linux_dsm_epyc7002/drivers/block/loop.c

2059 lines
50 KiB
C
Raw Normal View History

/*
* linux/drivers/block/loop.c
*
* Written by Theodore Ts'o, 3/29/93
*
* Copyright 1993 by Theodore Ts'o. Redistribution of this file is
* permitted under the GNU General Public License.
*
* DES encryption plus some minor changes by Werner Almesberger, 30-MAY-1993
* more DES encryption plus IDEA encryption by Nicholas J. Leon, June 20, 1996
*
* Modularized and updated for 1.1.16 kernel - Mitch Dsouza 28th May 1994
* Adapted for 1.3.59 kernel - Andries Brouwer, 1 Feb 1996
*
* Fixed do_loop_request() re-entrancy - Vincent.Renardias@waw.com Mar 20, 1997
*
* Added devfs support - Richard Gooch <rgooch@atnf.csiro.au> 16-Jan-1998
*
* Handle sparse backing files correctly - Kenn Humborg, Jun 28, 1998
*
* Loadable modules and other fixes by AK, 1998
*
* Make real block number available to downstream transfer functions, enables
* CBC (and relatives) mode encryption requiring unique IVs per data block.
* Reed H. Petty, rhp@draper.net
*
* Maximum number of loop devices now dynamic via max_loop module parameter.
* Russell Kroll <rkroll@exploits.org> 19990701
*
* Maximum number of loop devices when compiled-in now selectable by passing
* max_loop=<1-255> to the kernel on boot.
* Erik I. Bolsø, <eriki@himolde.no>, Oct 31, 1999
*
* Completely rewrite request handling to be make_request_fn style and
* non blocking, pushing work to a helper thread. Lots of fixes from
* Al Viro too.
* Jens Axboe <axboe@suse.de>, Nov 2000
*
* Support up to 256 loop devices
* Heinz Mauelshagen <mge@sistina.com>, Feb 2002
*
* Support for falling back on the write file operation when the address space
* operations write_begin is not available on the backing filesystem.
* Anton Altaparmakov, 16 Feb 2005
*
* Still To Fix:
* - Advisory locking is ignored here.
* - Should use an own CAP_* category instead of CAP_SYS_ADMIN
*
*/
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/stat.h>
#include <linux/errno.h>
#include <linux/major.h>
#include <linux/wait.h>
#include <linux/blkdev.h>
#include <linux/blkpg.h>
#include <linux/init.h>
#include <linux/swap.h>
#include <linux/slab.h>
#include <linux/compat.h>
#include <linux/suspend.h>
#include <linux/freezer.h>
#include <linux/mutex.h>
#include <linux/writeback.h>
#include <linux/completion.h>
#include <linux/highmem.h>
#include <linux/kthread.h>
#include <linux/splice.h>
#include <linux/sysfs.h>
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
#include <linux/miscdevice.h>
#include <linux/falloc.h>
#include <linux/uio.h>
#include "loop.h"
#include <asm/uaccess.h>
static DEFINE_IDR(loop_index_idr);
static DEFINE_MUTEX(loop_index_mutex);
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
static int max_part;
static int part_shift;
static int transfer_xor(struct loop_device *lo, int cmd,
struct page *raw_page, unsigned raw_off,
struct page *loop_page, unsigned loop_off,
int size, sector_t real_block)
{
char *raw_buf = kmap_atomic(raw_page) + raw_off;
char *loop_buf = kmap_atomic(loop_page) + loop_off;
char *in, *out, *key;
int i, keysize;
if (cmd == READ) {
in = raw_buf;
out = loop_buf;
} else {
in = loop_buf;
out = raw_buf;
}
key = lo->lo_encrypt_key;
keysize = lo->lo_encrypt_key_size;
for (i = 0; i < size; i++)
*out++ = *in++ ^ key[(i & 511) % keysize];
kunmap_atomic(loop_buf);
kunmap_atomic(raw_buf);
cond_resched();
return 0;
}
static int xor_init(struct loop_device *lo, const struct loop_info64 *info)
{
if (unlikely(info->lo_encrypt_key_size <= 0))
return -EINVAL;
return 0;
}
static struct loop_func_table none_funcs = {
.number = LO_CRYPT_NONE,
};
static struct loop_func_table xor_funcs = {
.number = LO_CRYPT_XOR,
.transfer = transfer_xor,
.init = xor_init
};
/* xfer_funcs[0] is special - its release function is never called */
static struct loop_func_table *xfer_funcs[MAX_LO_CRYPT] = {
&none_funcs,
&xor_funcs
};
static loff_t get_size(loff_t offset, loff_t sizelimit, struct file *file)
{
loff_t loopsize;
/* Compute loopsize in bytes */
loopsize = i_size_read(file->f_mapping->host);
if (offset > 0)
loopsize -= offset;
/* offset is beyond i_size, weird but possible */
if (loopsize < 0)
return 0;
if (sizelimit > 0 && sizelimit < loopsize)
loopsize = sizelimit;
/*
* Unfortunately, if we want to do I/O on the device,
* the number of 512-byte sectors has to fit into a sector_t.
*/
return loopsize >> 9;
}
static loff_t get_loop_size(struct loop_device *lo, struct file *file)
{
return get_size(lo->lo_offset, lo->lo_sizelimit, file);
}
static void __loop_update_dio(struct loop_device *lo, bool dio)
{
struct file *file = lo->lo_backing_file;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
unsigned short sb_bsize = 0;
unsigned dio_align = 0;
bool use_dio;
if (inode->i_sb->s_bdev) {
sb_bsize = bdev_logical_block_size(inode->i_sb->s_bdev);
dio_align = sb_bsize - 1;
}
/*
* We support direct I/O only if lo_offset is aligned with the
* logical I/O size of backing device, and the logical block
* size of loop is bigger than the backing device's and the loop
* needn't transform transfer.
*
* TODO: the above condition may be loosed in the future, and
* direct I/O may be switched runtime at that time because most
* of requests in sane appplications should be PAGE_SIZE algined
*/
if (dio) {
if (queue_logical_block_size(lo->lo_queue) >= sb_bsize &&
!(lo->lo_offset & dio_align) &&
mapping->a_ops->direct_IO &&
!lo->transfer)
use_dio = true;
else
use_dio = false;
} else {
use_dio = false;
}
if (lo->use_dio == use_dio)
return;
/* flush dirty pages before changing direct IO */
vfs_fsync(file, 0);
/*
* The flag of LO_FLAGS_DIRECT_IO is handled similarly with
* LO_FLAGS_READ_ONLY, both are set from kernel, and losetup
* will get updated by ioctl(LOOP_GET_STATUS)
*/
blk_mq_freeze_queue(lo->lo_queue);
lo->use_dio = use_dio;
if (use_dio)
lo->lo_flags |= LO_FLAGS_DIRECT_IO;
else
lo->lo_flags &= ~LO_FLAGS_DIRECT_IO;
blk_mq_unfreeze_queue(lo->lo_queue);
}
static int
figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
{
loff_t size = get_size(offset, sizelimit, lo->lo_backing_file);
sector_t x = (sector_t)size;
struct block_device *bdev = lo->lo_device;
if (unlikely((loff_t)x != size))
return -EFBIG;
if (lo->lo_offset != offset)
lo->lo_offset = offset;
if (lo->lo_sizelimit != sizelimit)
lo->lo_sizelimit = sizelimit;
set_capacity(lo->lo_disk, x);
bd_set_size(bdev, (loff_t)get_capacity(bdev->bd_disk) << 9);
/* let user-space know about the new size */
kobject_uevent(&disk_to_dev(bdev->bd_disk)->kobj, KOBJ_CHANGE);
return 0;
}
static inline int
lo_do_transfer(struct loop_device *lo, int cmd,
struct page *rpage, unsigned roffs,
struct page *lpage, unsigned loffs,
int size, sector_t rblock)
{
int ret;
ret = lo->transfer(lo, cmd, rpage, roffs, lpage, loffs, size, rblock);
if (likely(!ret))
return 0;
printk_ratelimited(KERN_ERR
"loop: Transfer error at byte offset %llu, length %i.\n",
(unsigned long long)rblock << 9, size);
return ret;
}
static int lo_write_bvec(struct file *file, struct bio_vec *bvec, loff_t *ppos)
{
struct iov_iter i;
ssize_t bw;
iov_iter_bvec(&i, ITER_BVEC, bvec, 1, bvec->bv_len);
file_start_write(file);
bw = vfs_iter_write(file, &i, ppos);
file_end_write(file);
if (likely(bw == bvec->bv_len))
return 0;
printk_ratelimited(KERN_ERR
"loop: Write error at byte offset %llu, length %i.\n",
(unsigned long long)*ppos, bvec->bv_len);
if (bw >= 0)
bw = -EIO;
return bw;
}
static int lo_write_simple(struct loop_device *lo, struct request *rq,
loff_t pos)
{
struct bio_vec bvec;
struct req_iterator iter;
int ret = 0;
rq_for_each_segment(bvec, rq, iter) {
ret = lo_write_bvec(lo->lo_backing_file, &bvec, &pos);
if (ret < 0)
break;
cond_resched();
}
return ret;
}
/*
* This is the slow, transforming version that needs to double buffer the
* data as it cannot do the transformations in place without having direct
* access to the destination pages of the backing file.
*/
static int lo_write_transfer(struct loop_device *lo, struct request *rq,
loff_t pos)
{
struct bio_vec bvec, b;
struct req_iterator iter;
struct page *page;
block: Convert bio_for_each_segment() to bvec_iter More prep work for immutable biovecs - with immutable bvecs drivers won't be able to use the biovec directly, they'll need to use helpers that take into account bio->bi_iter.bi_bvec_done. This updates callers for the new usage without changing the implementation yet. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Paul Clements <Paul.Clements@steeleye.com> Cc: Jim Paris <jim@jtan.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Nagalakshmi Nandigama <Nagalakshmi.Nandigama@lsi.com> Cc: Sreekanth Reddy <Sreekanth.Reddy@lsi.com> Cc: support@lsi.com Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Tejun Heo <tj@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Matthew Wilcox <matthew.r.wilcox@intel.com> Cc: Keith Busch <keith.busch@intel.com> Cc: Stephen Hemminger <shemminger@vyatta.com> Cc: Quoc-Son Anh <quoc-sonx.anh@intel.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Jerome Marchand <jmarchan@redhat.com> Cc: Seth Jennings <sjenning@linux.vnet.ibm.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: "Darrick J. Wong" <darrick.wong@oracle.com> Cc: Chris Metcalf <cmetcalf@tilera.com> Cc: Jan Kara <jack@suse.cz> Cc: linux-m68k@lists.linux-m68k.org Cc: linuxppc-dev@lists.ozlabs.org Cc: drbd-user@lists.linbit.com Cc: nbd-general@lists.sourceforge.net Cc: cbe-oss-dev@lists.ozlabs.org Cc: xen-devel@lists.xensource.com Cc: virtualization@lists.linux-foundation.org Cc: linux-raid@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: DL-MPTFusionLinux@lsi.com Cc: linux-scsi@vger.kernel.org Cc: devel@driverdev.osuosl.org Cc: linux-fsdevel@vger.kernel.org Cc: cluster-devel@redhat.com Cc: linux-mm@kvack.org Acked-by: Geoff Levand <geoff@infradead.org>
2013-11-24 08:19:00 +07:00
int ret = 0;
page = alloc_page(GFP_NOIO);
if (unlikely(!page))
return -ENOMEM;
rq_for_each_segment(bvec, rq, iter) {
ret = lo_do_transfer(lo, WRITE, page, 0, bvec.bv_page,
bvec.bv_offset, bvec.bv_len, pos >> 9);
if (unlikely(ret))
break;
b.bv_page = page;
b.bv_offset = 0;
b.bv_len = bvec.bv_len;
ret = lo_write_bvec(lo->lo_backing_file, &b, &pos);
if (ret < 0)
break;
}
__free_page(page);
return ret;
}
static int lo_read_simple(struct loop_device *lo, struct request *rq,
loff_t pos)
{
struct bio_vec bvec;
struct req_iterator iter;
struct iov_iter i;
ssize_t len;
rq_for_each_segment(bvec, rq, iter) {
iov_iter_bvec(&i, ITER_BVEC, &bvec, 1, bvec.bv_len);
len = vfs_iter_read(lo->lo_backing_file, &i, &pos);
if (len < 0)
return len;
flush_dcache_page(bvec.bv_page);
if (len != bvec.bv_len) {
struct bio *bio;
__rq_for_each_bio(bio, rq)
zero_fill_bio(bio);
break;
}
cond_resched();
}
return 0;
}
static int lo_read_transfer(struct loop_device *lo, struct request *rq,
loff_t pos)
{
struct bio_vec bvec, b;
struct req_iterator iter;
struct iov_iter i;
struct page *page;
ssize_t len;
int ret = 0;
page = alloc_page(GFP_NOIO);
if (unlikely(!page))
return -ENOMEM;
rq_for_each_segment(bvec, rq, iter) {
loff_t offset = pos;
b.bv_page = page;
b.bv_offset = 0;
b.bv_len = bvec.bv_len;
iov_iter_bvec(&i, ITER_BVEC, &b, 1, b.bv_len);
len = vfs_iter_read(lo->lo_backing_file, &i, &pos);
if (len < 0) {
ret = len;
goto out_free_page;
}
ret = lo_do_transfer(lo, READ, page, 0, bvec.bv_page,
bvec.bv_offset, len, offset >> 9);
if (ret)
goto out_free_page;
flush_dcache_page(bvec.bv_page);
if (len != bvec.bv_len) {
struct bio *bio;
__rq_for_each_bio(bio, rq)
zero_fill_bio(bio);
break;
}
}
ret = 0;
out_free_page:
__free_page(page);
return ret;
}
static int lo_discard(struct loop_device *lo, struct request *rq, loff_t pos)
{
/*
* We use punch hole to reclaim the free space used by the
* image a.k.a. discard. However we do not support discard if
* encryption is enabled, because it may give an attacker
* useful information.
