* [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate
@ 2021-09-23 2:37 Ming Lei
2021-09-23 13:31 ` Jan Kara
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Ming Lei @ 2021-09-23 2:37 UTC (permalink / raw)
To: Jens Axboe; +Cc: linux-block, Ming Lei, Jan Kara
When running ->fallocate(), blkdev_fallocate() should hold
mapping->invalidate_lock to prevent page cache from being accessed,
otherwise stale data may be read in page cache.
Without this patch, blktests block/009 fails sometimes. With this patch,
block/009 can pass always.
Also as Jan pointed out, no pages can be created in the discarded area
while you are holding the invalidate_lock, so remove the 2nd
truncate_bdev_range().
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
V2:
- include <linux/fs.h> for avoiding implicit declaration of function
filemap_invalidate_lock
- remove 2nd truncate_bdev_range() as suggested by Jan
block/fops.c | 21 ++++++++++-----------
1 file changed, 10 insertions(+), 11 deletions(-)
diff --git a/block/fops.c b/block/fops.c
index ffce6f6c68dd..1e970c247e0e 100644
--- a/block/fops.c
+++ b/block/fops.c
@@ -14,6 +14,7 @@
#include <linux/task_io_accounting_ops.h>
#include <linux/falloc.h>
#include <linux/suspend.h>
+#include <linux/fs.h>
#include "blk.h"
static struct inode *bdev_file_inode(struct file *file)
@@ -553,7 +554,8 @@ static ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
static long blkdev_fallocate(struct file *file, int mode, loff_t start,
loff_t len)
{
- struct block_device *bdev = I_BDEV(bdev_file_inode(file));
+ struct inode *inode = bdev_file_inode(file);
+ struct block_device *bdev = I_BDEV(inode);
loff_t end = start + len - 1;
loff_t isize;
int error;
@@ -580,10 +582,12 @@ static long blkdev_fallocate(struct file *file, int mode, loff_t start,
if ((start | len) & (bdev_logical_block_size(bdev) - 1))
return -EINVAL;
+ filemap_invalidate_lock(inode->i_mapping);
+
/* Invalidate the page cache, including dirty pages. */
error = truncate_bdev_range(bdev, file->f_mode, start, end);
if (error)
- return error;
+ goto fail;
switch (mode) {
case FALLOC_FL_ZERO_RANGE:
@@ -600,17 +604,12 @@ static long blkdev_fallocate(struct file *file, int mode, loff_t start,
GFP_KERNEL, 0);
break;
default:
- return -EOPNOTSUPP;
+ error = -EOPNOTSUPP;
}
- if (error)
- return error;
- /*
- * Invalidate the page cache again; if someone wandered in and dirtied
- * a page, we just discard it - userspace has no way of knowing whether
- * the write happened before or after discard completing...
- */
- return truncate_bdev_range(bdev, file->f_mode, start, end);
+ fail:
+ filemap_invalidate_unlock(inode->i_mapping);
+ return error;
}
const struct file_operations def_blk_fops = {
--
2.31.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate
2021-09-23 2:37 [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate Ming Lei
@ 2021-09-23 13:31 ` Jan Kara
2021-09-24 17:07 ` Jens Axboe
2021-11-09 10:46 ` Shinichiro Kawasaki
2 siblings, 0 replies; 4+ messages in thread
From: Jan Kara @ 2021-09-23 13:31 UTC (permalink / raw)
To: Ming Lei; +Cc: Jens Axboe, linux-block, Jan Kara
On Thu 23-09-21 10:37:51, Ming Lei wrote:
> When running ->fallocate(), blkdev_fallocate() should hold
> mapping->invalidate_lock to prevent page cache from being accessed,
> otherwise stale data may be read in page cache.
>
> Without this patch, blktests block/009 fails sometimes. With this patch,
> block/009 can pass always.
>
> Also as Jan pointed out, no pages can be created in the discarded area
> while you are holding the invalidate_lock, so remove the 2nd
> truncate_bdev_range().
