ide: fix locking for manual DMA enable/disable ("hdparm -d")

Since hwif->ide_dma_check and hwif->ide_dma_on never queue any commands
(ide_config_drive_speed() sets transfer mode using polling and has no error
recovery) we are safe with setting hwgroup->busy for the time while DMA
setting for a drive is changed (so it won't race against I/O commands in fly).

I audited briefly all ->ide_dma_check/->ide_dma_on/->tuneproc/->speedproc
implementations and they all look OK wrt to this change.

This patch finally allowed me to close kernel bugzilla bug #8169
(once again thanks to Patrick Horn for reporting the issue & testing patches).

Cc: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
This commit is contained in:
Bartlomiej Zolnierkiewicz 2007-03-26 23:03:19 +02:00
parent f68d9320cd
commit 8799620400
1 changed files with 30 additions and 7 deletions

View File

@ -1124,17 +1124,40 @@ static int set_io_32bit(ide_drive_t *drive, int arg)
static int set_using_dma (ide_drive_t *drive, int arg)
{
#ifdef CONFIG_BLK_DEV_IDEDMA
ide_hwif_t *hwif = drive->hwif;
int err = -EPERM;
if (!drive->id || !(drive->id->capability & 1))
return -EPERM;
if (HWIF(drive)->ide_dma_check == NULL)
return -EPERM;
goto out;
if (hwif->ide_dma_check == NULL)
goto out;
err = -EBUSY;
if (ide_spin_wait_hwgroup(drive))
goto out;
/*
* set ->busy flag, unlock and let it ride
*/
hwif->hwgroup->busy = 1;
spin_unlock_irq(&ide_lock);
err = 0;
if (arg) {
if (ide_set_dma(drive))
return -EIO;
if (HWIF(drive)->ide_dma_on(drive)) return -EIO;
if (ide_set_dma(drive) || hwif->ide_dma_on(drive))
err = -EIO;
} else
ide_dma_off(drive);
return 0;
/*
* lock, clear ->busy flag and unlock before leaving
*/
spin_lock_irq(&ide_lock);
hwif->hwgroup->busy = 0;
spin_unlock_irq(&ide_lock);
out:
return err;
#else
return -EPERM;
#endif