Use the kthread_ API instead of opencoding lots of hairy code for kernel
thread creation and teardown.
Also use wake_up_process instead of an additional per-socket waitqueue.
Signed-off-by: Christoph Hellwig <[email protected]>
Index: linux-2.6/drivers/pcmcia/cs.c
===================================================================
--- linux-2.6.orig/drivers/pcmcia/cs.c 2006-01-06 10:44:16.000000000 +0100
+++ linux-2.6/drivers/pcmcia/cs.c 2006-02-14 18:34:14.000000000 +0100
@@ -29,6 +29,7 @@
#include <linux/pm.h>
#include <linux/pci.h>
#include <linux/device.h>
+#include <linux/kthread.h>
#include <asm/system.h>
#include <asm/irq.h>
@@ -235,22 +236,16 @@
INIT_LIST_HEAD(&socket->cis_cache);
init_completion(&socket->socket_released);
- init_completion(&socket->thread_done);
- init_waitqueue_head(&socket->thread_wait);
init_MUTEX(&socket->skt_sem);
spin_lock_init(&socket->thread_lock);
- ret = kernel_thread(pccardd, socket, CLONE_KERNEL);
- if (ret < 0)
+ socket->thread = kthread_run(pccardd, socket, "pccardd");
+ if (IS_ERR(socket->thread)) {
+ ret = PTR_ERR(socket->thread);
goto err;
-
- wait_for_completion(&socket->thread_done);
- if(!socket->thread) {
- printk(KERN_WARNING "PCMCIA: warning: socket thread for socket %p did not start\n", socket);
- return -EIO;
}
- pcmcia_parse_events(socket, SS_DETECT);
+ pcmcia_parse_events(socket, SS_DETECT);
return 0;
err:
@@ -273,10 +268,8 @@
cs_dbg(socket, 0, "pcmcia_unregister_socket(0x%p)\n", socket->ops);
if (socket->thread) {
- init_completion(&socket->thread_done);
+ kthread_stop(socket->thread);
socket->thread = NULL;
- wake_up(&socket->thread_wait);
- wait_for_completion(&socket->thread_done);
}
release_cis_mem(socket);
@@ -630,12 +623,8 @@
static int pccardd(void *__skt)
{
struct pcmcia_socket *skt = __skt;
- DECLARE_WAITQUEUE(wait, current);
int ret;
- daemonize("pccardd");
-
- skt->thread = current;
skt->socket = dead_socket;
skt->ops->init(skt);
skt->ops->set_socket(skt, &skt->socket);
@@ -645,13 +634,9 @@
if (ret) {
printk(KERN_WARNING "PCMCIA: unable to register socket 0x%p\n",
skt);
- skt->thread = NULL;
- complete_and_exit(&skt->thread_done, 0);
+ return 0;
}
- add_wait_queue(&skt->thread_wait, &wait);
- complete(&skt->thread_done);
-
for (;;) {
unsigned long flags;
unsigned int events;
@@ -677,7 +662,7 @@
continue;
}
- if (!skt->thread)
+ if (!kthread_should_stop())
break;
schedule();
@@ -686,12 +671,10 @@
/* make sure we are running before we exit */
set_current_state(TASK_RUNNING);
- remove_wait_queue(&skt->thread_wait, &wait);
-
/* remove from the device core */
class_device_unregister(&skt->dev);
- complete_and_exit(&skt->thread_done, 0);
+ return 0;
}
/*
@@ -706,7 +689,7 @@
s->thread_events |= events;
spin_unlock(&s->thread_lock);
- wake_up(&s->thread_wait);
+ wake_up_process(s->thread);
}
} /* pcmcia_parse_events */
EXPORT_SYMBOL(pcmcia_parse_events);
Index: linux-2.6/include/pcmcia/ss.h
===================================================================
--- linux-2.6.orig/include/pcmcia/ss.h 2006-01-06 10:44:17.000000000 +0100
+++ linux-2.6/include/pcmcia/ss.h 2006-02-14 18:32:41.000000000 +0100
@@ -244,8 +244,6 @@
struct semaphore skt_sem; /* protects socket h/w state */
struct task_struct *thread;
- struct completion thread_done;
- wait_queue_head_t thread_wait;
spinlock_t thread_lock; /* protects thread_events */
unsigned int thread_events;
On Tue, Feb 14, 2006 at 06:50:16PM +0100, Christoph Hellwig wrote:
> Use the kthread_ API instead of opencoding lots of hairy code for kernel
> thread creation and teardown.
>
> Also use wake_up_process instead of an additional per-socket waitqueue.
Nack. This breaks an error path - look at what happens when we fail to
register a socket - we set socket->thread to NULL and exit. That is
then picked up by the "socket thread for socket %p did not start" code
which you decided to remove.
IOW, this introduces bugs into what was bug free code.