>
> Cc: Jan Kara <jack@suse.cz>
> Signed-off-by: Ming Lei <ming.lei@redhat.com>
Looks good. Feel free to add:
Reviewed-by: Jan Kara <jack@suse.cz>
Honza
> ---
> V2:
> - include <linux/fs.h> for avoiding implicit declaration of function
> filemap_invalidate_lock
> - remove 2nd truncate_bdev_range() as suggested by Jan
>
> block/fops.c | 21 ++++++++++-----------
> 1 file changed, 10 insertions(+), 11 deletions(-)
>
> diff --git a/block/fops.c b/block/fops.c
> index ffce6f6c68dd..1e970c247e0e 100644
> --- a/block/fops.c
> +++ b/block/fops.c
> @@ -14,6 +14,7 @@
> #include <linux/task_io_accounting_ops.h>
> #include <linux/falloc.h>
> #include <linux/suspend.h>
> +#include <linux/fs.h>
> #include "blk.h"
>
> static struct inode *bdev_file_inode(struct file *file)
> @@ -553,7 +554,8 @@ static ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
> static long blkdev_fallocate(struct file *file, int mode, loff_t start,
> loff_t len)
> {
> - struct block_device *bdev = I_BDEV(bdev_file_inode(file));
> + struct inode *inode = bdev_file_inode(file);
> + struct block_device *bdev = I_BDEV(inode);
> loff_t end = start + len - 1;
> loff_t isize;
> int error;
> @@ -580,10 +582,12 @@ static long blkdev_fallocate(struct file *file, int mode, loff_t start,
> if ((start | len) & (bdev_logical_block_size(bdev) - 1))
> return -EINVAL;
>
> + filemap_invalidate_lock(inode->i_mapping);
> +
> /* Invalidate the page cache, including dirty pages. */
> error = truncate_bdev_range(bdev, file->f_mode, start, end);
> if (error)
> - return error;
> + goto fail;
>
> switch (mode) {
> case FALLOC_FL_ZERO_RANGE:
> @@ -600,17 +604,12 @@ static long blkdev_fallocate(struct file *file, int mode, loff_t start,
> GFP_KERNEL, 0);
> break;
> default:
> - return -EOPNOTSUPP;
> + error = -EOPNOTSUPP;
> }
> - if (error)
> - return error;
>
> - /*
> - * Invalidate the page cache again; if someone wandered in and dirtied
> - * a page, we just discard it - userspace has no way of knowing whether
> - * the write happened before or after discard completing...
> - */
> - return truncate_bdev_range(bdev, file->f_mode, start, end);
> + fail:
> + filemap_invalidate_unlock(inode->i_mapping);
> + return error;
> }
>
> const struct file_operations def_blk_fops = {
> --
> 2.31.1
>
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate
2021-09-23 2:37 [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate Ming Lei
2021-09-23 13:31 ` Jan Kara
@ 2021-09-24 17:07 ` Jens Axboe
2021-11-09 10:46 ` Shinichiro Kawasaki
2 siblings, 0 replies; 4+ messages in thread
From: Jens Axboe @ 2021-09-24 17:07 UTC (permalink / raw)
To: Ming Lei; +Cc: linux-block, Jan Kara
On 9/22/21 8:37 PM, Ming Lei wrote:
> When running ->fallocate(), blkdev_fallocate() should hold
> mapping->invalidate_lock to prevent page cache from being accessed,
> otherwise stale data may be read in page cache.
>
> Without this patch, blktests block/009 fails sometimes. With this patch,
> block/009 can pass always.
>
> Also as Jan pointed out, no pages can be created in the discarded area
> while you are holding the invalidate_lock, so remove the 2nd
> truncate_bdev_range().
Applied, thanks.
--
Jens Axboe
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate
2021-09-23 2:37 [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate Ming Lei
2021-09-23 13:31 ` Jan Kara
2021-09-24 17:07 ` Jens Axboe
@ 2021-11-09 10:46 ` Shinichiro Kawasaki
2 siblings, 0 replies; 4+ messages in thread
From: Shinichiro Kawasaki @ 2021-11-09 10:46 UTC (permalink / raw)
To: Ming Lei; +Cc: Jens Axboe, linux-block, Jan Kara, Damien Le Moal
On Sep 23, 2021 / 10:37, Ming Lei wrote:
> When running ->fallocate(), blkdev_fallocate() should hold
> mapping->invalidate_lock to prevent page cache from being accessed,
> otherwise stale data may be read in page cache.
>
> Without this patch, blktests block/009 fails sometimes. With this patch,
> block/009 can pass always.
>
> Also as Jan pointed out, no pages can be created in the discarded area
> while you are holding the invalidate_lock, so remove the 2nd
> truncate_bdev_range().
Hello Ming, Jan, thanks for the fix.
Unfortunately, I still observe block/009 failure on the kernel version 5.15.0,
which includes this fix patch. I found that BLKDISCARD ioctl has the same
problem. I modified blk_ioctl_discard() in same manner, and the block/009
failure goes away. I also found that BLKZEROOUT has the same issue. I will
post two patches for these ioctl. Your reviews will be appreciated.
--
Best Regards,
Shin'ichiro Kawasaki
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2021-11-09 10:46 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-09-23 2:37 [PATCH V2] block: hold ->invalidate_lock in blkdev_fallocate Ming Lei
2021-09-23 13:31 ` Jan Kara
2021-09-24 17:07 ` Jens Axboe
2021-11-09 10:46 ` Shinichiro Kawasaki
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).