ 991.246590]
[ 991.246590] but task is already holding lock:
[ 991.246590] (&mm->mmap_sem){++++++}, at: [<ffffffff8105bb0a>] __do_page_fault+0x17a/0x550
[ 991.246590]
[ 991.246590] which lock already depends on the new lock.
[ 991.246590]
[ 991.246590]
[ 991.246590] the existing dependency chain (in reverse order) is:
[ 991.246590]
[ 991.246590] -> #1 (&mm->mmap_sem){++++++}:
[ 991.246590] lock_acquire+0xa5/0x250
[ 991.246590] __might_fault+0x68/0x90
[ 991.246590] copy_page_to_iter+0xc4/0x310
[ 991.246590] generic_file_read_iter+0x325/0x7d0
[ 991.246590] nfs_file_read+0x7c/0xa0 [nfs]
[ 991.246590] __vfs_read+0xe1/0x130
[ 991.246590] vfs_read+0xa8/0x150
[ 991.246590] SyS_read+0x58/0xd0
[ 991.246590] entry_SYSCALL_64_fastpath+0x1f/0xbe
[ 991.246590]
[ 991.246590] -> #0 (&sb->s_type->i_mutex_key#16){++++++}:
[ 991.246590] __lock_acquire+0x15e1/0x1940
[ 991.246590] lock_acquire+0xa5/0x250
[ 991.246590] down_read+0x3e/0x70
[ 991.246590] nfs_start_io_direct+0x1e/0x70 [nfs]
[ 991.246590] nfs_file_direct_write+0x1b6/0x290 [nfs]
[ 991.246590] nfs_file_write+0x169/0x1f0 [nfs]
[ 991.246590] __swap_writepage+0x121/0x2f0
[ 991.246590] swap_writepage+0x34/0x90
[ 991.246590] pageout.isra.18+0xf8/0x3c0
[ 991.246590] shrink_page_list+0x779/0xac0
[ 991.246590] shrink_inactive_list+0x200/0x580
[ 991.246590] shrink_node_memcg+0x367/0x750
[ 991.246590] shrink_node+0xf7/0x2f0
[ 991.246590] do_try_to_free_pages+0xd7/0x350
[ 991.246590] try_to_free_mem_cgroup_pages+0x111/0x390
[ 991.246590] try_charge+0x14b/0xa20
[ 991.246590] mem_cgroup_try_charge+0x87/0x480
[ 991.246590] __handle_mm_fault+0xbc0/0x11b0
[ 991.246590] handle_mm_fault+0x174/0x340
[ 991.246590] __do_page_fault+0x290/0x550
[ 991.246590] trace_do_page_fault+0x9a/0x260
[ 991.246590] do_async_page_fault+0x4f/0x70
[ 991.246590] async_page_fault+0x28/0x30
[ 991.246590]
[ 991.246590] other info that might help us debug this:
[ 991.246590]
[ 991.246590] Possible unsafe locking scenario:
[ 991.246590]
[ 991.246590] CPU0 CPU1
[ 991.246590] ---- ----
[ 991.246590] lock(&mm->mmap_sem);
[ 991.246590] lock(&sb->s_type->i_mutex_key#16);
[ 991.246590] lock(&mm->mmap_sem);
[ 991.246590] lock(&sb->s_type->i_mutex_key#16);
[ 991.246590]
[ 991.246590] *** DEADLOCK ***
[ 991.246590]
[ 991.246590] 1 lock held by swapme/626:
[ 991.246590] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8105bb0a>] __do_page_fault+0x17a/0x550
[ 991.246590]
[ 991.246590] stack backtrace:
[ 991.246590] CPU: 2 PID: 626 Comm: swapme Not tainted 4.12.0-rc4-00006-g0e2e3e2ba974 #3
[ 991.246590] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-20170228_101828-anatol 04/01/2014
[ 991.246590] Call Trace:
[ 991.246590] dump_stack+0x8e/0xcd
[ 991.246590] print_circular_bug+0x1f8/0x2e0
[ 991.246590] __lock_acquire+0x15e1/0x1940
[ 991.246590] lock_acquire+0xa5/0x250
[ 991.246590] ? lock_acquire+0xa5/0x250
[ 991.246590] ? nfs_start_io_direct+0x1e/0x70 [nfs]
[ 991.246590] down_read+0x3e/0x70
[ 991.246590] ? nfs_start_io_direct+0x1e/0x70 [nfs]
[ 991.246590] nfs_start_io_direct+0x1e/0x70 [nfs]
[ 991.246590] nfs_file_direct_write+0x1b6/0x290 [nfs]
[ 991.246590] nfs_file_write+0x169/0x1f0 [nfs]
[ 991.246590] ? SyS_madvise+0x870/0x870
[ 991.246590] __swap_writepage+0x121/0x2f0
[ 991.246590] swap_writepage+0x34/0x90
[ 991.246590] pageout.isra.18+0xf8/0x3c0
[ 991.246590] shrink_page_list+0x779/0xac0
[ 991.246590] shrink_inactive_list+0x200/0x580
[ 991.246590] ? mark_lock+0x5d0/0x670
[ 991.246590] shrink_node_memcg+0x367/0x750
[ 991.246590] ? mem_cgroup_iter+0x1c1/0x760
[ 991.246590] shrink_node+0xf7/0x2f0
[ 991.246590] ? shrink_node+0xf7/0x2f0
[ 991.246590] do_try_to_free_pages+0xd7/0x350
[ 991.246590] try_to_free_mem_cgroup_pages+0x111/0x390
[ 991.246590] try_charge+0x14b/0xa20
[ 991.246590] ? get_mem_cgroup_from_mm+0xfc/0x2d0
[ 991.246590] mem_cgroup_try_charge+0x87/0x480
[ 991.246590] __handle_mm_fault+0xbc0/0x11b0
[ 991.246590] handle_mm_fault+0x174/0x340
[ 991.246590] __do_page_fault+0x290/0x550
[ 991.246590] trace_do_page_fault+0x9a/0x260
[ 991.246590] do_async_page_fault+0x4f/0x70
[ 991.246590] async_page_fault+0x28/0x30
[ 991.246590] RIP: 0033:0x400b72
[ 991.246590] RSP: 002b:00007fffa3a2ed60 EFLAGS: 00010206
[ 991.246590] RAX: 00007f78598cf000 RBX: 00007f78598cf000 RCX: 00007f787969fa8a
[ 991.246590] RDX: 0000000000000000 RSI: 0000000000001000 RDI: 0000000000000000
[ 991.246590] RBP: 0000000000001000 R08: ffffffffffffffff R09: 0000000000000000
[ 991.246590] R10: 0000000000000022 R11: 0000000000000246 R12: 0000000000040000
[ 991.246590] R13: 00007f78598d0000 R14: 00000000000001ff R15: 000000000001fee5
prev parent reply other threads:[~2017-06-09 0:06 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-08 21:35 status of swapfiles on Btrfs Chris Murphy
2017-06-09 0:06 ` Omar Sandoval [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170609000649.GA32051@vader \
--to=osandov@osandov.com \
--cc=dsterba@suse.cz \
--cc=hch@lst.de \
--cc=linux-btrfs@vger.kernel.org \
--cc=lists@colorremedies.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.