rfcomm deadlock fix
PATCH 1/1 rfcomm deadlock fix
There's logic in __rfcomm_dlc_close:
rfcomm_dlc_lock(d); d->state = BT_CLOSED; d->state_changed(d, err); rfcomm_dlc_unlock(d);
In rfcomm_dev_state_change, it's possible that rfcomm_dev_put try to take the dlc lock, then we will deadlock.
Here fixed it by unlock dlc before rfcomm_dev_get in rfcomm_dev_state_change.
why not unlock just before rfcomm_dev_put? it's because there's another problem. rfcomm_dev_get/rfcomm_dev_del will take rfcomm_dev_lock, but in rfcomm_dev_add the lock order is : rfcomm_dev_lock --> dlc lock
so I unlock dlc before the taken of rfcomm_dev_lock.
Actually it's a regression caused by commit1905f6c736cb618e07eca0c96e60e3c024023428, the dlc state_change could be two
callbacks : rfcomm_sk_state_change and rfcomm_dev_state_change. I missed the rfcomm_sk_state_change that time.
Thanks Arjan van de Ven <[email protected]> for the effort in commit 4c8411f8c115def968820a4df6658ccfd55d7f1a
but he missed the rfcomm_dev_state_change lock issue.
---net/bluetooth/rfcomm/tty.c | 9 +++++++++
1 files changed, 9 insertions(+)
net/bluetooth/rfcomm/tty.c +9 -0
@@ -566,11 +566,20 @@ static void rfcomm_dev_state_change(struif (dlc->state ==BT_CLOSED ) {if (!dev->tty) {if (test_bit (RFCOMM_RELEASE_ONHUP , &dev->flags)) { +/* Drop DLC lock here to avoid deadlock +* 1. rfcomm_dev_get will take rfcomm_dev_lock +* but in rfcomm_dev_add there's lock order: +* rfcomm_dev_lock -> dlc lock +* 2. rfcomm_dev_put will deaklock if it's +* the last reference +*/ +rfcomm_dlc_unlock (dlc);if (rfcomm_dev_get (dev->id) ==NULL )return ;rfcomm_dev_del (dev);rfcomm_dev_put (dev); +rfcomm_dlc_lock (dlc); } }else tty_hangup (dev->tty);
Sorry, I sent a wrong version, I will sent again after a while.
Quoted 56 lines · Dave Young
There's logic in
__rfcomm_dlc_close:rfcomm_dlc_lock(d); d->state = BT_CLOSED; d->state_changed(d, err); rfcomm_dlc_unlock(d);In
rfcomm_dev_state_change, it's possible thatrfcomm_dev_puttry to take the dlc lock, then we will deadlock.Here fixed it by unlock dlc before
rfcomm_dev_getinrfcomm_dev_state_change.why not unlock just before
rfcomm_dev_put? it's because there's another problem. rfcomm_dev_get/rfcomm_dev_del will takerfcomm_dev_lock, but inrfcomm_dev_addthe lock order is :rfcomm_dev_lock--> dlc lockso I unlock dlc before the taken of
rfcomm_dev_lock.Actually it's a regression caused by commit
1905f6c736cb618e07eca0c96e60e3c024023428, the dlcstate_changecould be two
callbacks :rfcomm_sk_state_changeandrfcomm_dev_state_change. I missed therfcomm_sk_state_changethat time.Thanks Arjan van de Ven <[email protected]> for the effort in commit
4c8411f8c115def968820a4df6658ccfd55d7f1a
but he missed therfcomm_dev_state_changelock issue.Signed-off-by Dave Young <[email protected]>---
net/bluetooth/rfcomm/tty.c| 9 +++++++++1 files changed, 9 insertions(+)diff -upr
linux/net/bluetooth/rfcomm/tty.clinux.new/net/bluetooth/rfcomm/tty.cnet/bluetooth/rfcomm/tty.c 2008-05-30 16:02:38.000000000 +0800 renamed from net/bluetooth/rfcomm/tty.c 2008-05-30 15:46:33.000000000 +0800 +9
-0@@ -566,11 +566,20 @@ static void rfcomm_dev_state_change(stru if (dlc->state == BT_CLOSED) { if (!dev->tty) { if (test_bit(RFCOMM_RELEASE_ONHUP, &dev->flags)) { + /* Drop DLC lock here to avoid deadlock + * 1. rfcomm_dev_get will take rfcomm_dev_lock + * but in rfcomm_dev_add there's lock order: + * rfcomm_dev_lock -> dlc lock + * 2. rfcomm_dev_put will deaklock if it's + * the last reference + */ + rfcomm_dlc_unlock(dlc); if (rfcomm_dev_get(dev->id) == NULL) return; rfcomm_dev_del(dev); rfcomm_dev_put(dev); + rfcomm_dlc_lock(dlc); } } else tty_hangup(dev->tty);
Signature · 3 lines
--- Regards dave
Marcel Holtmann
maintainerHi Dave,
Quoted 1 line
Sorry, I sent a wrong version, I will sent again after a while.
also do s/deaklock/deadlock/ :)
Regards
Marcel
Quoted 5 lines · Marcel Holtmann
Hi Dave,
Quoted 1 line
Sorry, I sent a wrong version, I will sent again after a while.
also do s/deaklock/deadlock/ :)
My typo, will send again.
Regards
dave