*/
struct file *file = lo->lo_backing_file;
int mode = FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE;
int ret;
if ((!file->f_op->fallocate) || lo->lo_encrypt_key_size) {
ret = -EOPNOTSUPP;
goto out;
}
ret = file->f_op->fallocate(file, mode, pos, blk_rq_bytes(rq));
if (unlikely(ret && ret != -EINVAL && ret != -EOPNOTSUPP))
ret = -EIO;
out:
return ret;
}
static int lo_req_flush(struct loop_device *lo, struct request *rq)
{
struct file *file = lo->lo_backing_file;
int ret = vfs_fsync(file, 0);
if (unlikely(ret && ret != -EINVAL))
ret = -EIO;
return ret;
}
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
static inline void handle_partial_read(struct loop_cmd *cmd, long bytes)
{
if (bytes < 0 || (cmd->rq->cmd_flags & REQ_WRITE))
return;
if (unlikely(bytes < blk_rq_bytes(cmd->rq))) {
struct bio *bio = cmd->rq->bio;
bio_advance(bio, bytes);
zero_fill_bio(bio);
}
}
static void lo_rw_aio_complete(struct kiocb *iocb, long ret, long ret2)
{
struct loop_cmd *cmd = container_of(iocb, struct loop_cmd, iocb);
struct request *rq = cmd->rq;
handle_partial_read(cmd, ret);
if (ret > 0)
ret = 0;
else if (ret < 0)
ret = -EIO;
blk_mq_complete_request(rq, ret);
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
}
static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd,
loff_t pos, bool rw)
{
struct iov_iter iter;
struct bio_vec *bvec;
struct bio *bio = cmd->rq->bio;
struct file *file = lo->lo_backing_file;
int ret;
/* nomerge for loop request queue */
WARN_ON(cmd->rq->bio != cmd->rq->biotail);
bvec = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
iov_iter_bvec(&iter, ITER_BVEC | rw, bvec,
bio_segments(bio), blk_rq_bytes(cmd->rq));
/*
* This bio may be started from the middle of the 'bvec'
* because of bio splitting, so offset from the bvec must
* be passed to iov iterator
*/
iter.iov_offset = bio->bi_iter.bi_bvec_done;
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
cmd->iocb.ki_pos = pos;
cmd->iocb.ki_filp = file;
cmd->iocb.ki_complete = lo_rw_aio_complete;
cmd->iocb.ki_flags = IOCB_DIRECT;
if (rw == WRITE)
ret = file->f_op->write_iter(&cmd->iocb, &iter);
else
ret = file->f_op->read_iter(&cmd->iocb, &iter);
if (ret != -EIOCBQUEUED)
cmd->iocb.ki_complete(&cmd->iocb, ret, 0);
return 0;
}
static inline int lo_rw_simple(struct loop_device *lo,
struct request *rq, loff_t pos, bool rw)
{
struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
if (cmd->use_aio)
return lo_rw_aio(lo, cmd, pos, rw);
/*
* lo_write_simple and lo_read_simple should have been covered
* by io submit style function like lo_rw_aio(), one blocker
* is that lo_read_simple() need to call flush_dcache_page after
* the page is written from kernel, and it isn't easy to handle
* this in io submit style function which submits all segments
* of the req at one time. And direct read IO doesn't need to
* run flush_dcache_page().
*/
if (rw == WRITE)
return lo_write_simple(lo, rq, pos);
else
return lo_read_simple(lo, rq, pos);
}
static int do_req_filebacked(struct loop_device *lo, struct request *rq)
{
loff_t pos;
int ret;
pos = ((loff_t) blk_rq_pos(rq) << 9) + lo->lo_offset;
if (rq->cmd_flags & REQ_WRITE) {
if (rq->cmd_flags & REQ_FLUSH)
ret = lo_req_flush(lo, rq);
else if (rq->cmd_flags & REQ_DISCARD)
ret = lo_discard(lo, rq, pos);
else if (lo->transfer)
ret = lo_write_transfer(lo, rq, pos);
else
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
ret = lo_rw_simple(lo, rq, pos, WRITE);
} else {
if (lo->transfer)
ret = lo_read_transfer(lo, rq, pos);
else
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
ret = lo_rw_simple(lo, rq, pos, READ);
}
return ret;
}
struct switch_request {
struct file *file;
struct completion wait;
};
static inline void loop_update_dio(struct loop_device *lo)
{
__loop_update_dio(lo, io_is_direct(lo->lo_backing_file) |
lo->use_dio);
}
/*
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
* Do the actual switch; called from the BIO completion routine
*/
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
static void do_loop_switch(struct loop_device *lo, struct switch_request *p)
{
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
struct file *file = p->file;
struct file *old_file = lo->lo_backing_file;
struct address_space *mapping;
[PATCH] optimise loop driver a bit Looks like locking can be optimised quite a lot. Increase lock widths slightly so lo_lock is taken fewer times per request. Also it was quite trivial to cover lo_pending with that lock, and remove the atomic requirement. This also makes memory ordering explicitly correct, which is nice (not that I particularly saw any mem ordering bugs). Test was reading 4 250MB files in parallel on ext2-on-tmpfs filesystem (1K block size, 4K page size). System is 2 socket Xeon with HT (4 thread). intel:/home/npiggin# umount /dev/loop0 ; mount /dev/loop0 /mnt/loop ; /usr/bin/time ./mtloop.sh Before: 0.24user 5.51system 0:02.84elapsed 202%CPU (0avgtext+0avgdata 0maxresident)k 0.19user 5.52system 0:02.88elapsed 198%CPU (0avgtext+0avgdata 0maxresident)k 0.19user 5.57system 0:02.89elapsed 198%CPU (0avgtext+0avgdata 0maxresident)k 0.22user 5.51system 0:02.90elapsed 197%CPU (0avgtext+0avgdata 0maxresident)k 0.19user 5.44system 0:02.91elapsed 193%CPU (0avgtext+0avgdata 0maxresident)k After: 0.07user 2.34system 0:01.68elapsed 143%CPU (0avgtext+0avgdata 0maxresident)k 0.06user 2.37system 0:01.68elapsed 144%CPU (0avgtext+0avgdata 0maxresident)k 0.06user 2.39system 0:01.68elapsed 145%CPU (0avgtext+0avgdata 0maxresident)k 0.06user 2.36system 0:01.68elapsed 144%CPU (0avgtext+0avgdata 0maxresident)k 0.06user 2.42system 0:01.68elapsed 147%CPU (0avgtext+0avgdata 0maxresident)k Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 14:09:06 +07:00
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
/* if no new file, only flush of queued bios requested */
if (!file)
return;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
mapping = file->f_mapping;
mapping_set_gfp_mask(old_file->f_mapping, lo->old_gfp_mask);
lo->lo_backing_file = file;
lo->lo_blocksize = S_ISBLK(mapping->host->i_mode) ?
mapping->host->i_bdev->bd_block_size : PAGE_SIZE;
lo->old_gfp_mask = mapping_gfp_mask(mapping);
mapping_set_gfp_mask(mapping, lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS));
loop_update_dio(lo);
}
/*
* loop_switch performs the hard work of switching a backing store.
* First it needs to flush existing IO, it does this by sending a magic
* BIO down the pipe. The completion of this BIO does the actual switch.
*/
static int loop_switch(struct loop_device *lo, struct file *file)
{
struct switch_request w;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
w.file = file;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
/* freeze queue and wait for completion of scheduled requests */
blk_mq_freeze_queue(lo->lo_queue);
/* do the switch action */
do_loop_switch(lo, &w);
/* unfreeze */
blk_mq_unfreeze_queue(lo->lo_queue);
return 0;
}
/*
* Helper to flush the IOs in loop, but keeping loop thread running
*/
static int loop_flush(struct loop_device *lo)
{
return loop_switch(lo, NULL);
}
static void loop_reread_partitions(struct loop_device *lo,
struct block_device *bdev)
{
int rc;
/*
* bd_mutex has been held already in release path, so don't
* acquire it if this function is called in such case.
*
* If the reread partition isn't from release path, lo_refcnt
* must be at least one and it can only become zero when the
* current holder is released.
*/
if (!atomic_read(&lo->lo_refcnt))
rc = __blkdev_reread_part(bdev);
else
rc = blkdev_reread_part(bdev);
if (rc)
pr_warn("%s: partition scan of loop%d (%s) failed (rc=%d)\n",
__func__, lo->lo_number, lo->lo_file_name, rc);
}
/*
* loop_change_fd switched the backing store of a loopback device to
* a new file. This is useful for operating system installers to free up
* the original file and in High Availability environments to switch to
* an alternative location for the content in case of server meltdown.
* This can only work if the loop device is used read-only, and if the
* new backing store is the same size and type as the old backing store.
