schedule inside spin_lock_irqsave
in [Kernel]
|
Prev: VFS: add memory barrier to sb_mark_clean and sb_mark_dirty
Next: Add a target to use the Coccinelle checker
From: Jiri Slaby on 6 Jun 2010 14:00 On 06/06/2010 02:43 PM, Richard Zidlicky wrote: > Hi, > > I have done a minimaly invasive patch for the stable 2.6.34 kernel and stress-tested > it for many hours, definitely seems to improve the behaviour. > > I have left out your beautification suggestion for now, want to do more playing with > other aspects of the driver. There still seem to be issues when the device is unplugged > while in use and such. > > --- linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c.rz 2010-06-03 21:58:11.000000000 +0200 > +++ linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c 2010-06-04 23:00:35.000000000 +0200 > @@ -1100,31 +1100,26 @@ > * > * @return pointer to descriptor on success, NULL on error. > */ > -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) > + > +struct smscore_buffer_t *get_entry(void) > { > struct smscore_buffer_t *cb = NULL; > unsigned long flags; > > - DEFINE_WAIT(wait); > - > spin_lock_irqsave(&coredev->bufferslock, flags); Sorry, maybe I'm just blind, but where is 'coredev' defined in this scope? You probably forgot to pass it to get_entry? How could this be compiled? Is there coredev defined globally? > - > - /* This function must return a valid buffer, since the buffer list is > - * finite, we check that there is an available buffer, if not, we wait > - * until such buffer become available. > - */ > - > - prepare_to_wait(&coredev->buffer_mng_waitq, &wait, TASK_INTERRUPTIBLE); > - > - if (list_empty(&coredev->buffers)) > - schedule(); > - > - finish_wait(&coredev->buffer_mng_waitq, &wait); > - > + if (!list_empty(&coredev->buffers)) { > cb = (struct smscore_buffer_t *) coredev->buffers.next; > list_del(&cb->entry); > - > + } > spin_unlock_irqrestore(&coredev->bufferslock, flags); > + return cb; > +} > + > +struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) > +{ > + struct smscore_buffer_t *cb = NULL; > + > + wait_event(coredev->buffer_mng_waitq, (cb = get_entry())); > > return cb; > } -- js -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Richard Zidlicky on 7 Jun 2010 09:00 On Sun, Jun 06, 2010 at 07:51:56PM +0200, Jiri Slaby wrote: > On 06/06/2010 02:43 PM, Richard Zidlicky wrote: > > Hi, > > > > I have done a minimaly invasive patch for the stable 2.6.34 kernel and stress-tested > > it for many hours, definitely seems to improve the behaviour. > > > > I have left out your beautification suggestion for now, want to do more playing with > > other aspects of the driver. There still seem to be issues when the device is unplugged > > while in use and such. > > > > --- linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c.rz 2010-06-03 21:58:11.000000000 +0200 > > +++ linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c 2010-06-04 23:00:35.000000000 +0200 > > @@ -1100,31 +1100,26 @@ > > * > > * @return pointer to descriptor on success, NULL on error. > > */ > > -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) > > + > > +struct smscore_buffer_t *get_entry(void) > > { > > struct smscore_buffer_t *cb = NULL; > > unsigned long flags; > > > > - DEFINE_WAIT(wait); > > - > > spin_lock_irqsave(&coredev->bufferslock, flags); > > Sorry, maybe I'm just blind, but where is 'coredev' defined in this > scope? You probably forgot to pass it to get_entry? > > How could this be compiled? Is there coredev defined globally? good catch. I think it failed and despite a different kernel id the old module was loaded. Here is the new version, this time lightly tested --- linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c.rz 2010-06-03 21:58:11.000000000 +0200 +++ linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c 2010-06-07 14:32:06.000000000 +0200 @@ -1100,31 +1100,26 @@ * * @return pointer to descriptor on success, NULL on error. */ -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) + +struct smscore_buffer_t *get_entry(struct smscore_device_t *coredev) { struct smscore_buffer_t *cb = NULL; unsigned long flags; - DEFINE_WAIT(wait); - spin_lock_irqsave(&coredev->bufferslock, flags); - - /* This function must return a valid buffer, since the buffer list is - * finite, we check that there is an available buffer, if not, we wait - * until such buffer become available. - */ - - prepare_to_wait(&coredev->buffer_mng_waitq, &wait, TASK_INTERRUPTIBLE); - - if (list_empty(&coredev->buffers)) - schedule(); - - finish_wait(&coredev->buffer_mng_waitq, &wait); - + if (!list_empty(&coredev->buffers)) { cb = (struct smscore_buffer_t *) coredev->buffers.next; list_del(&cb->entry); - + } spin_unlock_irqrestore(&coredev->bufferslock, flags); + return cb; +} + +struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) +{ + struct smscore_buffer_t *cb = NULL; + + wait_event(coredev->buffer_mng_waitq, (cb = get_entry(coredev))); return cb; } Richard -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Mauro Carvalho Chehab on 6 Jul 2010 08:50
Em 07-06-2010 10:00, Richard Zidlicky escreveu: > On Sun, Jun 06, 2010 at 07:51:56PM +0200, Jiri Slaby wrote: >> On 06/06/2010 02:43 PM, Richard Zidlicky wrote: >>> Hi, >>> >>> I have done a minimaly invasive patch for the stable 2.6.34 kernel and stress-tested >>> it for many hours, definitely seems to improve the behaviour. >>> >>> I have left out your beautification suggestion for now, want to do more playing with >>> other aspects of the driver. There still seem to be issues when the device is unplugged >>> while in use and such. >>> >>> --- linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c.rz 2010-06-03 21:58:11.000000000 +0200 >>> +++ linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c 2010-06-04 23:00:35.000000000 +0200 >>> @@ -1100,31 +1100,26 @@ >>> * >>> * @return pointer to descriptor on success, NULL on error. >>> */ >>> -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) >>> + >>> +struct smscore_buffer_t *get_entry(void) >>> { >>> struct smscore_buffer_t *cb = NULL; >>> unsigned long flags; >>> >>> - DEFINE_WAIT(wait); >>> - >>> spin_lock_irqsave(&coredev->bufferslock, flags); >> >> Sorry, maybe I'm just blind, but where is 'coredev' defined in this >> scope? You probably forgot to pass it to get_entry? >> >> How could this be compiled? Is there coredev defined globally? > > good catch. I think it failed and despite a different kernel id the old module was > loaded. > > Here is the new version, this time lightly tested Could you please fix the indentation and send your SOB for this patch? Cheers, Mauro. > > --- linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c.rz 2010-06-03 21:58:11.000000000 +0200 > +++ linux-2.6.34/drivers/media/dvb/siano/smscoreapi.c 2010-06-07 14:32:06.000000000 +0200 > @@ -1100,31 +1100,26 @@ > * > * @return pointer to descriptor on success, NULL on error. > */ > -struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) > + > +struct smscore_buffer_t *get_entry(struct smscore_device_t *coredev) > { > struct smscore_buffer_t *cb = NULL; > unsigned long flags; > > - DEFINE_WAIT(wait); > - > spin_lock_irqsave(&coredev->bufferslock, flags); > - > - /* This function must return a valid buffer, since the buffer list is > - * finite, we check that there is an available buffer, if not, we wait > - * until such buffer become available. > - */ > - > - prepare_to_wait(&coredev->buffer_mng_waitq, &wait, TASK_INTERRUPTIBLE); > - > - if (list_empty(&coredev->buffers)) > - schedule(); > - > - finish_wait(&coredev->buffer_mng_waitq, &wait); > - > + if (!list_empty(&coredev->buffers)) { > cb = (struct smscore_buffer_t *) coredev->buffers.next; > list_del(&cb->entry); > - > + } > spin_unlock_irqrestore(&coredev->bufferslock, flags); > + return cb; > +} > + > +struct smscore_buffer_t *smscore_getbuffer(struct smscore_device_t *coredev) > +{ > + struct smscore_buffer_t *cb = NULL; > + > + wait_event(coredev->buffer_mng_waitq, (cb = get_entry(coredev))); > > return cb; > } > > > Richard > -- > To unsubscribe from this list: send the line "unsubscribe linux-media" in > the body of a message to majordomo(a)vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |