NET: phy_device, fix lock imbalance
Subsystems (3)
PATCH 1/2 NET: phy_device, fix lock imbalance
Don't forget to unlock a mutex in phy_scan_fixups on a fail path.
--- drivers/net/phy/phy_device.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-)
drivers/net/phy/phy_device.c +3 -1
@@ -134,8 +134,10 @@ int phy_scan_fixups(struct phy_device *phydev) err = fixup->run (phydev); -if (err <0 ) +if (err <0 ) { +mutex_unlock (&phy_fixup_lock);return err; + } } }mutex_unlock (&phy_fixup_lock);
Signature · 1 line
1.6.3.2
PATCH 2/2 NET: sungem, use spin_trylock_irqsave
Use spin_trylock_irqsave instead of open-coded
local_irq_save+spin_trylock.
Impact: cleanup
--- drivers/net/sungem.c | 4 +--- 1 files changed, 1 insertions(+), 3 deletions(-)
drivers/net/sungem.c +1 -3
@@ -1032,10 +1032,8 @@ static int gem_start_xmit(struct sk_buff *skb, struct net_device *dev) (csum_stuff_off << 21)); } -local_irq_save (flags); -if (!spin_trylock (&gp->tx_lock)) { +if (!spin_trylock_irqsave (&gp->tx_lock)) {/* Tell upper layer to requeue */ -local_irq_restore (flags);return NETDEV_TX_LOCKED ; }/* We raced with gem_do_stop() */
Signature · 1 line
1.6.3.2
David Miller
maintainerRe: [PATCH 2/2] NET: sungem, use spin_trylock_irqsave
Please DO NOT combine cleanups with real bug fixes into a patch series.
Bug fixes go to one tree, the rest go to another tree, and that is especially the case this late in the RC series.
A series of patches are a group of changes that are supposed to be related and are expected to be applied all to the same tree. That is not true of the patches you have posted here.
David Miller
maintainerRe: [PATCH 2/2] NET: sungem, use spin_trylock_irqsave
In reply to Jiri SlabyFrom: Jiri Slaby <[email protected]>
Date: Mon, 13 Jul 2009 23:23:40 +0200
drivers/net/sungem.c static int gem_start_xmit(struct sk_buff *skb, struct net_device *dev) · line 1035
Quoted 2 lines
drivers/net/sungem.c +1
-1-if (!spin_trylock (&gp->tx_lock)) { +if (!spin_trylock_irqsave (&gp->tx_lock)) {
I'm about to give you a royal flaming.
Any idea why?
(hint: type 'make' before you send patches!!!!)
Re: [PATCH 2/2] NET: sungem, use spin_trylock_irqsave
Quoted 4 lines · David Miller
drivers/net/sungem.c static int gem_start_xmit(struct sk_buff *skb, struct net_device *dev) · line 1035
Quoted 2 lines
drivers/net/sungem.c +1
-1-if (!spin_trylock (&gp->tx_lock)) { +if (!spin_trylock_irqsave (&gp->tx_lock)) {(hint: type 'make' before you send patches!!!!)
Grr, I did but my bash_history shows that I built wrong subtree. My apologies.
From: Jiri Slaby <[email protected]>
Date: Mon, 13 Jul 2009 23:23:39 +0200
Quoted 3 lines
Don't forget to unlock a mutex in
phy_scan_fixupson a fail path.Signed-off-by Jiri Slaby <[email protected]>
Applied.