*/
static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
unsigned int arg)
{
struct file *file, *old_file;
struct inode *inode;
int error;
error = -ENXIO;
if (lo->lo_state != Lo_bound)
goto out;
/* the loop device has to be read-only */
error = -EINVAL;
if (!(lo->lo_flags & LO_FLAGS_READ_ONLY))
goto out;
error = -EBADF;
file = fget(arg);
if (!file)
goto out;
inode = file->f_mapping->host;
old_file = lo->lo_backing_file;
error = -EINVAL;
if (!S_ISREG(inode->i_mode) && !S_ISBLK(inode->i_mode))
goto out_putf;
/* size of the new backing store needs to be the same */
if (get_loop_size(lo, file) != get_loop_size(lo, old_file))
goto out_putf;
/* and ... switch */
error = loop_switch(lo, file);
if (error)
goto out_putf;
fput(old_file);
2011-08-24 01:12:04 +07:00
if (lo->lo_flags & LO_FLAGS_PARTSCAN)
loop_reread_partitions(lo, bdev);
return 0;
out_putf:
fput(file);
out:
return error;
}
static inline int is_loop_device(struct file *file)
{
struct inode *i = file->f_mapping->host;
return i && S_ISBLK(i->i_mode) && MAJOR(i->i_rdev) == LOOP_MAJOR;
}
/* loop sysfs attributes */
static ssize_t loop_attr_show(struct device *dev, char *page,
ssize_t (*callback)(struct loop_device *, char *))
{
struct gendisk *disk = dev_to_disk(dev);
struct loop_device *lo = disk->private_data;
return callback(lo, page);
}
#define LOOP_ATTR_RO(_name) \
static ssize_t loop_attr_##_name##_show(struct loop_device *, char *); \
static ssize_t loop_attr_do_show_##_name(struct device *d, \
struct device_attribute *attr, char *b) \
{ \
return loop_attr_show(d, b, loop_attr_##_name##_show); \
} \
static struct device_attribute loop_attr_##_name = \
__ATTR(_name, S_IRUGO, loop_attr_do_show_##_name, NULL);
static ssize_t loop_attr_backing_file_show(struct loop_device *lo, char *buf)
{
ssize_t ret;
char *p = NULL;
spin_lock_irq(&lo->lo_lock);
if (lo->lo_backing_file)
p = file_path(lo->lo_backing_file, buf, PAGE_SIZE - 1);
spin_unlock_irq(&lo->lo_lock);
if (IS_ERR_OR_NULL(p))
ret = PTR_ERR(p);
else {
ret = strlen(p);
memmove(buf, p, ret);
buf[ret++] = '\n';
buf[ret] = 0;
}
return ret;
}
static ssize_t loop_attr_offset_show(struct loop_device *lo, char *buf)
{
return sprintf(buf, "%llu\n", (unsigned long long)lo->lo_offset);
}
static ssize_t loop_attr_sizelimit_show(struct loop_device *lo, char *buf)
{
return sprintf(buf, "%llu\n", (unsigned long long)lo->lo_sizelimit);
}
static ssize_t loop_attr_autoclear_show(struct loop_device *lo, char *buf)
{
int autoclear = (lo->lo_flags & LO_FLAGS_AUTOCLEAR);
return sprintf(buf, "%s\n", autoclear ? "1" : "0");
}
2011-08-24 01:12:04 +07:00
static ssize_t loop_attr_partscan_show(struct loop_device *lo, char *buf)
{
int partscan = (lo->lo_flags & LO_FLAGS_PARTSCAN);
return sprintf(buf, "%s\n", partscan ? "1" : "0");
}
static ssize_t loop_attr_dio_show(struct loop_device *lo, char *buf)
{
int dio = (lo->lo_flags & LO_FLAGS_DIRECT_IO);
return sprintf(buf, "%s\n", dio ? "1" : "0");
}
LOOP_ATTR_RO(backing_file);
LOOP_ATTR_RO(offset);
LOOP_ATTR_RO(sizelimit);
LOOP_ATTR_RO(autoclear);
2011-08-24 01:12:04 +07:00
LOOP_ATTR_RO(partscan);
LOOP_ATTR_RO(dio);
static struct attribute *loop_attrs[] = {
&loop_attr_backing_file.attr,
&loop_attr_offset.attr,
&loop_attr_sizelimit.attr,
&loop_attr_autoclear.attr,
2011-08-24 01:12:04 +07:00
&loop_attr_partscan.attr,
&loop_attr_dio.attr,
NULL,
};
static struct attribute_group loop_attribute_group = {
.name = "loop",
.attrs= loop_attrs,
};
static int loop_sysfs_init(struct loop_device *lo)
{
return sysfs_create_group(&disk_to_dev(lo->lo_disk)->kobj,
&loop_attribute_group);
}
static void loop_sysfs_exit(struct loop_device *lo)
{
sysfs_remove_group(&disk_to_dev(lo->lo_disk)->kobj,
&loop_attribute_group);
}
static void loop_config_discard(struct loop_device *lo)
{
struct file *file = lo->lo_backing_file;
struct inode *inode = file->f_mapping->host;
struct request_queue *q = lo->lo_queue;
/*
* We use punch hole to reclaim the free space used by the
* image a.k.a. discard. However we do not support discard if
* encryption is enabled, because it may give an attacker
* useful information.
*/
if ((!file->f_op->fallocate) ||
lo->lo_encrypt_key_size) {
q->limits.discard_granularity = 0;
q->limits.discard_alignment = 0;
blk_queue_max_discard_sectors(q, 0);
q->limits.discard_zeroes_data = 0;
queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q);
return;
}
q->limits.discard_granularity = inode->i_sb->s_blocksize;
q->limits.discard_alignment = 0;
blk_queue_max_discard_sectors(q, UINT_MAX >> 9);
q->limits.discard_zeroes_data = 1;
queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
}
static void loop_unprepare_queue(struct loop_device *lo)
{
flush_kthread_worker(&lo->worker);
kthread_stop(lo->worker_task);
}
static int loop_prepare_queue(struct loop_device *lo)
{
init_kthread_worker(&lo->worker);
lo->worker_task = kthread_run(kthread_worker_fn,
&lo->worker, "loop%d", lo->lo_number);
if (IS_ERR(lo->worker_task))
return -ENOMEM;
set_user_nice(lo->worker_task, MIN_NICE);
return 0;
}
static int loop_set_fd(struct loop_device *lo, fmode_t mode,
struct block_device *bdev, unsigned int arg)
{
struct file *file, *f;
struct inode *inode;
struct address_space *mapping;
unsigned lo_blocksize;
int lo_flags = 0;
int error;
loff_t size;
/* This is safe, since we have a reference from open(). */
__module_get(THIS_MODULE);
error = -EBADF;
file = fget(arg);
if (!file)
goto out;
error = -EBUSY;
if (lo->lo_state != Lo_unbound)
goto out_putf;
/* Avoid recursion */
f = file;
while (is_loop_device(f)) {
struct loop_device *l;
if (f->f_mapping->host->i_bdev == bdev)
goto out_putf;
l = f->f_mapping->host->i_bdev->bd_disk->private_data;
if (l->lo_state == Lo_unbound) {
error = -EINVAL;
goto out_putf;
}
f = l->lo_backing_file;
}
mapping = file->f_mapping;
inode = mapping->host;
error = -EINVAL;
if (!S_ISREG(inode->i_mode) && !S_ISBLK(inode->i_mode))
goto out_putf;
if (!(file->f_mode & FMODE_WRITE) || !(mode & FMODE_WRITE) ||
!file->f_op->write_iter)
lo_flags |= LO_FLAGS_READ_ONLY;
lo_blocksize = S_ISBLK(inode->i_mode) ?
inode->i_bdev->bd_block_size : PAGE_SIZE;
error = -EFBIG;
size = get_loop_size(lo, file);
if ((loff_t)(sector_t)size != size)
goto out_putf;
error = loop_prepare_queue(lo);
if (error)
goto out_putf;
error = 0;
set_device_ro(bdev, (lo_flags & LO_FLAGS_READ_ONLY) != 0);
lo->use_dio = false;
lo->lo_blocksize = lo_blocksize;
lo->lo_device = bdev;
lo->lo_flags = lo_flags;
lo->lo_backing_file = file;
lo->transfer = NULL;
lo->ioctl = NULL;
lo->lo_sizelimit = 0;
lo->old_gfp_mask = mapping_gfp_mask(mapping);
mapping_set_gfp_mask(mapping, lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS));
if (!(lo_flags & LO_FLAGS_READ_ONLY) && file->f_op->fsync)
blk_queue_write_cache(lo->lo_queue, true, false);
loop_update_dio(lo);
set_capacity(lo->lo_disk, size);
bd_set_size(bdev, size << 9);
loop_sysfs_init(lo);
/* let user-space know about the new size */
kobject_uevent(&disk_to_dev(bdev->bd_disk)->kobj, KOBJ_CHANGE);
set_blocksize(bdev, lo_blocksize);
lo->lo_state = Lo_bound;
2011-08-24 01:12:04 +07:00
if (part_shift)
lo->lo_flags |= LO_FLAGS_PARTSCAN;
if (lo->lo_flags & LO_FLAGS_PARTSCAN)
loop_reread_partitions(lo, bdev);
loop: prevent bdev freeing while device in use struct block_device lifecycle is defined by its inode (see fs/block_dev.c) - block_device allocated first time we access /dev/loopXX and deallocated on bdev_destroy_inode. When we create the device "losetup /dev/loopXX afile" we want that block_device stay alive until we destroy the loop device with "losetup -d". But because we do not hold /dev/loopXX inode its counter goes 0, and inode/bdev can be destroyed at any moment. Usually it happens at memory pressure or when user drops inode cache (like in the test below). When later in loop_clr_fd() we want to use bdev we have use-after-free error with following stack: BUG: unable to handle kernel NULL pointer dereference at 0000000000000280 bd_set_size+0x10/0xa0 loop_clr_fd+0x1f8/0x420 [loop] lo_ioctl+0x200/0x7e0 [loop] lo_compat_ioctl+0x47/0xe0 [loop] compat_blkdev_ioctl+0x341/0x1290 do_filp_open+0x42/0xa0 compat_sys_ioctl+0xc1/0xf20 do_sys_open+0x16e/0x1d0 sysenter_dispatch+0x7/0x1a To prevent use-after-free we need to grab the device in loop_set_fd() and put it later in loop_clr_fd(). The issue is reprodusible on current Linus head and v3.3. Here is the test: dd if=/dev/zero of=loop.file bs=1M count=1 while [ true ]; do losetup /dev/loop0 loop.file echo 2 > /proc/sys/vm/drop_caches losetup -d /dev/loop0 done [ Doing bdgrab/bput in loop_set_fd/loop_clr_fd is safe, because every time we call loop_set_fd() we check that loop_device->lo_state is Lo_unbound and set it to Lo_bound If somebody will try to set_fd again it will get EBUSY. And if we try to loop_clr_fd() on unbound loop device we'll get ENXIO. loop_set_fd/loop_clr_fd (and any other loop ioctl) is called under loop_device->lo_ctl_mutex. ] Signed-off-by: Anatol Pomozov <anatol.pomozov@gmail.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-01 23:47:56 +07:00
/* Grab the block_device to prevent its destruction after we
* put /dev/loopXX inode. Later in loop_clr_fd() we bdput(bdev).
*/
bdgrab(bdev);
return 0;
out_putf:
fput(file);
out:
/* This is safe: open() is still holding a reference. */
module_put(THIS_MODULE);
return error;
}
static int
loop_release_xfer(struct loop_device *lo)
{
int err = 0;
struct loop_func_table *xfer = lo->lo_encryption;
if (xfer) {
if (xfer->release)
err = xfer->release(lo);
lo->transfer = NULL;
lo->lo_encryption = NULL;
module_put(xfer->owner);
}
return err;
}
static int
loop_init_xfer(struct loop_device *lo, struct loop_func_table *xfer,
const struct loop_info64 *i)
{
int err = 0;
if (xfer) {
struct module *owner = xfer->owner;
if (!try_module_get(owner))
return -EINVAL;
if (xfer->init)
err = xfer->init(lo, i);
if (err)
module_put(owner);
else
lo->lo_encryption = xfer;
}
return err;
}
static int loop_clr_fd(struct loop_device *lo)
{
struct file *filp = lo->lo_backing_file;
gfp_t gfp = lo->old_gfp_mask;
struct block_device *bdev = lo->lo_device;
if (lo->lo_state != Lo_bound)
return -ENXIO;
loop: Make explicit loop device destruction lazy xfstests has always had random failures of tests due to loop devices failing to be torn down and hence leaving filesytems that cannot be unmounted. This causes test runs to immediately stop. Over the past 6 or 7 years we've added hacks like explicit unmount -d commands for loop mounts, losetup -d after unmount -d fails, etc, but still the problems persist. Recently, the frequency of loop related failures increased again to the point that xfstests 259 will reliably fail with a stray loop device that was not torn down. That is despite the fact the test is above as simple as it gets - loop 5 or 6 times running mkfs.xfs with different paramters: lofile=$(losetup -f) losetup $lofile "$testfile" "$MKFS_XFS_PROG" -b size=512 $lofile >/dev/null || echo "mkfs failed!" sync losetup -d $lofile And losteup -d $lofile is failing with EBUSY on 1-3 of these loops every time the test is run. Turns out that blkid is running simultaneously with losetup -d, and so it sees an elevated reference count and returns EBUSY. But why is blkid running? It's obvious, isn't it? udev has decided to try and find out what is on the block device as a result of a creation notification. And it is racing with mkfs, so might still be scanning the device when mkfs finishes and we try to tear it down. So, make losetup -d force autoremove behaviour. That is, when the last reference goes away, tear down the device. xfstests wants it *gone*, not causing random teardown failures when we know that all the operations the tests have specifically run on the device have completed and are no longer referencing the loop device. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-09-28 15:42:23 +07:00
/*
* If we've explicitly asked to tear down the loop device,
* and it has an elevated reference count, set it for auto-teardown when
* the last reference goes away. This stops $!~#$@ udev from
* preventing teardown because it decided that it needs to run blkid on
* the loopback device whenever they appear. xfstests is notorious for
* failing tests because blkid via udev races with a losetup
* <dev>/do something like mkfs/losetup -d <dev> causing the losetup -d
* command to fail with EBUSY.
