* [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC
@ 2014-03-05 13:15 Veaceslav Falico
2014-03-06 2:02 ` Ding Tianhong
2014-03-06 13:22 ` Veaceslav Falico
0 siblings, 2 replies; 3+ messages in thread
From: Veaceslav Falico @ 2014-03-05 13:15 UTC (permalink / raw)
To: netdev
Cc: Veaceslav Falico, Thomas Glanzmann, Ding Tianhong, Jay Vosburgh,
Andy Gospodarek
Currently we're using GFP_KERNEL, however there are some path(s) where we
can hold some spinlocks, specifically bond->curr_slave_lock:
[ 4.722916] BUG: sleeping function called from invalid context at mm/slub.c:965
[ 4.724438] in_atomic(): 1, irqs_disabled(): 0, pid: 940, name: ifup-eth
[ 4.726034] 5 locks held by ifup-eth/940:
...snip...
[ 4.734646] #4: (&bond->curr_slave_lock){+...+.}, at: [<ffffffffa00badc6>] bond_enslave+0xda6/0xdd0 [bonding]
...snip...
[ 4.759081] [<ffffffffa00b6f11>] bond_change_active_slave+0x191/0x3b0 [bonding]
[ 4.760917] [<ffffffffa00b7227>] bond_select_active_slave+0xf7/0x1d0 [bonding]
[ 4.762751] [<ffffffffa00badce>] bond_enslave+0xdae/0xdd0 [bonding]
...snip...
As it's out of hot path and is a really rare event - change the gfp_t flags
to GFP_ATOMIC to avoid sleeping under spinlock.
CC: Thomas Glanzmann <thomas@glanzmann.de>
CC: Ding Tianhong <dingtianhong@huawei.com>
CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
---
drivers/net/bonding/bonding.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
index 4303628..4cd451d 100644
--- a/drivers/net/bonding/bonding.h
+++ b/drivers/net/bonding/bonding.h
@@ -292,7 +292,7 @@ static inline void bond_set_active_slave(struct slave *slave)
{
if (slave->backup) {
slave->backup = 0;
- rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
+ rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
}
}
@@ -300,7 +300,7 @@ static inline void bond_set_backup_slave(struct slave *slave)
{
if (!slave->backup) {
slave->backup = 1;
- rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
+ rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
}
}
--
1.8.4
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC
2014-03-05 13:15 [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC Veaceslav Falico
@ 2014-03-06 2:02 ` Ding Tianhong
2014-03-06 13:22 ` Veaceslav Falico
1 sibling, 0 replies; 3+ messages in thread
From: Ding Tianhong @ 2014-03-06 2:02 UTC (permalink / raw)
To: Veaceslav Falico, netdev; +Cc: Thomas Glanzmann, Jay Vosburgh, Andy Gospodarek
On 2014/3/5 21:15, Veaceslav Falico wrote:
> Currently we're using GFP_KERNEL, however there are some path(s) where we
> can hold some spinlocks, specifically bond->curr_slave_lock:
>
> [ 4.722916] BUG: sleeping function called from invalid context at mm/slub.c:965
> [ 4.724438] in_atomic(): 1, irqs_disabled(): 0, pid: 940, name: ifup-eth
> [ 4.726034] 5 locks held by ifup-eth/940:
> ...snip...
> [ 4.734646] #4: (&bond->curr_slave_lock){+...+.}, at: [<ffffffffa00badc6>] bond_enslave+0xda6/0xdd0 [bonding]
> ...snip...
> [ 4.759081] [<ffffffffa00b6f11>] bond_change_active_slave+0x191/0x3b0 [bonding]
> [ 4.760917] [<ffffffffa00b7227>] bond_select_active_slave+0xf7/0x1d0 [bonding]
> [ 4.762751] [<ffffffffa00badce>] bond_enslave+0xdae/0xdd0 [bonding]
> ...snip...
>
> As it's out of hot path and is a really rare event - change the gfp_t flags
> to GFP_ATOMIC to avoid sleeping under spinlock.
