From: Pavel Emelyanov <xemul@openvz.org>
To: Jay Vosburgh <fubar@us.ibm.com>, David Miller <davem@davemloft.net>
Cc: bonding-devel@lists.sourceforge.net,
Linux Netdev List <netdev@vger.kernel.org>
Subject: [PATCH 8/8][BONDING]: Deadlock between bonding_store_bonds and bond_destroy_sysfs.
Date: Tue, 29 Apr 2008 19:11:48 +0400 [thread overview]
Message-ID: <48173AB4.4080009@openvz.org> (raw)
In-Reply-To: <481735F4.30909@openvz.org>
The sysfs layer has an internal protection, that ensures, that
all the process sitting inside ->sore/->show callback exits
before the appropriate entry is unregistered (the calltraces
are rather big, but I can provide them if required).
On the other hand, bonding takes rtnl_lock in
a) the bonding_store_bonds, i.e. in ->store callback,
b) module exit before calling the sysfs unregister routines.
Thus, the classical AB-BA deadlock may occur. To reproduce run
# while :; do modprobe bonding; rmmod bonding; done
and
# while :; do echo '+bond%d' > /sys/class/net/bonding_masters ; done
in parallel.
The fix is to move the bond_destroy_sysfs out of the rtnl_lock,
but _before_ bond_free_all to make sure no bonding devices exist
after module unload.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
---
drivers/net/bonding/bond_main.c | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 4d852ab..7e7b0b8 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -4984,9 +4984,10 @@ err:
destroy_workqueue(bond->wq);
}
+ bond_destroy_sysfs();
+
rtnl_lock();
bond_free_all();
- bond_destroy_sysfs();
rtnl_unlock();
out:
return res;
@@ -4998,9 +4999,10 @@ static void __exit bonding_exit(void)
unregister_netdevice_notifier(&bond_netdev_notifier);
unregister_inetaddr_notifier(&bond_inetaddr_notifier);
+ bond_destroy_sysfs();
+
rtnl_lock();
bond_free_all();
- bond_destroy_sysfs();
rtnl_unlock();
}
--
1.5.3.4
next prev parent reply other threads:[~2008-04-29 15:49 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-04-29 14:51 [PATCH 0/8][BONDING]: Some BUGs and deadlocks fixes and a bit of concomitant cleanups Pavel Emelyanov
2008-04-29 14:54 ` [PATCH 1/8][BONDING]: Do not call free_netdev for already registered device Pavel Emelyanov
2008-04-29 14:56 ` [PATCH 2/8][BONDING]: Don't create bondings with % in their names Pavel Emelyanov
2008-04-29 14:58 ` [PATCH 3/8][BONDING]: Merge two calls to dev_alloc_name Pavel Emelyanov
2008-04-29 15:01 ` [PATCH 4/8][BONDING]: Remove redundant argument from bond_create Pavel Emelyanov
2008-04-29 15:02 ` [PATCH 5/8][BONDING]: Lost semaphores unlock in one of bonding_store_bonds error paths Pavel Emelyanov
2008-04-29 15:05 ` [PATCH 6/8][BONDING]: Relax unneeded _safe lists iterations Pavel Emelyanov
2008-04-29 15:06 ` [PATCH 7/8][BONDING]: Remove unneeded list_empty checks Pavel Emelyanov
2008-04-29 15:11 ` Pavel Emelyanov [this message]
2008-05-03 0:06 ` [PATCH 0/8][BONDING]: Some BUGs and deadlocks fixes and a bit of concomitant cleanups David Miller
2008-05-03 0:25 ` Jay Vosburgh
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=48173AB4.4080009@openvz.org \
--to=xemul@openvz.org \
--cc=bonding-devel@lists.sourceforge.net \
--cc=davem@davemloft.net \
--cc=fubar@us.ibm.com \
--cc=netdev@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).