*/
if (atomic_read(&lo->lo_refcnt) > 1) {
loop: Make explicit loop device destruction lazy xfstests has always had random failures of tests due to loop devices failing to be torn down and hence leaving filesytems that cannot be unmounted. This causes test runs to immediately stop. Over the past 6 or 7 years we've added hacks like explicit unmount -d commands for loop mounts, losetup -d after unmount -d fails, etc, but still the problems persist. Recently, the frequency of loop related failures increased again to the point that xfstests 259 will reliably fail with a stray loop device that was not torn down. That is despite the fact the test is above as simple as it gets - loop 5 or 6 times running mkfs.xfs with different paramters: lofile=$(losetup -f) losetup $lofile "$testfile" "$MKFS_XFS_PROG" -b size=512 $lofile >/dev/null || echo "mkfs failed!" sync losetup -d $lofile And losteup -d $lofile is failing with EBUSY on 1-3 of these loops every time the test is run. Turns out that blkid is running simultaneously with losetup -d, and so it sees an elevated reference count and returns EBUSY. But why is blkid running? It's obvious, isn't it? udev has decided to try and find out what is on the block device as a result of a creation notification. And it is racing with mkfs, so might still be scanning the device when mkfs finishes and we try to tear it down. So, make losetup -d force autoremove behaviour. That is, when the last reference goes away, tear down the device. xfstests wants it *gone*, not causing random teardown failures when we know that all the operations the tests have specifically run on the device have completed and are no longer referencing the loop device. Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2012-09-28 15:42:23 +07:00
lo->lo_flags |= LO_FLAGS_AUTOCLEAR;
mutex_unlock(&lo->lo_ctl_mutex);
return 0;
}
if (filp == NULL)
return -EINVAL;
/* freeze request queue during the transition */
blk_mq_freeze_queue(lo->lo_queue);
spin_lock_irq(&lo->lo_lock);
lo->lo_state = Lo_rundown;
lo->lo_backing_file = NULL;
spin_unlock_irq(&lo->lo_lock);
loop_release_xfer(lo);
lo->transfer = NULL;
lo->ioctl = NULL;
lo->lo_device = NULL;
lo->lo_encryption = NULL;
lo->lo_offset = 0;
lo->lo_sizelimit = 0;
lo->lo_encrypt_key_size = 0;
memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
memset(lo->lo_file_name, 0, LO_NAME_SIZE);
loop: prevent bdev freeing while device in use struct block_device lifecycle is defined by its inode (see fs/block_dev.c) - block_device allocated first time we access /dev/loopXX and deallocated on bdev_destroy_inode. When we create the device "losetup /dev/loopXX afile" we want that block_device stay alive until we destroy the loop device with "losetup -d". But because we do not hold /dev/loopXX inode its counter goes 0, and inode/bdev can be destroyed at any moment. Usually it happens at memory pressure or when user drops inode cache (like in the test below). When later in loop_clr_fd() we want to use bdev we have use-after-free error with following stack: BUG: unable to handle kernel NULL pointer dereference at 0000000000000280 bd_set_size+0x10/0xa0 loop_clr_fd+0x1f8/0x420 [loop] lo_ioctl+0x200/0x7e0 [loop] lo_compat_ioctl+0x47/0xe0 [loop] compat_blkdev_ioctl+0x341/0x1290 do_filp_open+0x42/0xa0 compat_sys_ioctl+0xc1/0xf20 do_sys_open+0x16e/0x1d0 sysenter_dispatch+0x7/0x1a To prevent use-after-free we need to grab the device in loop_set_fd() and put it later in loop_clr_fd(). The issue is reprodusible on current Linus head and v3.3. Here is the test: dd if=/dev/zero of=loop.file bs=1M count=1 while [ true ]; do losetup /dev/loop0 loop.file echo 2 > /proc/sys/vm/drop_caches losetup -d /dev/loop0 done [ Doing bdgrab/bput in loop_set_fd/loop_clr_fd is safe, because every time we call loop_set_fd() we check that loop_device->lo_state is Lo_unbound and set it to Lo_bound If somebody will try to set_fd again it will get EBUSY. And if we try to loop_clr_fd() on unbound loop device we'll get ENXIO. loop_set_fd/loop_clr_fd (and any other loop ioctl) is called under loop_device->lo_ctl_mutex. ] Signed-off-by: Anatol Pomozov <anatol.pomozov@gmail.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-01 23:47:56 +07:00
if (bdev) {
bdput(bdev);
invalidate_bdev(bdev);
loop: prevent bdev freeing while device in use struct block_device lifecycle is defined by its inode (see fs/block_dev.c) - block_device allocated first time we access /dev/loopXX and deallocated on bdev_destroy_inode. When we create the device "losetup /dev/loopXX afile" we want that block_device stay alive until we destroy the loop device with "losetup -d". But because we do not hold /dev/loopXX inode its counter goes 0, and inode/bdev can be destroyed at any moment. Usually it happens at memory pressure or when user drops inode cache (like in the test below). When later in loop_clr_fd() we want to use bdev we have use-after-free error with following stack: BUG: unable to handle kernel NULL pointer dereference at 0000000000000280 bd_set_size+0x10/0xa0 loop_clr_fd+0x1f8/0x420 [loop] lo_ioctl+0x200/0x7e0 [loop] lo_compat_ioctl+0x47/0xe0 [loop] compat_blkdev_ioctl+0x341/0x1290 do_filp_open+0x42/0xa0 compat_sys_ioctl+0xc1/0xf20 do_sys_open+0x16e/0x1d0 sysenter_dispatch+0x7/0x1a To prevent use-after-free we need to grab the device in loop_set_fd() and put it later in loop_clr_fd(). The issue is reprodusible on current Linus head and v3.3. Here is the test: dd if=/dev/zero of=loop.file bs=1M count=1 while [ true ]; do losetup /dev/loop0 loop.file echo 2 > /proc/sys/vm/drop_caches losetup -d /dev/loop0 done [ Doing bdgrab/bput in loop_set_fd/loop_clr_fd is safe, because every time we call loop_set_fd() we check that loop_device->lo_state is Lo_unbound and set it to Lo_bound If somebody will try to set_fd again it will get EBUSY. And if we try to loop_clr_fd() on unbound loop device we'll get ENXIO. loop_set_fd/loop_clr_fd (and any other loop ioctl) is called under loop_device->lo_ctl_mutex. ] Signed-off-by: Anatol Pomozov <anatol.pomozov@gmail.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-01 23:47:56 +07:00
}
set_capacity(lo->lo_disk, 0);
loop_sysfs_exit(lo);
if (bdev) {
bd_set_size(bdev, 0);
/* let user-space know about this change */
kobject_uevent(&disk_to_dev(bdev->bd_disk)->kobj, KOBJ_CHANGE);
}
mapping_set_gfp_mask(filp->f_mapping, gfp);
lo->lo_state = Lo_unbound;
/* This is safe: open() is still holding a reference. */
module_put(THIS_MODULE);
blk_mq_unfreeze_queue(lo->lo_queue);
if (lo->lo_flags & LO_FLAGS_PARTSCAN && bdev)
loop_reread_partitions(lo, bdev);
2011-08-24 01:12:04 +07:00
lo->lo_flags = 0;
if (!part_shift)
lo->lo_disk->flags |= GENHD_FL_NO_PART_SCAN;
loop_unprepare_queue(lo);
loop: fix circular locking in loop_clr_fd() With CONFIG_PROVE_LOCKING enabled $ losetup /dev/loop0 file $ losetup -o 32256 /dev/loop1 /dev/loop0 $ losetup -d /dev/loop1 $ losetup -d /dev/loop0 triggers a [ INFO: possible circular locking dependency detected ] I think this warning is a false positive. Open/close on a loop device acquires bd_mutex of the device before acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of a device, if backing file is a device and this is the last reference to the file being dropped . But it is guaranteed that it is impossible to have a circular list of backing devices.(say loop2->loop1->loop0->loop2 is not possible), which guarantees that this can never deadlock. So this warning should be suppressed. It is very difficult to annotate lockdep not to warn here in the correct way. A simple way to silence lockdep could be to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some other real bugs. @@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode, struct loop_device *lo = bdev->bd_disk->private_data; int err; - mutex_lock(&lo->lo_ctl_mutex); + mutex_lock_nested(&lo->lo_ctl_mutex, 1); switch (cmd) { case LOOP_SET_FD: err = loop_set_fd(lo, mode, bdev, arg); Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be a better solution. Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex held, so no lockdep annotation is required. If you do not like the special handling of the lo_ctl_mutex just for the LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside each of the individual ioctl handlers and I could send you another patch. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 18:33:41 +07:00
mutex_unlock(&lo->lo_ctl_mutex);
/*
* Need not hold lo_ctl_mutex to fput backing file.
* Calling fput holding lo_ctl_mutex triggers a circular
* lock dependency possibility warning as fput can take
* bd_mutex which is usually taken before lo_ctl_mutex.
