From: Trond Myklebust <trond.myklebust@primarydata.com>
To: linux-nfs@vger.kernel.org
Subject: [PATCH 2/3] pNFS: Fix a deadlock when coalescing writes and returning the layout
Date: Tue, 2 May 2017 12:38:08 -0400 [thread overview]
Message-ID: <20170502163812.6211-5-trond.myklebust@primarydata.com> (raw)
In-Reply-To: <20170502163812.6211-4-trond.myklebust@primarydata.com>
Consider the following deadlock:
Process P1 Process P2 Process P3
========== ========== ==========
lock_page(page)
lseg = pnfs_update_layout(inode)
lo = NFS_I(inode)->layout
pnfs_error_mark_layout_for_return(lo)
lock_page(page)
lseg = pnfs_update_layout(inode)
In this scenario,
- P1 has declared the layout to be in error, but P2 holds a reference to
a layout segment on that inode, so the layoutreturn is deferred.
- P2 is waiting for a page lock held by P3.
- P3 is asking for a new layout segment, but is blocked waiting
for the layoutreturn.
The fix is to ensure that pnfs_error_mark_layout_for_return() does
not set the NFS_LAYOUT_RETURN flag, which blocks P3. Instead, we allow
the latter to call LAYOUTGET so that it can make progress and unblock
P2.
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
---
fs/nfs/pnfs.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index cea1e838efae..adc6ec28d4b5 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -2063,8 +2063,6 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
return;
}
pnfs_set_plh_return_info(lo, range.iomode, 0);
- /* Block LAYOUTGET */
- set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags);
/*
* mark all matching lsegs so that we are sure to have no live
* segments at hand when sending layoutreturn. See pnfs_put_lseg()
--
2.9.3
next prev parent reply other threads:[~2017-05-02 16:38 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-02 16:38 [PATCH 0/3] Fix up a couple of issues around layout handling Trond Myklebust
2017-05-02 16:38 ` [RFC PATCH 1/5] SUNRPC: Allow creation of RPC clients with multiple connections Trond Myklebust
2017-05-02 16:38 ` [PATCH 1/3] pNFS: Don't clear the layout return info if there are segments to return Trond Myklebust
2017-05-02 16:38 ` [RFC PATCH 2/5] NFS: Add a mount option to specify number of TCP connections to use Trond Myklebust
2017-05-02 16:38 ` Trond Myklebust [this message]
2017-05-02 16:38 ` [RFC PATCH 3/5] NFSv4: Allow multiple connections to NFSv4.x (x>0) servers Trond Myklebust
2017-05-02 16:38 ` [PATCH 3/3] pNFS: Fix a typo in pnfs_generic_alloc_ds_commits Trond Myklebust
2017-05-02 16:38 ` [RFC PATCH 4/5] pNFS: Allow multiple connections to the DS Trond Myklebust
2017-05-02 16:38 ` [RFC PATCH 5/5] NFS: Display the "nconnect" mount option if it is set Trond Myklebust
2017-05-02 16:41 [PATCH 0/3] Fix up a couple of issues around layout handling Trond Myklebust
2017-05-02 16:41 ` [PATCH 1/3] pNFS: Don't clear the layout return info if there are segments to return Trond Myklebust
2017-05-02 16:41 ` [PATCH 2/3] pNFS: Fix a deadlock when coalescing writes and returning the layout Trond Myklebust
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=20170502163812.6211-5-trond.myklebust@primarydata.com \
--to=trond.myklebust@primarydata.com \
--cc=linux-nfs@vger.kernel.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).