On Thu, 19 Jul 2012, Jeff Moyer wrote:
> Mikulas Patocka <mpato...@redhat.com> writes:
>
> > On Tue, 17 Jul 2012, Jeff Moyer wrote:
> >
>
> >> > This is the patch that fixes this crash: it takes a rw-semaphore around
> >> > all direct-IO path.
> >> >
> >> > (note that if someone is concerned about performance, the rw-semaphore
> >> > could be made per-cpu --- take it for read on the current CPU and take
> >> > it
> >> > for write on all CPUs).
> >>
> >> Here we go again. :-) I believe we had at one point tried taking a rw
> >> semaphore around GUP inside of the direct I/O code path to fix the fork
> >> vs. GUP race (that still exists today). When testing that, the overhead
> >> of the semaphore was *way* too high to be considered an acceptable
> >> solution. I've CC'd Larry Woodman, Andrea, and Kosaki Motohiro who all
> >> worked on that particular bug. Hopefully they can give better
> >> quantification of the slowdown than my poor memory.
> >>
> >> Cheers,
> >> Jeff
> >
> > Both down_read and up_read together take 82 ticks on Core2, 69 ticks on
> > AMD K10, 62 ticks on UltraSparc2 if the target is in L1 cache. So, if
> > percpu rw_semaphores were used, it would slow down only by this amount.
>
> Sorry, I'm not familiar with per-cpu rw semaphores. Where are they
> implemented?
Here I'm resending the upstream patches with per rw-semaphores - percpu
rw-semaphores are implemented in the next patch.
(For Jeff: you can use your patch for RHEL-6 that you did for perfocmance
testing, with the change that I proposed).
Mikulas
---
blockdev: fix a crash when block size is changed and I/O is issued
simultaneously
The kernel may crash when block size is changed and I/O is issued
simultaneously.
Because some subsystems (udev or lvm) may read any block device anytime,
the bug actually puts any code that changes a block device size in
jeopardy.
The crash can be reproduced if you place "msleep(1000)" to
blkdev_get_blocks just before "bh->b_size = max_blocks <<
inode->i_blkbits;".
Then, run "dd if=/dev/ram0 of=/dev/null bs=4k count=1 iflag=direct"
While it is waiting in msleep, run "blockdev --setbsz 2048 /dev/ram0"
You get a BUG.
The direct and non-direct I/O is written with the assumption that block
size does not change. It doesn't seem practical to fix these crashes
one-by-one there may be many crash possibilities when block size changes
at a certain place and it is impossible to find them all and verify the
code.
This patch introduces a new rw-lock bd_block_size_semaphore. The lock is
taken for read during I/O. It is taken for write when changing block
size. Consequently, block size can't be changed while I/O is being
submitted.
For asynchronous I/O, the patch only prevents block size change while
the I/O is being submitted. The block size can change when the I/O is in
progress or when the I/O is being finished. This is acceptable because
there are no accesses to block size when asynchronous I/O is being
finished.
The patch prevents block size changing while the device is mapped with
mmap.
Signed-off-by: Mikulas Patocka <mpato...@redhat.com>
---
drivers/char/raw.c | 2 -
fs/block_dev.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++++++--
include/linux/fs.h | 4 +++
3 files changed, 63 insertions(+), 3 deletions(-)
Index: linux-3.5-rc6-devel/include/linux/fs.h
===================================================================
--- linux-3.5-rc6-devel.orig/include/linux/fs.h 2012-07-16 20:20:12.000000000
+0200
+++ linux-3.5-rc6-devel/include/linux/fs.h 2012-07-16 01:29:21.000000000
+0200
@@ -713,6 +713,8 @@ struct block_device {
int bd_fsfreeze_count;
/* Mutex for freeze */
struct mutex bd_fsfreeze_mutex;
+ /* A semaphore that prevents I/O while block size is being changed */
+ struct rw_semaphore bd_block_size_semaphore;
};
/*
@@ -2414,6 +2416,8 @@ extern int generic_segment_checks(const
unsigned long *nr_segs, size_t *count, int access_flags);
/* fs/block_dev.c */
+extern ssize_t blkdev_aio_read(struct kiocb *iocb, const struct iovec *iov,
+ unsigned long nr_segs, loff_t pos);
extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos);
extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
Index: linux-3.5-rc6-devel/fs/block_dev.c