*/
fput(filp);
return 0;
}
static int
loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
{
int err;
struct loop_func_table *xfer;
kuid_t uid = current_uid();
if (lo->lo_encrypt_key_size &&
!uid_eq(lo->lo_key_owner, uid) &&
!capable(CAP_SYS_ADMIN))
return -EPERM;
if (lo->lo_state != Lo_bound)
return -ENXIO;
if ((unsigned int) info->lo_encrypt_key_size > LO_KEY_SIZE)
return -EINVAL;
err = loop_release_xfer(lo);
if (err)
return err;
if (info->lo_encrypt_type) {
unsigned int type = info->lo_encrypt_type;
if (type >= MAX_LO_CRYPT)
return -EINVAL;
xfer = xfer_funcs[type];
if (xfer == NULL)
return -EINVAL;
} else
xfer = NULL;
err = loop_init_xfer(lo, xfer, info);
if (err)
return err;
if (lo->lo_offset != info->lo_offset ||
lo->lo_sizelimit != info->lo_sizelimit)
if (figure_loop_size(lo, info->lo_offset, info->lo_sizelimit))
return -EFBIG;
loop_config_discard(lo);
memcpy(lo->lo_file_name, info->lo_file_name, LO_NAME_SIZE);
memcpy(lo->lo_crypt_name, info->lo_crypt_name, LO_NAME_SIZE);
lo->lo_file_name[LO_NAME_SIZE-1] = 0;
lo->lo_crypt_name[LO_NAME_SIZE-1] = 0;
if (!xfer)
xfer = &none_funcs;
lo->transfer = xfer->transfer;
lo->ioctl = xfer->ioctl;
if ((lo->lo_flags & LO_FLAGS_AUTOCLEAR) !=
(info->lo_flags & LO_FLAGS_AUTOCLEAR))
lo->lo_flags ^= LO_FLAGS_AUTOCLEAR;
2011-08-24 01:12:04 +07:00
if ((info->lo_flags & LO_FLAGS_PARTSCAN) &&
!(lo->lo_flags & LO_FLAGS_PARTSCAN)) {
lo->lo_flags |= LO_FLAGS_PARTSCAN;
lo->lo_disk->flags &= ~GENHD_FL_NO_PART_SCAN;
loop_reread_partitions(lo, lo->lo_device);
2011-08-24 01:12:04 +07:00
}
lo->lo_encrypt_key_size = info->lo_encrypt_key_size;
lo->lo_init[0] = info->lo_init[0];
lo->lo_init[1] = info->lo_init[1];
if (info->lo_encrypt_key_size) {
memcpy(lo->lo_encrypt_key, info->lo_encrypt_key,
info->lo_encrypt_key_size);
lo->lo_key_owner = uid;
}
/* update dio if lo_offset or transfer is changed */
__loop_update_dio(lo, lo->use_dio);
return 0;
}
static int
loop_get_status(struct loop_device *lo, struct loop_info64 *info)
{
struct file *file = lo->lo_backing_file;
struct kstat stat;
int error;
if (lo->lo_state != Lo_bound)
return -ENXIO;
error = vfs_getattr(&file->f_path, &stat);
if (error)
return error;
memset(info, 0, sizeof(*info));
info->lo_number = lo->lo_number;
info->lo_device = huge_encode_dev(stat.dev);
info->lo_inode = stat.ino;
info->lo_rdevice = huge_encode_dev(lo->lo_device ? stat.rdev : stat.dev);
info->lo_offset = lo->lo_offset;
info->lo_sizelimit = lo->lo_sizelimit;
info->lo_flags = lo->lo_flags;
memcpy(info->lo_file_name, lo->lo_file_name, LO_NAME_SIZE);
memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
info->lo_encrypt_type =
lo->lo_encryption ? lo->lo_encryption->number : 0;
if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
lo->lo_encrypt_key_size);
}
return 0;
}
static void
loop_info64_from_old(const struct loop_info *info, struct loop_info64 *info64)
{
memset(info64, 0, sizeof(*info64));
info64->lo_number = info->lo_number;
info64->lo_device = info->lo_device;
info64->lo_inode = info->lo_inode;
info64->lo_rdevice = info->lo_rdevice;
info64->lo_offset = info->lo_offset;
info64->lo_sizelimit = 0;
info64->lo_encrypt_type = info->lo_encrypt_type;
info64->lo_encrypt_key_size = info->lo_encrypt_key_size;
info64->lo_flags = info->lo_flags;
info64->lo_init[0] = info->lo_init[0];
info64->lo_init[1] = info->lo_init[1];
if (info->lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
memcpy(info64->lo_crypt_name, info->lo_name, LO_NAME_SIZE);
else
memcpy(info64->lo_file_name, info->lo_name, LO_NAME_SIZE);
memcpy(info64->lo_encrypt_key, info->lo_encrypt_key, LO_KEY_SIZE);
}
static int
loop_info64_to_old(const struct loop_info64 *info64, struct loop_info *info)
{
memset(info, 0, sizeof(*info));
info->lo_number = info64->lo_number;
info->lo_device = info64->lo_device;
info->lo_inode = info64->lo_inode;
info->lo_rdevice = info64->lo_rdevice;
info->lo_offset = info64->lo_offset;
info->lo_encrypt_type = info64->lo_encrypt_type;
info->lo_encrypt_key_size = info64->lo_encrypt_key_size;
info->lo_flags = info64->lo_flags;
info->lo_init[0] = info64->lo_init[0];
info->lo_init[1] = info64->lo_init[1];
if (info->lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
memcpy(info->lo_name, info64->lo_crypt_name, LO_NAME_SIZE);
else
memcpy(info->lo_name, info64->lo_file_name, LO_NAME_SIZE);
memcpy(info->lo_encrypt_key, info64->lo_encrypt_key, LO_KEY_SIZE);
/* error in case values were truncated */
if (info->lo_device != info64->lo_device ||
info->lo_rdevice != info64->lo_rdevice ||
info->lo_inode != info64->lo_inode ||
info->lo_offset != info64->lo_offset)
return -EOVERFLOW;
return 0;
}
static int
loop_set_status_old(struct loop_device *lo, const struct loop_info __user *arg)
{
struct loop_info info;
struct loop_info64 info64;
if (copy_from_user(&info, arg, sizeof (struct loop_info)))
return -EFAULT;
loop_info64_from_old(&info, &info64);
return loop_set_status(lo, &info64);
}
static int
loop_set_status64(struct loop_device *lo, const struct loop_info64 __user *arg)
{
struct loop_info64 info64;
if (copy_from_user(&info64, arg, sizeof (struct loop_info64)))
return -EFAULT;
return loop_set_status(lo, &info64);
}
static int
loop_get_status_old(struct loop_device *lo, struct loop_info __user *arg) {
struct loop_info info;
struct loop_info64 info64;
int err = 0;
if (!arg)
err = -EINVAL;
if (!err)
err = loop_get_status(lo, &info64);
if (!err)
err = loop_info64_to_old(&info64, &info);
if (!err && copy_to_user(arg, &info, sizeof(info)))
err = -EFAULT;
return err;
}
static int
loop_get_status64(struct loop_device *lo, struct loop_info64 __user *arg) {
struct loop_info64 info64;
int err = 0;
if (!arg)
err = -EINVAL;
if (!err)
err = loop_get_status(lo, &info64);
if (!err && copy_to_user(arg, &info64, sizeof(info64)))
err = -EFAULT;
return err;
}
loop: add ioctl to resize a loop device Add the ability to 'resize' the loop device on the fly. One practical application is a loop file with XFS filesystem, already mounted: You can easily enlarge the file (append some bytes) and then call ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the new size and you can use xfs_growfs later on, which will allow you to use full capacity of the loop file without the need to unmount. Test app: #include <linux/fs.h> #include <linux/loop.h> #include <sys/ioctl.h> #include <sys/stat.h> #include <sys/types.h> #include <assert.h> #include <errno.h> #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _GNU_SOURCE #include <getopt.h> char *me; void usage(FILE *f) { fprintf(f, "%s [options] loop_dev [backend_file]\n" "-s, --set new_size_in_bytes\n" "\twhen backend_file is given, " "it will be expanded too while keeping the original contents\n", me); } struct option opts[] = { { .name = "set", .has_arg = 1, .flag = NULL, .val = 's' }, { .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' } }; void err_size(char *name, __u64 old) { fprintf(stderr, "size must be larger than current %s (%llu)\n", name, old); } int main(int argc, char *argv[]) { int fd, err, c, i, bfd; ssize_t ssz; size_t sz; __u64 old, new, append; char a[BUFSIZ]; struct stat st; FILE *out; char *backend, *dev; err = EINVAL; out = stderr; me = argv[0]; new = 0; while ((c = getopt_long(argc, argv, "s:h", opts, &i)) != -1) { switch (c) { case 's': errno = 0; new = strtoull(optarg, NULL, 0); if (errno) { err = errno; perror(argv[i]); goto out; } break; case 'h': err = 0; out = stdout; goto err; default: perror(argv[i]); goto err; } } if (optind < argc) dev = argv[optind++]; else goto err; fd = open(dev, O_RDONLY); if (fd < 0) { err = errno; perror(dev); goto out; } err = ioctl(fd, BLKGETSIZE64, &old); if (err) { err = errno; perror("ioctl BLKGETSIZE64"); goto out; } if (!new) { printf("%llu\n", old); goto out; } if (new < old) { err = EINVAL; err_size(dev, old); goto out; } if (optind < argc) { backend = argv[optind++]; bfd = open(backend, O_WRONLY|O_APPEND); if (bfd < 0) { err = errno; perror(backend); goto out; } err = fstat(bfd, &st); if (err) { err = errno; perror(backend); goto out; } if (new < st.st_size) { err = EINVAL; err_size(backend, st.st_size); goto out; } append = new - st.st_size; sz = sizeof(a); while (append > 0) { if (append < sz) sz = append; ssz = write(bfd, a, sz); if (ssz != sz) { err = errno; perror(backend); goto out; } append -= sz; } err = fsync(bfd); if (err) { err = errno; perror(backend); goto out; } } err = ioctl(fd, LOOP_SET_CAPACITY, new); if (err) { err = errno; perror("ioctl LOOP_SET_CAPACITY"); } goto out; err: usage(out); out: return err; } Signed-off-by: J. R. Okajima <hooanon05@yahoo.co.jp> Signed-off-by: Tomas Matejicek <tomas@slax.org> Cc: <util-linux-ng@vger.kernel.org> Cc: Karel Zak <kzak@redhat.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: <linux-api@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-01 05:23:43 +07:00
static int loop_set_capacity(struct loop_device *lo, struct block_device *bdev)
{
if (unlikely(lo->lo_state != Lo_bound))
return -ENXIO;
loop: add ioctl to resize a loop device Add the ability to 'resize' the loop device on the fly. One practical application is a loop file with XFS filesystem, already mounted: You can easily enlarge the file (append some bytes) and then call ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the new size and you can use xfs_growfs later on, which will allow you to use full capacity of the loop file without the need to unmount. Test app: #include <linux/fs.h> #include <linux/loop.h> #include <sys/ioctl.h> #include <sys/stat.h> #include <sys/types.h> #include <assert.h> #include <errno.h> #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _GNU_SOURCE #include <getopt.h> char *me; void usage(FILE *f) { fprintf(f, "%s [options] loop_dev [backend_file]\n" "-s, --set new_size_in_bytes\n" "\twhen backend_file is given, " "it will be expanded too while keeping the original contents\n", me); } struct option opts[] = { { .name = "set", .has_arg = 1, .flag = NULL, .val = 's' }, { .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' } }; void err_size(char *name, __u64 old) { fprintf(stderr, "size must be larger than current %s (%llu)\n", name, old); } int main(int argc, char *argv[]) { int fd, err, c, i, bfd; ssize_t ssz; size_t sz; __u64 old, new, append; char a[BUFSIZ]; struct stat st; FILE *out; char *backend, *dev; err = EINVAL; out = stderr; me = argv[0]; new = 0; while ((c = getopt_long(argc, argv, "s:h", opts, &i)) != -1) { switch (c) { case 's': errno = 0; new = strtoull(optarg, NULL, 0); if (errno) { err = errno; perror(argv[i]); goto out; } break; case 'h': err = 0; out = stdout; goto err; default: perror(argv[i]); goto err; } } if (optind < argc) dev = argv[optind++]; else goto err; fd = open(dev, O_RDONLY); if (fd < 0) { err = errno; perror(dev); goto out; } err = ioctl(fd, BLKGETSIZE64, &old); if (err) { err = errno; perror("ioctl BLKGETSIZE64"); goto out; } if (!new) { printf("%llu\n", old); goto out; } if (new < old) { err = EINVAL; err_size(dev, old); goto out; } if (optind < argc) { backend = argv[optind++]; bfd = open(backend, O_WRONLY|O_APPEND); if (bfd < 0) { err = errno; perror(backend); goto out; } err = fstat(bfd, &st); if (err) { err = errno; perror(backend); goto out; } if (new < st.st_size) { err = EINVAL; err_size(backend, st.st_size); goto out; } append = new - st.st_size; sz = sizeof(a); while (append > 0) { if (append < sz) sz = append; ssz = write(bfd, a, sz); if (ssz != sz) { err = errno; perror(backend); goto out; } append -= sz; } err = fsync(bfd); if (err) { err = errno; perror(backend); goto out; } } err = ioctl(fd, LOOP_SET_CAPACITY, new); if (err) { err = errno; perror("ioctl LOOP_SET_CAPACITY"); } goto out; err: usage(out); out: return err; } Signed-off-by: J. R. Okajima <hooanon05@yahoo.co.jp> Signed-off-by: Tomas Matejicek <tomas@slax.org> Cc: <util-linux-ng@vger.kernel.org> Cc: Karel Zak <kzak@redhat.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: <linux-api@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-01 05:23:43 +07:00