> Signed-off-by: Christoph Hellwig <[email protected]>
>
> Index: linux-2.6/drivers/pcmcia/cs.c
> ===================================================================
> --- linux-2.6.orig/drivers/pcmcia/cs.c 2006-01-06 10:44:16.000000000 +0100
> +++ linux-2.6/drivers/pcmcia/cs.c 2006-02-14 18:34:14.000000000 +0100
> @@ -29,6 +29,7 @@
> #include <linux/pm.h>
> #include <linux/pci.h>
> #include <linux/device.h>
> +#include <linux/kthread.h>
> #include <asm/system.h>
> #include <asm/irq.h>
>
> @@ -235,22 +236,16 @@
> INIT_LIST_HEAD(&socket->cis_cache);
>
> init_completion(&socket->socket_released);
> - init_completion(&socket->thread_done);
> - init_waitqueue_head(&socket->thread_wait);
> init_MUTEX(&socket->skt_sem);
> spin_lock_init(&socket->thread_lock);
>
> - ret = kernel_thread(pccardd, socket, CLONE_KERNEL);
> - if (ret < 0)
> + socket->thread = kthread_run(pccardd, socket, "pccardd");
> + if (IS_ERR(socket->thread)) {
> + ret = PTR_ERR(socket->thread);
> goto err;
> -
> - wait_for_completion(&socket->thread_done);
> - if(!socket->thread) {
> - printk(KERN_WARNING "PCMCIA: warning: socket thread for socket %p did not start\n", socket);
> - return -EIO;
> }
> - pcmcia_parse_events(socket, SS_DETECT);
>
> + pcmcia_parse_events(socket, SS_DETECT);
> return 0;
>
> err:
> @@ -273,10 +268,8 @@
> cs_dbg(socket, 0, "pcmcia_unregister_socket(0x%p)\n", socket->ops);
>
> if (socket->thread) {
> - init_completion(&socket->thread_done);
> + kthread_stop(socket->thread);
> socket->thread = NULL;
> - wake_up(&socket->thread_wait);
> - wait_for_completion(&socket->thread_done);
> }
> release_cis_mem(socket);
>
> @@ -630,12 +623,8 @@
> static int pccardd(void *__skt)
> {
> struct pcmcia_socket *skt = __skt;
> - DECLARE_WAITQUEUE(wait, current);
> int ret;
>
> - daemonize("pccardd");
> -
> - skt->thread = current;
> skt->socket = dead_socket;
> skt->ops->init(skt);
> skt->ops->set_socket(skt, &skt->socket);
> @@ -645,13 +634,9 @@
> if (ret) {
> printk(KERN_WARNING "PCMCIA: unable to register socket 0x%p\n",
> skt);
> - skt->thread = NULL;
> - complete_and_exit(&skt->thread_done, 0);
> + return 0;
> }
>
> - add_wait_queue(&skt->thread_wait, &wait);
> - complete(&skt->thread_done);
> -
> for (;;) {
> unsigned long flags;
> unsigned int events;
> @@ -677,7 +662,7 @@
> continue;
> }
>
> - if (!skt->thread)
> + if (!kthread_should_stop())
> break;
>
> schedule();
> @@ -686,12 +671,10 @@
> /* make sure we are running before we exit */
> set_current_state(TASK_RUNNING);
>
> - remove_wait_queue(&skt->thread_wait, &wait);
> -
> /* remove from the device core */
> class_device_unregister(&skt->dev);
>
> - complete_and_exit(&skt->thread_done, 0);
> + return 0;
> }
>
> /*
> @@ -706,7 +689,7 @@
> s->thread_events |= events;
> spin_unlock(&s->thread_lock);
>
> - wake_up(&s->thread_wait);
> + wake_up_process(s->thread);
> }
> } /* pcmcia_parse_events */
> EXPORT_SYMBOL(pcmcia_parse_events);
> Index: linux-2.6/include/pcmcia/ss.h
> ===================================================================
> --- linux-2.6.orig/include/pcmcia/ss.h 2006-01-06 10:44:17.000000000 +0100
> +++ linux-2.6/include/pcmcia/ss.h 2006-02-14 18:32:41.000000000 +0100
> @@ -244,8 +244,6 @@
> struct semaphore skt_sem; /* protects socket h/w state */
>
> struct task_struct *thread;
> - struct completion thread_done;
> - wait_queue_head_t thread_wait;
> spinlock_t thread_lock; /* protects thread_events */
> unsigned int thread_events;
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
On Tue, Feb 14, 2006 at 07:37:13PM +0000, Russell King wrote:
> On Tue, Feb 14, 2006 at 06:50:16PM +0100, Christoph Hellwig wrote:
> > Use the kthread_ API instead of opencoding lots of hairy code for kernel
> > thread creation and teardown.
> >
> > Also use wake_up_process instead of an additional per-socket waitqueue.
>
> Nack. This breaks an error path - look at what happens when we fail to
> register a socket - we set socket->thread to NULL and exit. That is
> then picked up by the "socket thread for socket %p did not start" code
> which you decided to remove.
Yes, you're right. Let's drop the patch for now.