>
> CC: Thomas Glanzmann <thomas@glanzmann.de>
> CC: Ding Tianhong <dingtianhong@huawei.com>
> CC: Jay Vosburgh <fubar@us.ibm.com>
> CC: Andy Gospodarek <andy@greyhouse.net>
> Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
> ---
> drivers/net/bonding/bonding.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
> index 4303628..4cd451d 100644
> --- a/drivers/net/bonding/bonding.h
> +++ b/drivers/net/bonding/bonding.h
> @@ -292,7 +292,7 @@ static inline void bond_set_active_slave(struct slave *slave)
> {
> if (slave->backup) {
> slave->backup = 0;
> - rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
> + rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
> }
> }
>
> @@ -300,7 +300,7 @@ static inline void bond_set_backup_slave(struct slave *slave)
> {
> if (!slave->backup) {
> slave->backup = 1;
> - rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
> + rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
> }
> }
>
Acked-by: Ding Tianhong <dingtianhong@huawei.com>
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC
2014-03-05 13:15 [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC Veaceslav Falico
2014-03-06 2:02 ` Ding Tianhong
@ 2014-03-06 13:22 ` Veaceslav Falico
1 sibling, 0 replies; 3+ messages in thread
From: Veaceslav Falico @ 2014-03-06 13:22 UTC (permalink / raw)
To: netdev; +Cc: Thomas Glanzmann, Ding Tianhong, Jay Vosburgh, Andy Gospodarek
On Wed, Mar 05, 2014 at 02:15:53PM +0100, Veaceslav Falico wrote:
>Currently we're using GFP_KERNEL, however there are some path(s) where we
>can hold some spinlocks, specifically bond->curr_slave_lock:
>
>[ 4.722916] BUG: sleeping function called from invalid context at mm/slub.c:965
>[ 4.724438] in_atomic(): 1, irqs_disabled(): 0, pid: 940, name: ifup-eth
>[ 4.726034] 5 locks held by ifup-eth/940:
>...snip...
>[ 4.734646] #4: (&bond->curr_slave_lock){+...+.}, at: [<ffffffffa00badc6>] bond_enslave+0xda6/0xdd0 [bonding]
>...snip...
>[ 4.759081] [<ffffffffa00b6f11>] bond_change_active_slave+0x191/0x3b0 [bonding]
>[ 4.760917] [<ffffffffa00b7227>] bond_select_active_slave+0xf7/0x1d0 [bonding]
>[ 4.762751] [<ffffffffa00badce>] bond_enslave+0xdae/0xdd0 [bonding]
>...snip...
>
>As it's out of hot path and is a really rare event - change the gfp_t flags
>to GFP_ATOMIC to avoid sleeping under spinlock.
>
>CC: Thomas Glanzmann <thomas@glanzmann.de>
>CC: Ding Tianhong <dingtianhong@huawei.com>
>CC: Jay Vosburgh <fubar@us.ibm.com>
>CC: Andy Gospodarek <andy@greyhouse.net>
>Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Self-NAK as the net got synced and there is a conflicting fix from
5e5b0665 ("bonding: Fix RTNL: assertion failed at net/core/rtnetlink.c for
802.3ad mode"), which adds a lot of complexity, postpones the notifications
(and makes them kind of useless cause the state might have been changed
again), and still doesn't fix all the cases.
I'll re-send the patchset.
>---
> drivers/net/bonding/bonding.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
>diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
>index 4303628..4cd451d 100644
>--- a/drivers/net/bonding/bonding.h
>+++ b/drivers/net/bonding/bonding.h
>@@ -292,7 +292,7 @@ static inline void bond_set_active_slave(struct slave *slave)
> {
> if (slave->backup) {
> slave->backup = 0;
>- rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
>+ rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
> }
> }
>
>@@ -300,7 +300,7 @@ static inline void bond_set_backup_slave(struct slave *slave)
> {
> if (!slave->backup) {
> slave->backup = 1;
>- rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_KERNEL);
>+ rtmsg_ifinfo(RTM_NEWLINK, slave->dev, 0, GFP_ATOMIC);
> }
> }
>
>--
>1.8.4
>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2014-03-06 13:22 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-03-05 13:15 [PATCH net-next] bonding: make slave status notifications GFP_ATOMIC Veaceslav Falico
2014-03-06 2:02 ` Ding Tianhong
2014-03-06 13:22 ` Veaceslav Falico
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).