return figure_loop_size(lo, lo->lo_offset, lo->lo_sizelimit);
loop: add ioctl to resize a loop device Add the ability to 'resize' the loop device on the fly. One practical application is a loop file with XFS filesystem, already mounted: You can easily enlarge the file (append some bytes) and then call ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the new size and you can use xfs_growfs later on, which will allow you to use full capacity of the loop file without the need to unmount. Test app: #include <linux/fs.h> #include <linux/loop.h> #include <sys/ioctl.h> #include <sys/stat.h> #include <sys/types.h> #include <assert.h> #include <errno.h> #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _GNU_SOURCE #include <getopt.h> char *me; void usage(FILE *f) { fprintf(f, "%s [options] loop_dev [backend_file]\n" "-s, --set new_size_in_bytes\n" "\twhen backend_file is given, " "it will be expanded too while keeping the original contents\n", me); } struct option opts[] = { { .name = "set", .has_arg = 1, .flag = NULL, .val = 's' }, { .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' } }; void err_size(char *name, __u64 old) { fprintf(stderr, "size must be larger than current %s (%llu)\n", name, old); } int main(int argc, char *argv[]) { int fd, err, c, i, bfd; ssize_t ssz; size_t sz; __u64 old, new, append; char a[BUFSIZ]; struct stat st; FILE *out; char *backend, *dev; err = EINVAL; out = stderr; me = argv[0]; new = 0; while ((c = getopt_long(argc, argv, "s:h", opts, &i)) != -1) { switch (c) { case 's': errno = 0; new = strtoull(optarg, NULL, 0); if (errno) { err = errno; perror(argv[i]); goto out; } break; case 'h': err = 0; out = stdout; goto err; default: perror(argv[i]); goto err; } } if (optind < argc) dev = argv[optind++]; else goto err; fd = open(dev, O_RDONLY); if (fd < 0) { err = errno; perror(dev); goto out; } err = ioctl(fd, BLKGETSIZE64, &old); if (err) { err = errno; perror("ioctl BLKGETSIZE64"); goto out; } if (!new) { printf("%llu\n", old); goto out; } if (new < old) { err = EINVAL; err_size(dev, old); goto out; } if (optind < argc) { backend = argv[optind++]; bfd = open(backend, O_WRONLY|O_APPEND); if (bfd < 0) { err = errno; perror(backend); goto out; } err = fstat(bfd, &st); if (err) { err = errno; perror(backend); goto out; } if (new < st.st_size) { err = EINVAL; err_size(backend, st.st_size); goto out; } append = new - st.st_size; sz = sizeof(a); while (append > 0) { if (append < sz) sz = append; ssz = write(bfd, a, sz); if (ssz != sz) { err = errno; perror(backend); goto out; } append -= sz; } err = fsync(bfd); if (err) { err = errno; perror(backend); goto out; } } err = ioctl(fd, LOOP_SET_CAPACITY, new); if (err) { err = errno; perror("ioctl LOOP_SET_CAPACITY"); } goto out; err: usage(out); out: return err; } Signed-off-by: J. R. Okajima <hooanon05@yahoo.co.jp> Signed-off-by: Tomas Matejicek <tomas@slax.org> Cc: <util-linux-ng@vger.kernel.org> Cc: Karel Zak <kzak@redhat.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: <linux-api@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-01 05:23:43 +07:00
}
static int loop_set_dio(struct loop_device *lo, unsigned long arg)
{
int error = -ENXIO;
if (lo->lo_state != Lo_bound)
goto out;
__loop_update_dio(lo, !!arg);
if (lo->use_dio == !!arg)
return 0;
error = -EINVAL;
out:
return error;
}
static int lo_ioctl(struct block_device *bdev, fmode_t mode,
unsigned int cmd, unsigned long arg)
{
struct loop_device *lo = bdev->bd_disk->private_data;
int err;
loop: fix circular locking in loop_clr_fd() With CONFIG_PROVE_LOCKING enabled $ losetup /dev/loop0 file $ losetup -o 32256 /dev/loop1 /dev/loop0 $ losetup -d /dev/loop1 $ losetup -d /dev/loop0 triggers a [ INFO: possible circular locking dependency detected ] I think this warning is a false positive. Open/close on a loop device acquires bd_mutex of the device before acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of a device, if backing file is a device and this is the last reference to the file being dropped . But it is guaranteed that it is impossible to have a circular list of backing devices.(say loop2->loop1->loop0->loop2 is not possible), which guarantees that this can never deadlock. So this warning should be suppressed. It is very difficult to annotate lockdep not to warn here in the correct way. A simple way to silence lockdep could be to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some other real bugs. @@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode, struct loop_device *lo = bdev->bd_disk->private_data; int err; - mutex_lock(&lo->lo_ctl_mutex); + mutex_lock_nested(&lo->lo_ctl_mutex, 1); switch (cmd) { case LOOP_SET_FD: err = loop_set_fd(lo, mode, bdev, arg); Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be a better solution. Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex held, so no lockdep annotation is required. If you do not like the special handling of the lo_ctl_mutex just for the LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside each of the individual ioctl handlers and I could send you another patch. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 18:33:41 +07:00
mutex_lock_nested(&lo->lo_ctl_mutex, 1);
switch (cmd) {
case LOOP_SET_FD:
err = loop_set_fd(lo, mode, bdev, arg);
break;
case LOOP_CHANGE_FD:
err = loop_change_fd(lo, bdev, arg);
break;
case LOOP_CLR_FD:
loop: fix circular locking in loop_clr_fd() With CONFIG_PROVE_LOCKING enabled $ losetup /dev/loop0 file $ losetup -o 32256 /dev/loop1 /dev/loop0 $ losetup -d /dev/loop1 $ losetup -d /dev/loop0 triggers a [ INFO: possible circular locking dependency detected ] I think this warning is a false positive. Open/close on a loop device acquires bd_mutex of the device before acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of a device, if backing file is a device and this is the last reference to the file being dropped . But it is guaranteed that it is impossible to have a circular list of backing devices.(say loop2->loop1->loop0->loop2 is not possible), which guarantees that this can never deadlock. So this warning should be suppressed. It is very difficult to annotate lockdep not to warn here in the correct way. A simple way to silence lockdep could be to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some other real bugs. @@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode, struct loop_device *lo = bdev->bd_disk->private_data; int err; - mutex_lock(&lo->lo_ctl_mutex); + mutex_lock_nested(&lo->lo_ctl_mutex, 1); switch (cmd) { case LOOP_SET_FD: err = loop_set_fd(lo, mode, bdev, arg); Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be a better solution. Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex held, so no lockdep annotation is required. If you do not like the special handling of the lo_ctl_mutex just for the LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside each of the individual ioctl handlers and I could send you another patch. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 18:33:41 +07:00
/* loop_clr_fd would have unlocked lo_ctl_mutex on success */
err = loop_clr_fd(lo);
loop: fix circular locking in loop_clr_fd() With CONFIG_PROVE_LOCKING enabled $ losetup /dev/loop0 file $ losetup -o 32256 /dev/loop1 /dev/loop0 $ losetup -d /dev/loop1 $ losetup -d /dev/loop0 triggers a [ INFO: possible circular locking dependency detected ] I think this warning is a false positive. Open/close on a loop device acquires bd_mutex of the device before acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of a device, if backing file is a device and this is the last reference to the file being dropped . But it is guaranteed that it is impossible to have a circular list of backing devices.(say loop2->loop1->loop0->loop2 is not possible), which guarantees that this can never deadlock. So this warning should be suppressed. It is very difficult to annotate lockdep not to warn here in the correct way. A simple way to silence lockdep could be to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some other real bugs. @@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode, struct loop_device *lo = bdev->bd_disk->private_data; int err; - mutex_lock(&lo->lo_ctl_mutex); + mutex_lock_nested(&lo->lo_ctl_mutex, 1); switch (cmd) { case LOOP_SET_FD: err = loop_set_fd(lo, mode, bdev, arg); Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be a better solution. Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex held, so no lockdep annotation is required. If you do not like the special handling of the lo_ctl_mutex just for the LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside each of the individual ioctl handlers and I could send you another patch. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 18:33:41 +07:00
if (!err)
goto out_unlocked;
break;
case LOOP_SET_STATUS:
err = -EPERM;
if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
err = loop_set_status_old(lo,
(struct loop_info __user *)arg);
break;
case LOOP_GET_STATUS:
err = loop_get_status_old(lo, (struct loop_info __user *) arg);
break;
case LOOP_SET_STATUS64:
err = -EPERM;
if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
err = loop_set_status64(lo,
(struct loop_info64 __user *) arg);
break;
case LOOP_GET_STATUS64:
err = loop_get_status64(lo, (struct loop_info64 __user *) arg);
break;
loop: add ioctl to resize a loop device Add the ability to 'resize' the loop device on the fly. One practical application is a loop file with XFS filesystem, already mounted: You can easily enlarge the file (append some bytes) and then call ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the new size and you can use xfs_growfs later on, which will allow you to use full capacity of the loop file without the need to unmount. Test app: #include <linux/fs.h> #include <linux/loop.h> #include <sys/ioctl.h> #include <sys/stat.h> #include <sys/types.h> #include <assert.h> #include <errno.h> #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _GNU_SOURCE #include <getopt.h> char *me; void usage(FILE *f) { fprintf(f, "%s [options] loop_dev [backend_file]\n" "-s, --set new_size_in_bytes\n" "\twhen backend_file is given, " "it will be expanded too while keeping the original contents\n", me); } struct option opts[] = { { .name = "set", .has_arg = 1, .flag = NULL, .val = 's' }, { .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' } }; void err_size(char *name, __u64 old) { fprintf(stderr, "size must be larger than current %s (%llu)\n", name, old); } int main(int argc, char *argv[]) { int fd, err, c, i, bfd; ssize_t ssz; size_t sz; __u64 old, new, append; char a[BUFSIZ]; struct stat st; FILE *out; char *backend, *dev; err = EINVAL; out = stderr; me = argv[0]; new = 0; while ((c = getopt_long(argc, argv, "s:h", opts, &i)) != -1) { switch (c) { case 's': errno = 0; new = strtoull(optarg, NULL, 0); if (errno) { err = errno; perror(argv[i]); goto out; } break; case 'h': err = 0; out = stdout; goto err; default: perror(argv[i]); goto err; } } if (optind < argc) dev = argv[optind++]; else goto err; fd = open(dev, O_RDONLY); if (fd < 0) { err = errno; perror(dev); goto out; } err = ioctl(fd, BLKGETSIZE64, &old); if (err) { err = errno; perror("ioctl BLKGETSIZE64"); goto out; } if (!new) { printf("%llu\n", old); goto out; } if (new < old) { err = EINVAL; err_size(dev, old); goto out; } if (optind < argc) { backend = argv[optind++]; bfd = open(backend, O_WRONLY|O_APPEND); if (bfd < 0) { err = errno; perror(backend); goto out; } err = fstat(bfd, &st); if (err) { err = errno; perror(backend); goto out; } if (new < st.st_size) { err = EINVAL; err_size(backend, st.st_size); goto out; } append = new - st.st_size; sz = sizeof(a); while (append > 0) { if (append < sz) sz = append; ssz = write(bfd, a, sz); if (ssz != sz) { err = errno; perror(backend); goto out; } append -= sz; } err = fsync(bfd); if (err) { err = errno; perror(backend); goto out; } } err = ioctl(fd, LOOP_SET_CAPACITY, new); if (err) { err = errno; perror("ioctl LOOP_SET_CAPACITY"); } goto out; err: usage(out); out: return err; } Signed-off-by: J. R. Okajima <hooanon05@yahoo.co.jp> Signed-off-by: Tomas Matejicek <tomas@slax.org> Cc: <util-linux-ng@vger.kernel.org> Cc: Karel Zak <kzak@redhat.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: <linux-api@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-01 05:23:43 +07:00