===================================================================
--- linux-3.5-rc6-devel.orig/fs/block_dev.c 2012-07-16 20:20:12.000000000
+0200
+++ linux-3.5-rc6-devel/fs/block_dev.c 2012-07-16 21:47:30.000000000 +0200
@@ -116,6 +116,8 @@ EXPORT_SYMBOL(invalidate_bdev);
int set_blocksize(struct block_device *bdev, int size)
{
+ struct address_space *mapping;
+
/* Size must be a power of two, and between 512 and PAGE_SIZE */
if (size > PAGE_SIZE || size < 512 || !is_power_of_2(size))
return -EINVAL;
@@ -124,6 +126,20 @@ int set_blocksize(struct block_device *b
if (size < bdev_logical_block_size(bdev))
return -EINVAL;
+ /* Prevent starting I/O or mapping the device */
+ down_write(&bdev->bd_block_size_semaphore);
+
+ /* Check that the block device is not memory mapped */
+ mapping = bdev->bd_inode->i_mapping;
+ mutex_lock(&mapping->i_mmap_mutex);
+ if (!prio_tree_empty(&mapping->i_mmap) ||
+ !list_empty(&mapping->i_mmap_nonlinear)) {
+ mutex_unlock(&mapping->i_mmap_mutex);
+ up_write(&bdev->bd_block_size_semaphore);
+ return -EBUSY;
+ }
+ mutex_unlock(&mapping->i_mmap_mutex);
+
/* Don't change the size if it is same as current */
if (bdev->bd_block_size != size) {
sync_blockdev(bdev);
@@ -131,6 +147,9 @@ int set_blocksize(struct block_device *b
bdev->bd_inode->i_blkbits = blksize_bits(size);
kill_bdev(bdev);
}
+
+ up_write(&bdev->bd_block_size_semaphore);
+
return 0;
}
@@ -472,6 +491,7 @@ static void init_once(void *foo)
inode_init_once(&ei->vfs_inode);
/* Initialize mutex for freeze. */
mutex_init(&bdev->bd_fsfreeze_mutex);
+ init_rwsem(&bdev->bd_block_size_semaphore);
}
static inline void __bd_forget(struct inode *inode)
@@ -1567,6 +1587,22 @@ static long block_ioctl(struct file *fil
return blkdev_ioctl(bdev, mode, cmd, arg);
}
+ssize_t blkdev_aio_read(struct kiocb *iocb, const struct iovec *iov,
+ unsigned long nr_segs, loff_t pos)
+{
+ ssize_t ret;
+ struct block_device *bdev = I_BDEV(iocb->ki_filp->f_mapping->host);
+
+ down_read(&bdev->bd_block_size_semaphore);
+
+ ret = generic_file_aio_read(iocb, iov, nr_segs, pos);
+
+ up_read(&bdev->bd_block_size_semaphore);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(blkdev_aio_read);
+
/*
* Write data to the block device. Only intended for the block device itself
* and the raw driver which basically is a fake block device.
@@ -1578,10 +1614,13 @@ ssize_t blkdev_aio_write(struct kiocb *i
unsigned long nr_segs, loff_t pos)
{
struct file *file = iocb->ki_filp;
+ struct block_device *bdev = I_BDEV(file->f_mapping->host);
ssize_t ret;
BUG_ON(iocb->ki_pos != pos);
+ down_read(&bdev->bd_block_size_semaphore);
+
ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
if (ret > 0 || ret == -EIOCBQUEUED) {
ssize_t err;
@@ -1590,10 +1629,27 @@ ssize_t blkdev_aio_write(struct kiocb *i
if (err < 0 && ret > 0)
ret = err;
}
+
+ up_read(&bdev->bd_block_size_semaphore);
+
return ret;
}
EXPORT_SYMBOL_GPL(blkdev_aio_write);
+int blkdev_mmap(struct file *file, struct vm_area_struct *vma)
+{
+ int ret;
+ struct block_device *bdev = I_BDEV(file->f_mapping->host);
+
+ down_read(&bdev->bd_block_size_semaphore);
+
+ ret = generic_file_mmap(file, vma);
+
+ up_read(&bdev->bd_block_size_semaphore);
+
+ return ret;
+}
+
/*
* Try to release a page associated with block device when the system
* is under memory pressure.
@@ -1624,9 +1680,9 @@ const struct file_operations def_blk_fop
.llseek = block_llseek,
.read = do_sync_read,
.write = do_sync_write,
- .aio_read = generic_file_aio_read,
+ .aio_read = blkdev_aio_read,
.aio_write = blkdev_aio_write,
- .mmap = generic_file_mmap,
+ .mmap = blkdev_mmap,
.fsync = blkdev_fsync,
.unlocked_ioctl = block_ioctl,
#ifdef CONFIG_COMPAT
Index: linux-3.5-rc6-devel/drivers/char/raw.c
===================================================================
--- linux-3.5-rc6-devel.orig/drivers/char/raw.c 2012-07-16 20:20:12.000000000
+0200
+++ linux-3.5-rc6-devel/drivers/char/raw.c 2012-07-16 01:30:04.000000000
+0200
@@ -285,7 +285,7 @@ static long raw_ctl_compat_ioctl(struct
static const struct file_operations raw_fops = {
.read = do_sync_read,
- .aio_read = generic_file_aio_read,
+ .aio_read = blkdev_aio_read,
.write = do_sync_write,
.aio_write = blkdev_aio_write,
.fsync = blkdev_fsync,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
