On Thu, 14 Apr 2011 10:48:30 +0400 Alexey Khoroshilov <khoroshilov@xxxxxxxxx>
wrote:
> mddev_lock() is an alias for mutex_lock_interruptible() at the moment,
> but there are places where returned value of mddev_lock() is not checked.
>
> The patch proposed introduces mddev_lock_interruptible(),
> which is an equivalent for old mddev_lock(), while mddev_lock()
> becomes an alias for uninterruptible mutex_lock().
> Calls of mddev_lock(), where the result is being checked,
> are replaced by mddev_lock_interruptible().
> All the other calls become uninterruptible.
Thanks for reporting this.
I don't agree with your solution though.
I would mark mddev_lock as __must_check__ (or however it is spelt) and fix
the errors produced.
I think the call in md_stop_writes is the only one where an uninterruptible
wait is needed - I'd probaby just change that to an explicit "mutex_lock",
but I wouldn't be against creating "mddev_lock_uninterruptible" for that case.
NeilBrown
>
> Found by Linux Driver Verification project (linuxtesting.org).
>
> Signed-off-by: Alexey Khoroshilov <khoroshilov@xxxxxxxxx>
> ---
> drivers/md/md.c | 21 +++++++++++++--------
> 1 files changed, 13 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index b12b377..cc6d183 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -623,7 +623,12 @@ static mddev_t * mddev_find(dev_t unit)
> goto retry;
> }
>
> -static inline int mddev_lock(mddev_t * mddev)
> +static inline void mddev_lock(mddev_t *mddev)
> +{
> + return mutex_lock(&mddev->reconfig_mutex);
> +}
> +
> +static inline int __must_check mddev_lock_interruptible(mddev_t *mddev)
> {
> return mutex_lock_interruptible(&mddev->reconfig_mutex);
> }
> @@ -2706,7 +2711,7 @@ rdev_attr_show(struct kobject *kobj, struct
> attribute *attr, char *page)
> if (!entry->show)
> return -EIO;
>
> - rv = mddev ? mddev_lock(mddev) : -EBUSY;
> + rv = mddev ? mddev_lock_interruptible(mddev) : -EBUSY;
> if (!rv) {
> if (rdev->mddev == NULL)
> rv = -EBUSY;
> @@ -2730,7 +2735,7 @@ rdev_attr_store(struct kobject *kobj, struct
> attribute *attr,
> return -EIO;
> if (!capable(CAP_SYS_ADMIN))
> return -EACCES;
> - rv = mddev ? mddev_lock(mddev): -EBUSY;
> + rv = mddev ? mddev_lock_interruptible(mddev) : -EBUSY;
> if (!rv) {
> if (rdev->mddev == NULL)
> rv = -EBUSY;
> @@ -4199,7 +4204,7 @@ md_attr_show(struct kobject *kobj, struct
> attribute *attr, char *page)
>
> if (!entry->show)
> return -EIO;
> - rv = mddev_lock(mddev);
> + rv = mddev_lock_interruptible(mddev);
> if (!rv) {
> rv = entry->show(mddev, page);
> mddev_unlock(mddev);
> @@ -4219,7 +4224,7 @@ md_attr_store(struct kobject *kobj, struct
> attribute *attr,
> return -EIO;
> if (!capable(CAP_SYS_ADMIN))
> return -EACCES;
> - rv = mddev_lock(mddev);
> + rv = mddev_lock_interruptible(mddev);
> if (mddev->hold_active == UNTIL_IOCTL)
> mddev->hold_active = 0;
> if (!rv) {
> @@ -4951,7 +4956,7 @@ static void autorun_devices(int part)
> "md: cannot allocate memory for md drive.\n");
> break;
> }
> - if (mddev_lock(mddev))
> + if (mddev_lock_interruptible(mddev))
> printk(KERN_WARNING "md: %s locked, cannot run\n",
> mdname(mddev));
> else if (mddev->raid_disks || mddev->major_version
> @@ -5771,7 +5776,7 @@ static int md_ioctl(struct block_device *bdev,
> fmode_t mode,
> goto abort;
> }
>
> - err = mddev_lock(mddev);
> + err = mddev_lock_interruptible(mddev);
> if (err) {
> printk(KERN_INFO
> "md: ioctl lock interrupted, reason %d, cmd %d\n",
> @@ -6379,7 +6384,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
> return 0;
> }
>
> - if (mddev_lock(mddev) < 0)
> + if (mddev_lock_interruptible(mddev) < 0)
> return -EINTR;
>
> if (mddev->pers || mddev->raid_disks || !list_empty(&mddev->disks)) {
> -- 1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
