[Quantal][PATCH 21/29] netpoll: revert 6bdb7fe3104 and fix be_poll() instead
Jesse Sung
jesse.sung at canonical.com
Thu Dec 13 06:49:54 UTC 2012
From: Amerigo Wang <amwang at redhat.com>
BugLink: https://launchpad.net/bugs/1083088
Against -net.
In the patch "netpoll: re-enable irq in poll_napi()", I tried to
fix the following warning:
[100718.051041] ------------[ cut here ]------------
[100718.051048] WARNING: at kernel/softirq.c:159 local_bh_enable_ip+0x7d/0xb0()
(Not tainted)
[100718.051049] Hardware name: ProLiant BL460c G7
...
[100718.051068] Call Trace:
[100718.051073] [<ffffffff8106b747>] ? warn_slowpath_common+0x87/0xc0
[100718.051075] [<ffffffff8106b79a>] ? warn_slowpath_null+0x1a/0x20
[100718.051077] [<ffffffff810747ed>] ? local_bh_enable_ip+0x7d/0xb0
[100718.051080] [<ffffffff8150041b>] ? _spin_unlock_bh+0x1b/0x20
[100718.051085] [<ffffffffa00ee974>] ? be_process_mcc+0x74/0x230 [be2net]
[100718.051088] [<ffffffffa00ea68c>] ? be_poll_tx_mcc+0x16c/0x290 [be2net]
[100718.051090] [<ffffffff8144fe76>] ? netpoll_poll_dev+0xd6/0x490
[100718.051095] [<ffffffffa01d24a5>] ? bond_poll_controller+0x75/0x80 [bonding]
[100718.051097] [<ffffffff8144fde5>] ? netpoll_poll_dev+0x45/0x490
[100718.051100] [<ffffffff81161b19>] ? ksize+0x19/0x80
[100718.051102] [<ffffffff81450437>] ? netpoll_send_skb_on_dev+0x157/0x240
by reenabling IRQ before calling ->poll, but it seems more
problems are introduced after that patch:
http://ozlabs.org/~akpm/stuff/IMG_20120824_122054.jpg
http://marc.info/?l=linux-netdev&m=134563282530588&w=2
So it is safe to fix be2net driver code directly.
This patch reverts the offending commit and fixes be_poll() by
avoid disabling BH there, this is okay because be_poll()
can be called either by poll_napi() which already disables
IRQ, or by net_rx_action() which already disables BH.
Reported-by: Andrew Morton <akpm at linux-foundation.org>
Reported-by: Sylvain Munaut <s.munaut at whatever-company.com>
Cc: Sylvain Munaut <s.munaut at whatever-company.com>
Cc: Andrew Morton <akpm at linux-foundation.org>
Cc: David Miller <davem at davemloft.net>
Cc: Sathya Perla <sathya.perla at emulex.com>
Cc: Subbu Seetharaman <subbu.seetharaman at emulex.com>
Cc: Ajit Khaparde <ajit.khaparde at emulex.com>
Signed-off-by: Cong Wang <amwang at redhat.com>
Tested-by: Sylvain Munaut <s.munaut at whatever-company.com>
Signed-off-by: David S. Miller <davem at davemloft.net>
(cherry picked from commit 072a9c48600409d72aeb0d5b29fbb75861a06631)
Conflicts:
net/core/netpoll.c
Signed-off-by: Wen-chien Jesse Sung <jesse.sung at canonical.com>
---
drivers/net/ethernet/emulex/benet/be_cmds.c | 6 ++++--
drivers/net/ethernet/emulex/benet/be_main.c | 2 ++
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index b6748da..6701998 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c
@@ -259,7 +259,7 @@ int be_process_mcc(struct be_adapter *adapter)
int num = 0, status = 0;
struct be_mcc_obj *mcc_obj = &adapter->mcc_obj;
- spin_lock_bh(&adapter->mcc_cq_lock);
+ spin_lock(&adapter->mcc_cq_lock);
while ((compl = be_mcc_compl_get(adapter))) {
if (compl->flags & CQE_FLAGS_ASYNC_MASK) {
/* Interpret flags as an async trailer */
@@ -280,7 +280,7 @@ int be_process_mcc(struct be_adapter *adapter)
if (num)
be_cq_notify(adapter, mcc_obj->cq.id, mcc_obj->rearm_cq, num);
- spin_unlock_bh(&adapter->mcc_cq_lock);
+ spin_unlock(&adapter->mcc_cq_lock);
return status;
}
@@ -295,7 +295,9 @@ static int be_mcc_wait_compl(struct be_adapter *adapter)
if (be_error(adapter))
return -EIO;
+ local_bh_disable();
status = be_process_mcc(adapter);
+ local_bh_enable();
if (atomic_read(&mcc_obj->q.used) == 0)
break;
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index f714fc1..50691ef 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -3761,7 +3761,9 @@ static void be_worker(struct work_struct *work)
/* when interrupts are not yet enabled, just reap any pending
* mcc completions */
if (!netif_running(adapter->netdev)) {
+ local_bh_disable();
be_process_mcc(adapter);
+ local_bh_enable();
goto reschedule;
}
--
1.7.10.4
More information about the kernel-team
mailing list