The patch titled
Subject: zram: rework reset and destroy path
has been added to the -mm tree. Its filename is
zram-rework-reset-and-destroy-path.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/zram-rework-reset-and-destroy-path.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/zram-rework-reset-and-destroy-path.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
Subject: zram: rework reset and destroy path
We need to return set_capacity(disk, 0) from reset_store() back to
zram_reset_device(), a catch by Ganesh Mahendran. Potentially, we can
race set_capacity() calls from init and reset paths.
The problem is that zram_reset_device() is also getting called from
zram_exit(), which performs operations in misleading reversed order -- we
first create_device() and then init it, while zram_exit() perform
destroy_device() first and then does zram_reset_device(). This is done to
remove sysfs group before we reset device, so we can continue with device
reset/destruction not being raced by sysfs attr write (f.e. disksize).
Apart from that, destroy_device() releases zram->disk (but we still have
->disk pointer), so we cannot acces zram->disk in later
zram_reset_device() call, which may cause additional errors in the future.
So, this patch rework and cleanup destroy path.
1) remove several unneeded goto labels in zram_init()
2) factor out zram_init() error path and zram_exit() into
destroy_devices() function, which takes the number of devices to
destroy as its argument.
3) remove sysfs group in destroy_devices() first, so we can reorder
operations -- reset device (as expected) goes before disk destroy and
queue cleanup. So we can always access ->disk in zram_reset_device().
4) and, finally, return set_capacity() back under ->init_lock.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
Reported-by: Ganesh Mahendran <opensource.ganesh@xxxxxxxxx>
Cc: Minchan Kim <minchan@xxxxxxxxxx>
Cc: Jerome Marchand <jmarchan@xxxxxxxxxx>
Cc: Nitin Gupta <ngupta@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
drivers/block/zram/zram_drv.c | 71 +++++++++++++-------------------
1 file changed, 30 insertions(+), 41 deletions(-)
diff -puN drivers/block/zram/zram_drv.c~zram-rework-reset-and-destroy-path drivers/block/zram/zram_drv.c
--- a/drivers/block/zram/zram_drv.c~zram-rework-reset-and-destroy-path
+++ a/drivers/block/zram/zram_drv.c
@@ -732,8 +732,9 @@ static void zram_reset_device(struct zra
zram->meta = NULL;
/* Reset stats */
memset(&zram->stats, 0, sizeof(zram->stats));
-
zram->disksize = 0;
+ set_capacity(zram->disk, 0);
+
up_write(&zram->init_lock);
}
@@ -826,7 +827,6 @@ static ssize_t reset_store(struct device
/* Make sure all pending I/O is finished */
fsync_bdev(bdev);
zram_reset_device(zram);
- set_capacity(zram->disk, 0);
mutex_unlock(&bdev->bd_mutex);
revalidate_disk(zram->disk);
@@ -1112,15 +1112,29 @@ out:
return ret;
}
-static void destroy_device(struct zram *zram)
+static void destroy_devices(unsigned int nr)
{
- sysfs_remove_group(&disk_to_dev(zram->disk)->kobj,
- &zram_disk_attr_group);
+ struct zram *zram;
+ unsigned int i;
- del_gendisk(zram->disk);
- put_disk(zram->disk);
+ for (i = 0; i < nr; i++) {
+ zram = &zram_devices[i];
+ /* remove sysfs first, so no one will perform disksize
+ * store while we destroying devices */
+ sysfs_remove_group(&disk_to_dev(zram->disk)->kobj,
+ &zram_disk_attr_group);
- blk_cleanup_queue(zram->queue);
+ zram_reset_device(zram);
+
+ del_gendisk(zram->disk);
+ put_disk(zram->disk);
+
+ blk_cleanup_queue(zram->queue);
+ }
+
+ kfree(zram_devices);
+ unregister_blkdev(zram_major, "zram");
+ pr_debug("Destroyed %u device(s)\n", nr);
}
static int __init zram_init(void)
@@ -1130,64 +1144,39 @@ static int __init zram_init(void)
if (num_devices > max_num_devices) {
pr_warn("Invalid value for num_devices: %u\n",
num_devices);
- ret = -EINVAL;
- goto out;
+ return -EINVAL;
}
zram_major = register_blkdev(0, "zram");
if (zram_major <= 0) {
pr_warn("Unable to get major number\n");
- ret = -EBUSY;
- goto out;
+ return -EBUSY;
}
/* Allocate the device array and initialize each one */
zram_devices = kzalloc(num_devices * sizeof(struct zram), GFP_KERNEL);
if (!zram_devices) {
ret = -ENOMEM;
- goto unregister;
+ goto out_error;
}
for (dev_id = 0; dev_id < num_devices; dev_id++) {
ret = create_device(&zram_devices[dev_id], dev_id);
if (ret)
- goto free_devices;
+ goto out_error;
}
- pr_info("Created %u device(s) ...\n", num_devices);
-
+ pr_info("Created %u device(s)\n", num_devices);
return 0;
-free_devices:
- while (dev_id)
- destroy_device(&zram_devices[--dev_id]);
- kfree(zram_devices);
-unregister:
- unregister_blkdev(zram_major, "zram");
-out:
+out_error:
+ destroy_devices(dev_id);
return ret;
}
static void __exit zram_exit(void)
{
- int i;
- struct zram *zram;
-
- for (i = 0; i < num_devices; i++) {
- zram = &zram_devices[i];
-
- destroy_device(zram);
- /*
- * Shouldn't access zram->disk after destroy_device
- * because destroy_device already released zram->disk.
- */
- zram_reset_device(zram);
- }
-
- unregister_blkdev(zram_major, "zram");
-
- kfree(zram_devices);
- pr_debug("Cleanup done!\n");
+ destroy_devices(num_devices);
}
module_init(zram_init);
_
Patches currently in -mm which might be from sergey.senozhatsky@xxxxxxxxx are
zram-clean-up-zram_meta_alloc.patch
zram-free-meta-table-in-zram_meta_free.patch
zram-fix-umount-reset_store-mount-race-condition.patch
zram-rework-reset-and-destroy-path.patch
zram-rework-reset-and-destroy-path-fix.patch
linux-next.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
