* [PATCH] irda: ali-ircc: Fix deadlock in ali_ircc_sir_change_speed()
@ 2015-09-11 21:34 Alexey Khoroshilov
2015-09-11 23:18 ` David Miller
0 siblings, 1 reply; 2+ messages in thread
From: Alexey Khoroshilov @ 2015-09-11 21:34 UTC (permalink / raw)
To: Samuel Ortiz, David S . Miller
Cc: Alexey Khoroshilov, netdev, linux-kernel, ldv-project
ali_ircc_sir_change_speed() is always called with self->lock held,
so acquiring the lock inside it leads to unavoidable deadlock.
Call graph:
ali_ircc_sir_change_speed() is called from ali_ircc_change_speed()
ali_ircc_fir_hard_xmit() under spin_lock_irqsave(&self->lock, flags);
ali_ircc_sir_hard_xmit() under spin_lock_irqsave(&self->lock, flags);
ali_ircc_net_ioctl() under spin_lock_irqsave(&self->lock, flags);
ali_ircc_dma_xmit_complete()
ali_ircc_fir_interrupt()
ali_ircc_interrupt() under spin_lock(&self->lock);
ali_ircc_sir_write_wakeup()
ali_ircc_sir_interrupt()
ali_ircc_interrupt() under spin_lock(&self->lock);
The patch removes spin_lock/unlock from ali_ircc_sir_change_speed().
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
---
drivers/net/irda/ali-ircc.c | 6 ------
1 file changed, 6 deletions(-)
diff --git a/drivers/net/irda/ali-ircc.c b/drivers/net/irda/ali-ircc.c
index 58ae11a14bb6..64bb44d5d867 100644
--- a/drivers/net/irda/ali-ircc.c
+++ b/drivers/net/irda/ali-ircc.c
@@ -1031,7 +1031,6 @@ static void ali_ircc_fir_change_speed(struct ali_ircc_cb *priv, __u32 baud)
static void ali_ircc_sir_change_speed(struct ali_ircc_cb *priv, __u32 speed)
{
struct ali_ircc_cb *self = priv;
- unsigned long flags;
int iobase;
int fcr; /* FIFO control reg */
int lcr; /* Line control reg */
@@ -1061,8 +1060,6 @@ static void ali_ircc_sir_change_speed(struct ali_ircc_cb *priv, __u32 speed)
/* Update accounting for new speed */
self->io.speed = speed;
- spin_lock_irqsave(&self->lock, flags);
-
divisor = 115200/speed;
fcr = UART_FCR_ENABLE_FIFO;
@@ -1089,9 +1086,6 @@ static void ali_ircc_sir_change_speed(struct ali_ircc_cb *priv, __u32 speed)
/* without this, the connection will be broken after come back from FIR speed,
but with this, the SIR connection is harder to established */
outb((UART_MCR_DTR | UART_MCR_RTS | UART_MCR_OUT2), iobase+UART_MCR);
-
- spin_unlock_irqrestore(&self->lock, flags);
-
}
static void ali_ircc_change_dongle_speed(struct ali_ircc_cb *priv, int speed)
--
1.9.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] irda: ali-ircc: Fix deadlock in ali_ircc_sir_change_speed()
2015-09-11 21:34 [PATCH] irda: ali-ircc: Fix deadlock in ali_ircc_sir_change_speed() Alexey Khoroshilov
@ 2015-09-11 23:18 ` David Miller
0 siblings, 0 replies; 2+ messages in thread
From: David Miller @ 2015-09-11 23:18 UTC (permalink / raw)
To: khoroshilov; +Cc: samuel, netdev, linux-kernel, ldv-project
From: Alexey Khoroshilov <khoroshilov@ispras.ru>
Date: Sat, 12 Sep 2015 00:34:48 +0300
> ali_ircc_sir_change_speed() is always called with self->lock held,
> so acquiring the lock inside it leads to unavoidable deadlock.
>
> Call graph:
> ali_ircc_sir_change_speed() is called from ali_ircc_change_speed()
> ali_ircc_fir_hard_xmit() under spin_lock_irqsave(&self->lock, flags);
> ali_ircc_sir_hard_xmit() under spin_lock_irqsave(&self->lock, flags);
> ali_ircc_net_ioctl() under spin_lock_irqsave(&self->lock, flags);
> ali_ircc_dma_xmit_complete()
> ali_ircc_fir_interrupt()
> ali_ircc_interrupt() under spin_lock(&self->lock);
> ali_ircc_sir_write_wakeup()
> ali_ircc_sir_interrupt()
> ali_ircc_interrupt() under spin_lock(&self->lock);
>
> The patch removes spin_lock/unlock from ali_ircc_sir_change_speed().
>
> Found by Linux Driver Verification project (linuxtesting.org).
>
> Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Looks good, applied, thanks.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2015-09-11 23:18 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-09-11 21:34 [PATCH] irda: ali-ircc: Fix deadlock in ali_ircc_sir_change_speed() Alexey Khoroshilov
2015-09-11 23:18 ` David Miller
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).