[PATCH] idle ide disk on resume

[PATCH] idle ide disk on resume

[Jens Axboe]

Hi,

I need this patch to survive suspend on my powerbook, if the drive is
sleeping when suspend is entered. Otherwise it freezes on resume when it
tries to read from the drive.

===== drivers/ide/ide-disk.c 1.86 vs edited =====
--- 1.86/drivers/ide/ide-disk.c	2004-06-05 22:15:29 +02:00
+++ edited/drivers/ide/ide-disk.c	2004-06-22 14:15:08 +02:00
@@ -1334,7 +1334,8 @@
 	idedisk_pm_flush_cache	= ide_pm_state_start_suspend,
 	idedisk_pm_standby,
 
-	idedisk_pm_restore_dma	= ide_pm_state_start_resume,
+	idedisk_pm_idle		= ide_pm_state_start_resume,
+	idedisk_pm_restore_dma,
 };
 
 static void idedisk_complete_power_step (ide_drive_t *drive, struct request *rq, u8 stat, u8 error)
@@ -1349,6 +1350,9 @@
 	case idedisk_pm_standby:	/* Suspend step 2 (standby) complete */
 		rq->pm->pm_step = ide_pm_state_completed;
 		break;
+	case idedisk_pm_idle:		/* resume step 1, idle drive */
+		rq->pm->pm_step = idedisk_pm_restore_dma;
+		break;
 	}
 }
 
@@ -1376,6 +1380,12 @@
 		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_STANDBYNOW1;
 		args->command_type = IDE_DRIVE_TASK_NO_DATA;
 		args->handler	   = &task_no_data_intr;
+		return do_rw_taskfile(drive, args);
+
+	case idedisk_pm_idle:
+		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_IDLEIMMEDIATE;
+		args->command_type = IDE_DRIVE_TASK_NO_DATA;
+		args->handler = task_no_data_intr;
 		return do_rw_taskfile(drive, args);
 
 	case idedisk_pm_restore_dma:	/* Resume step 1 (restore DMA) */

-- 
Jens Axboe

Re: [PATCH] idle ide disk on resume

[Benjamin Herrenschmidt]

On Tue, 2004-06-22 at 07:17, Jens Axboe wrote:
> Hi,
> 
> I need this patch to survive suspend on my powerbook, if the drive is
> sleeping when suspend is entered. Otherwise it freezes on resume when it
> tries to read from the drive.

Interesting, I didn't experience that, I suppose I never suspended the
machine with the disk sleeping ;)

Looks good though, Bart, can you submit to Linus / Andrew please ?

Ben.

> ===== drivers/ide/ide-disk.c 1.86 vs edited =====
> --- 1.86/drivers/ide/ide-disk.c	2004-06-05 22:15:29 +02:00
> +++ edited/drivers/ide/ide-disk.c	2004-06-22 14:15:08 +02:00
> @@ -1334,7 +1334,8 @@
>  	idedisk_pm_flush_cache	= ide_pm_state_start_suspend,
>  	idedisk_pm_standby,
>  
> -	idedisk_pm_restore_dma	= ide_pm_state_start_resume,
> +	idedisk_pm_idle		= ide_pm_state_start_resume,
> +	idedisk_pm_restore_dma,
>  };
>  
>  static void idedisk_complete_power_step (ide_drive_t *drive, struct request *rq, u8 stat, u8 error)
> @@ -1349,6 +1350,9 @@
>  	case idedisk_pm_standby:	/* Suspend step 2 (standby) complete */
>  		rq->pm->pm_step = ide_pm_state_completed;
>  		break;
> +	case idedisk_pm_idle:		/* resume step 1, idle drive */
> +		rq->pm->pm_step = idedisk_pm_restore_dma;
> +		break;
>  	}
>  }
>  
> @@ -1376,6 +1380,12 @@
>  		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_STANDBYNOW1;
>  		args->command_type = IDE_DRIVE_TASK_NO_DATA;
>  		args->handler	   = &task_no_data_intr;
> +		return do_rw_taskfile(drive, args);
> +
> +	case idedisk_pm_idle:
> +		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_IDLEIMMEDIATE;
> +		args->command_type = IDE_DRIVE_TASK_NO_DATA;
> +		args->handler = task_no_data_intr;
>  		return do_rw_taskfile(drive, args);
>  
>  	case idedisk_pm_restore_dma:	/* Resume step 1 (restore DMA) */
-- 
Benjamin Herrenschmidt <benh@kernel.crashing.org>

Re: [PATCH] idle ide disk on resume

[Peter Münster]

On Tue, 22 Jun 2004, Jens Axboe wrote:

> I need this patch to survive suspend on my powerbook, if the drive is
> sleeping when suspend is entered. Otherwise it freezes on resume when it
> tries to read from the drive.

Hello,
unfortunately, with this patch the disk always wakes up when resuming (from
ACPI S1), even if it does not need to.
Could it be possible, to get previous behaviour, that is keeping disk
sleeping when resuming?
Kind regards, Peter

-- 
http://pmrb.free.fr/contact/
_____________________________________
FilmSearch engine: http://f-s.sf.net/

Re: [PATCH] idle ide disk on resume

[Jens Axboe]

On Sat, Aug 21 2004, Peter Münster wrote:
> On Tue, 22 Jun 2004, Jens Axboe wrote:
> 
> > I need this patch to survive suspend on my powerbook, if the drive is
> > sleeping when suspend is entered. Otherwise it freezes on resume when it
> > tries to read from the drive.
> 
> Hello,
> unfortunately, with this patch the disk always wakes up when resuming (from
> ACPI S1), even if it does not need to.
> Could it be possible, to get previous behaviour, that is keeping disk
> sleeping when resuming?

We can probably kill the patch completely, I'm not so sure the missing
idle command was the reason for the drive hang. But my problematic case
was resuming with a previously suspended drive, so...

-- 
Jens Axboe