case LOOP_SET_CAPACITY:
err = -EPERM;
if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
err = loop_set_capacity(lo, bdev);
break;
case LOOP_SET_DIRECT_IO:
err = -EPERM;
if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
err = loop_set_dio(lo, arg);
break;
default:
err = lo->ioctl ? lo->ioctl(lo, cmd, arg) : -EINVAL;
}
mutex_unlock(&lo->lo_ctl_mutex);
loop: fix circular locking in loop_clr_fd() With CONFIG_PROVE_LOCKING enabled $ losetup /dev/loop0 file $ losetup -o 32256 /dev/loop1 /dev/loop0 $ losetup -d /dev/loop1 $ losetup -d /dev/loop0 triggers a [ INFO: possible circular locking dependency detected ] I think this warning is a false positive. Open/close on a loop device acquires bd_mutex of the device before acquiring lo_ctl_mutex of the same device. For ioctl(LOOP_CLR_FD) after acquiring lo_ctl_mutex, fput on the backing_file might acquire the bd_mutex of a device, if backing file is a device and this is the last reference to the file being dropped . But it is guaranteed that it is impossible to have a circular list of backing devices.(say loop2->loop1->loop0->loop2 is not possible), which guarantees that this can never deadlock. So this warning should be suppressed. It is very difficult to annotate lockdep not to warn here in the correct way. A simple way to silence lockdep could be to mark the lo_ctl_mutex in ioctl to be a sub class, but this might mask some other real bugs. @@ -1164,7 +1164,7 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode, struct loop_device *lo = bdev->bd_disk->private_data; int err; - mutex_lock(&lo->lo_ctl_mutex); + mutex_lock_nested(&lo->lo_ctl_mutex, 1); switch (cmd) { case LOOP_SET_FD: err = loop_set_fd(lo, mode, bdev, arg); Or actually marking the bd_mutex after lo_ctl_mutex as a sub class could be a better solution. Luckily it is easy to avoid calling fput on backing file with lo_ctl_mutex held, so no lockdep annotation is required. If you do not like the special handling of the lo_ctl_mutex just for the LOOP_CLR_FD ioctl in lo_ioctl(), the mutex handling could be moved inside each of the individual ioctl handlers and I could send you another patch. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-03-24 18:33:41 +07:00
out_unlocked:
return err;
}
#ifdef CONFIG_COMPAT
struct compat_loop_info {
compat_int_t lo_number; /* ioctl r/o */
compat_dev_t lo_device; /* ioctl r/o */
compat_ulong_t lo_inode; /* ioctl r/o */
compat_dev_t lo_rdevice; /* ioctl r/o */
compat_int_t lo_offset;
compat_int_t lo_encrypt_type;
compat_int_t lo_encrypt_key_size; /* ioctl w/o */
compat_int_t lo_flags; /* ioctl r/o */
char lo_name[LO_NAME_SIZE];
unsigned char lo_encrypt_key[LO_KEY_SIZE]; /* ioctl w/o */
compat_ulong_t lo_init[2];
char reserved[4];
};
/*
* Transfer 32-bit compatibility structure in userspace to 64-bit loop info
* - noinlined to reduce stack space usage in main part of driver
*/
static noinline int
loop_info64_from_compat(const struct compat_loop_info __user *arg,
struct loop_info64 *info64)
{
struct compat_loop_info info;
if (copy_from_user(&info, arg, sizeof(info)))
return -EFAULT;
memset(info64, 0, sizeof(*info64));
info64->lo_number = info.lo_number;
info64->lo_device = info.lo_device;
info64->lo_inode = info.lo_inode;
info64->lo_rdevice = info.lo_rdevice;
info64->lo_offset = info.lo_offset;
info64->lo_sizelimit = 0;
info64->lo_encrypt_type = info.lo_encrypt_type;
info64->lo_encrypt_key_size = info.lo_encrypt_key_size;
info64->lo_flags = info.lo_flags;
info64->lo_init[0] = info.lo_init[0];
info64->lo_init[1] = info.lo_init[1];
if (info.lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
memcpy(info64->lo_crypt_name, info.lo_name, LO_NAME_SIZE);
else
memcpy(info64->lo_file_name, info.lo_name, LO_NAME_SIZE);
memcpy(info64->lo_encrypt_key, info.lo_encrypt_key, LO_KEY_SIZE);
return 0;
}
/*
* Transfer 64-bit loop info to 32-bit compatibility structure in userspace
* - noinlined to reduce stack space usage in main part of driver
*/
static noinline int
loop_info64_to_compat(const struct loop_info64 *info64,
struct compat_loop_info __user *arg)
{
struct compat_loop_info info;
memset(&info, 0, sizeof(info));
info.lo_number = info64->lo_number;
info.lo_device = info64->lo_device;
info.lo_inode = info64->lo_inode;
info.lo_rdevice = info64->lo_rdevice;
info.lo_offset = info64->lo_offset;
info.lo_encrypt_type = info64->lo_encrypt_type;
info.lo_encrypt_key_size = info64->lo_encrypt_key_size;
info.lo_flags = info64->lo_flags;
info.lo_init[0] = info64->lo_init[0];
info.lo_init[1] = info64->lo_init[1];
if (info.lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
memcpy(info.lo_name, info64->lo_crypt_name, LO_NAME_SIZE);
else
memcpy(info.lo_name, info64->lo_file_name, LO_NAME_SIZE);
memcpy(info.lo_encrypt_key, info64->lo_encrypt_key, LO_KEY_SIZE);
/* error in case values were truncated */
if (info.lo_device != info64->lo_device ||
info.lo_rdevice != info64->lo_rdevice ||
info.lo_inode != info64->lo_inode ||
info.lo_offset != info64->lo_offset ||
info.lo_init[0] != info64->lo_init[0] ||
info.lo_init[1] != info64->lo_init[1])
return -EOVERFLOW;
if (copy_to_user(arg, &info, sizeof(info)))
return -EFAULT;
return 0;
}
static int
loop_set_status_compat(struct loop_device *lo,
const struct compat_loop_info __user *arg)
{
struct loop_info64 info64;
int ret;
ret = loop_info64_from_compat(arg, &info64);
if (ret < 0)
return ret;
return loop_set_status(lo, &info64);
}
static int
loop_get_status_compat(struct loop_device *lo,
struct compat_loop_info __user *arg)
{
struct loop_info64 info64;
int err = 0;
if (!arg)
err = -EINVAL;
if (!err)
err = loop_get_status(lo, &info64);
if (!err)
err = loop_info64_to_compat(&info64, arg);
return err;
}
static int lo_compat_ioctl(struct block_device *bdev, fmode_t mode,
unsigned int cmd, unsigned long arg)
{
struct loop_device *lo = bdev->bd_disk->private_data;
int err;
switch(cmd) {
case LOOP_SET_STATUS:
mutex_lock(&lo->lo_ctl_mutex);
err = loop_set_status_compat(
lo, (const struct compat_loop_info __user *) arg);
mutex_unlock(&lo->lo_ctl_mutex);
break;
case LOOP_GET_STATUS:
mutex_lock(&lo->lo_ctl_mutex);
err = loop_get_status_compat(
lo, (struct compat_loop_info __user *) arg);
mutex_unlock(&lo->lo_ctl_mutex);
break;
loop: add ioctl to resize a loop device Add the ability to 'resize' the loop device on the fly. One practical application is a loop file with XFS filesystem, already mounted: You can easily enlarge the file (append some bytes) and then call ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the new size and you can use xfs_growfs later on, which will allow you to use full capacity of the loop file without the need to unmount. Test app: #include <linux/fs.h> #include <linux/loop.h> #include <sys/ioctl.h> #include <sys/stat.h> #include <sys/types.h> #include <assert.h> #include <errno.h> #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #define _GNU_SOURCE #include <getopt.h> char *me; void usage(FILE *f) { fprintf(f, "%s [options] loop_dev [backend_file]\n" "-s, --set new_size_in_bytes\n" "\twhen backend_file is given, " "it will be expanded too while keeping the original contents\n", me); } struct option opts[] = { { .name = "set", .has_arg = 1, .flag = NULL, .val = 's' }, { .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' } }; void err_size(char *name, __u64 old) { fprintf(stderr, "size must be larger than current %s (%llu)\n", name, old); } int main(int argc, char *argv[]) { int fd, err, c, i, bfd; ssize_t ssz; size_t sz; __u64 old, new, append; char a[BUFSIZ]; struct stat st; FILE *out; char *backend, *dev; err = EINVAL; out = stderr; me = argv[0]; new = 0; while ((c = getopt_long(argc, argv, "s:h", opts, &i)) != -1) { switch (c) { case 's': errno = 0; new = strtoull(optarg, NULL, 0); if (errno) { err = errno; perror(argv[i]); goto out; } break; case 'h': err = 0; out = stdout; goto err; default: perror(argv[i]); goto err; } } if (optind < argc) dev = argv[optind++]; else goto err; fd = open(dev, O_RDONLY); if (fd < 0) { err = errno; perror(dev); goto out; } err = ioctl(fd, BLKGETSIZE64, &old); if (err) { err = errno; perror("ioctl BLKGETSIZE64"); goto out; } if (!new) { printf("%llu\n", old); goto out; } if (new < old) { err = EINVAL; err_size(dev, old); goto out; } if (optind < argc) { backend = argv[optind++]; bfd = open(backend, O_WRONLY|O_APPEND); if (bfd < 0) { err = errno; perror(backend); goto out; } err = fstat(bfd, &st); if (err) { err = errno; perror(backend); goto out; } if (new < st.st_size) { err = EINVAL; err_size(backend, st.st_size); goto out; } append = new - st.st_size; sz = sizeof(a); while (append > 0) { if (append < sz) sz = append; ssz = write(bfd, a, sz); if (ssz != sz) { err = errno; perror(backend); goto out; } append -= sz; } err = fsync(bfd); if (err) { err = errno; perror(backend); goto out; } } err = ioctl(fd, LOOP_SET_CAPACITY, new); if (err) { err = errno; perror("ioctl LOOP_SET_CAPACITY"); } goto out; err: usage(out); out: return err; } Signed-off-by: J. R. Okajima <hooanon05@yahoo.co.jp> Signed-off-by: Tomas Matejicek <tomas@slax.org> Cc: <util-linux-ng@vger.kernel.org> Cc: Karel Zak <kzak@redhat.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: <linux-api@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-01 05:23:43 +07:00
case LOOP_SET_CAPACITY:
case LOOP_CLR_FD:
case LOOP_GET_STATUS64:
case LOOP_SET_STATUS64:
arg = (unsigned long) compat_ptr(arg);
case LOOP_SET_FD:
case LOOP_CHANGE_FD:
err = lo_ioctl(bdev, mode, cmd, arg);
break;
default:
err = -ENOIOCTLCMD;
break;
}
return err;
}
#endif
static int lo_open(struct block_device *bdev, fmode_t mode)
{
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
struct loop_device *lo;
int err = 0;
mutex_lock(&loop_index_mutex);
lo = bdev->bd_disk->private_data;
if (!lo) {
err = -ENXIO;
goto out;
}
atomic_inc(&lo->lo_refcnt);
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
out:
mutex_unlock(&loop_index_mutex);
return err;
}
static void lo_release(struct gendisk *disk, fmode_t mode)
{
struct loop_device *lo = disk->private_data;
int err;
if (atomic_dec_return(&lo->lo_refcnt))
return;
mutex_lock(&lo->lo_ctl_mutex);
if (lo->lo_flags & LO_FLAGS_AUTOCLEAR) {
/*
* In autoclear mode, stop the loop thread
* and remove configuration after last close.
*/
err = loop_clr_fd(lo);
if (!err)
return;
} else {
/*
* Otherwise keep thread (if running) and config,
* but flush possible ongoing bios in thread.
*/
loop_flush(lo);
}
mutex_unlock(&lo->lo_ctl_mutex);
}
static const struct block_device_operations lo_fops = {
.owner = THIS_MODULE,
.open = lo_open,
.release = lo_release,
.ioctl = lo_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = lo_compat_ioctl,
#endif
};
/*
* And now the modules code and kernel interface.
*/
static int max_loop;
module_param(max_loop, int, S_IRUGO);
MODULE_PARM_DESC(max_loop, "Maximum number of loop devices");
module_param(max_part, int, S_IRUGO);
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
MODULE_PARM_DESC(max_part, "Maximum number of partitions per loop device");
MODULE_LICENSE("GPL");
MODULE_ALIAS_BLOCKDEV_MAJOR(LOOP_MAJOR);
int loop_register_transfer(struct loop_func_table *funcs)
{
unsigned int n = funcs->number;
if (n >= MAX_LO_CRYPT || xfer_funcs[n])
return -EINVAL;
xfer_funcs[n] = funcs;
return 0;
}
static int unregister_transfer_cb(int id, void *ptr, void *data)
{
struct loop_device *lo = ptr;
struct loop_func_table *xfer = data;
mutex_lock(&lo->lo_ctl_mutex);
if (lo->lo_encryption == xfer)
loop_release_xfer(lo);
mutex_unlock(&lo->lo_ctl_mutex);
return 0;
}
int loop_unregister_transfer(int number)
{
unsigned int n = number;
struct loop_func_table *xfer;
if (n == 0 || n >= MAX_LO_CRYPT || (xfer = xfer_funcs[n]) == NULL)
return -EINVAL;
xfer_funcs[n] = NULL;
idr_for_each(&loop_index_idr, &unregister_transfer_cb, xfer);
return 0;
}
EXPORT_SYMBOL(loop_register_transfer);
EXPORT_SYMBOL(loop_unregister_transfer);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
static int loop_queue_rq(struct blk_mq_hw_ctx *hctx,
const struct blk_mq_queue_data *bd)
{
struct loop_cmd *cmd = blk_mq_rq_to_pdu(bd->rq);
struct loop_device *lo = cmd->rq->q->queuedata;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
blk_mq_start_request(bd->rq);
if (lo->lo_state != Lo_bound)
return -EIO;
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
if (lo->use_dio && !(cmd->rq->cmd_flags & (REQ_FLUSH |
REQ_DISCARD)))
cmd->use_aio = true;
else
cmd->use_aio = false;
queue_kthread_work(&lo->worker, &cmd->work);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
return BLK_MQ_RQ_QUEUE_OK;
}
static void loop_handle_cmd(struct loop_cmd *cmd)
{
const bool write = cmd->rq->cmd_flags & REQ_WRITE;
struct loop_device *lo = cmd->rq->q->queuedata;
int ret = 0;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
if (write && (lo->lo_flags & LO_FLAGS_READ_ONLY)) {
ret = -EIO;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
goto failed;
}
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
ret = do_req_filebacked(lo, cmd->rq);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
failed:
block: loop: support DIO & AIO There are at least 3 advantages to use direct I/O and AIO on read/write loop's backing file: 1) double cache can be avoided, then memory usage gets decreased a lot 2) not like user space direct I/O, there isn't cost of pinning pages 3) avoid context switch for obtaining good throughput - in buffered file read, random I/O top throughput is often obtained only if they are submitted concurrently from lots of tasks; but for sequential I/O, most of times they can be hit from page cache, so concurrent submissions often introduce unnecessary context switch and can't improve throughput much. There was such discussion[1] to use non-blocking I/O to improve the problem for application. - with direct I/O and AIO, concurrent submissions can be avoided and random read throughput can't be affected meantime xfstests(-g auto, ext4) is basically passed when running with direct I/O(aio), one exception is generic/232, but it failed in loop buffered I/O(4.2-rc6-next-20150814) too. Follows the fio test result for performance purpose: 4 jobs fio test inside ext4 file system over loop block 1) How to run - KVM: 4 VCPUs, 2G RAM - linux kernel: 4.2-rc6-next-20150814(base) with the patchset - the loop block is over one image on SSD. - linux psync, 4 jobs, size 1500M, ext4 over loop block - test result: IOPS from fio output 2) Throughput(IOPS) becomes a bit better with direct I/O(aio) ------------------------------------------------------------- test cases |randread |read |randwrite |write | ------------------------------------------------------------- base |8015 |113811 |67442 |106978 ------------------------------------------------------------- base+loop aio |8136 |125040 |67811 |111376 ------------------------------------------------------------- - somehow, it should be caused by more page cache avaiable for application or one extra page copy is avoided in case of direct I/O 3) context switch - context switch decreased by ~50% with loop direct I/O(aio) compared with loop buffered I/O(4.2-rc6-next-20150814) 4) memory usage from /proc/meminfo ------------------------------------------------------------- | Buffers | Cached ------------------------------------------------------------- base | > 760MB | ~950MB ------------------------------------------------------------- base+loop direct I/O(aio) | < 5MB | ~1.6GB ------------------------------------------------------------- - so there are much more page caches available for application with direct I/O [1] https://lwn.net/Articles/612483/ Signed-off-by: Ming Lei <ming.lei@canonical.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-08-17 09:31:51 +07:00
/* complete non-aio request */
if (!cmd->use_aio || ret)
blk_mq_complete_request(cmd->rq, ret ? -EIO : 0);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
}
static void loop_queue_work(struct kthread_work *work)
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
{
struct loop_cmd *cmd =
container_of(work, struct loop_cmd, work);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
loop_handle_cmd(cmd);
}
static int loop_init_request(void *data, struct request *rq,
unsigned int hctx_idx, unsigned int request_idx,
unsigned int numa_node)
{
struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
cmd->rq = rq;
init_kthread_work(&cmd->work, loop_queue_work);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
return 0;
}
static struct blk_mq_ops loop_mq_ops = {
.queue_rq = loop_queue_rq,
.map_queue = blk_mq_map_queue,
.init_request = loop_init_request,
};
static int loop_add(struct loop_device **l, int i)
{
struct loop_device *lo;
struct gendisk *disk;
int err;
err = -ENOMEM;
lo = kzalloc(sizeof(*lo), GFP_KERNEL);
if (!lo)
goto out;
loop: fix crash when using unassigned loop device When the loop module is loaded, it creates 8 loop devices /dev/loop[0-7]. The devices have no request routine and thus, when they are used without being assigned, a crash happens. For example, these commands cause crash (assuming there are no used loop devices): Kernel Fault: Code=26 regs=000000007f420980 (Addr=0000000000000010) CPU: 1 PID: 50 Comm: kworker/1:1 Not tainted 3.11.0 #1 Workqueue: ksnaphd do_metadata [dm_snapshot] task: 000000007fcf4078 ti: 000000007f420000 task.ti: 000000007f420000 [ 116.319988] YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI PSW: 00001000000001001111111100001111 Not tainted r00-03 000000ff0804ff0f 00000000408bf5d0 00000000402d8204 000000007b7ff6c0 r04-07 00000000408a95d0 000000007f420950 000000007b7ff6c0 000000007d06c930 r08-11 000000007f4205c0 0000000000000001 000000007f4205c0 000000007f4204b8 r12-15 0000000000000010 0000000000000000 0000000000000000 0000000000000000 r16-19 000000001108dd48 000000004061cd7c 000000007d859800 000000000800000f r20-23 0000000000000000 0000000000000008 0000000000000000 0000000000000000 r24-27 00000000ffffffff 000000007b7ff6c0 000000007d859800 00000000408a95d0 r28-31 0000000000000000 000000007f420950 000000007f420980 000000007f4208e8 sr00-03 0000000000000000 0000000000000000 0000000000000000 0000000000303000 sr04-07 0000000000000000 0000000000000000 0000000000000000 0000000000000000 [ 117.549988] IASQ: 0000000000000000 0000000000000000 IAOQ: 00000000402d82fc 00000000402d8300 IIR: 53820020 ISR: 0000000000000000 IOR: 0000000000000010 CPU: 1 CR30: 000000007f420000 CR31: ffffffffffffffff ORIG_R28: 0000000000000001 IAOQ[0]: generic_make_request+0x11c/0x1a0 IAOQ[1]: generic_make_request+0x120/0x1a0 RP(r2): generic_make_request+0x24/0x1a0 Backtrace: [<00000000402d83f0>] submit_bio+0x70/0x140 [<0000000011087c4c>] dispatch_io+0x234/0x478 [dm_mod] [<0000000011087f44>] sync_io+0xb4/0x190 [dm_mod] [<00000000110883bc>] dm_io+0x2c4/0x310 [dm_mod] [<00000000110bfcd0>] do_metadata+0x28/0xb0 [dm_snapshot] [<00000000401591d8>] process_one_work+0x160/0x460 [<0000000040159bc0>] worker_thread+0x300/0x478 [<0000000040161a70>] kthread+0x118/0x128 [<0000000040104020>] end_fault_vector+0x20/0x28 [<0000000040177220>] task_tick_fair+0x420/0x4d0 [<00000000401aa048>] invoke_rcu_core+0x50/0x60 [<00000000401ad5b8>] rcu_check_callbacks+0x210/0x8d8 [<000000004014aaa0>] update_process_times+0xa8/0xc0 [<00000000401ab86c>] rcu_process_callbacks+0x4b4/0x598 [<0000000040142408>] __do_softirq+0x250/0x2c0 [<00000000401789d0>] find_busiest_group+0x3c0/0xc70 [ 119.379988] Kernel panic - not syncing: Kernel Fault Rebooting in 1 seconds.. Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Cc: stable@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2013-10-16 03:14:38 +07:00
lo->lo_state = Lo_unbound;
/* allocate id, if @id >= 0, we're requesting that specific id */
if (i >= 0) {
err = idr_alloc(&loop_index_idr, lo, i, i + 1, GFP_KERNEL);
if (err == -ENOSPC)
err = -EEXIST;
} else {
err = idr_alloc(&loop_index_idr, lo, 0, 0, GFP_KERNEL);
}
if (err < 0)
goto out_free_dev;
i = err;
err = -ENOMEM;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
lo->tag_set.ops = &loop_mq_ops;
lo->tag_set.nr_hw_queues = 1;
lo->tag_set.queue_depth = 128;
lo->tag_set.numa_node = NUMA_NO_NODE;
lo->tag_set.cmd_size = sizeof(struct loop_cmd);
lo->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_SG_MERGE;
lo->tag_set.driver_data = lo;
err = blk_mq_alloc_tag_set(&lo->tag_set);
if (err)
loop: fix crash if blk_alloc_queue fails loop: fix crash if blk_alloc_queue fails If blk_alloc_queue fails, loop_add cleans up, but it doesn't clean up the identifier allocated with idr_alloc. That causes crash on module unload in idr_for_each(&loop_index_idr, &loop_exit_cb, NULL); where we attempt to remove non-existed device with that id. BUG: unable to handle kernel NULL pointer dereference at 0000000000000380 IP: [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 PGD 43d399067 PUD 43d0ad067 PMD 0 Oops: 0000 [#1] PREEMPT SMP Modules linked in: loop(-) dm_snapshot dm_zero dm_mirror dm_region_hash dm_log dm_loop dm_mod ip6table_filter ip6_tables uvesafb cfbcopyarea cfbimgblt cfbfillrect fbcon font bitblit fbcon_rotate fbcon_cw fbcon_ud fbcon_ccw softcursor fb fbdev msr ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 xt_state ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp llc tun ipv6 cpufreq_userspace cpufreq_stats cpufreq_ondemand cpufreq_conservative cpufreq_powersave spadfs fuse hid_generic usbhid hid raid0 md_mod dmi_sysfs nf_nat_ftp nf_nat nf_conntrack_ftp nf_conntrack snd_usb_audio snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd_page_alloc lm85 hwmon_vid snd_hwdep snd_usbmidi_lib snd_rawmidi snd soundcore acpi_cpufreq ohci_hcd freq_table tg3 ehci_pci mperf ehci_hcd kvm_amd kvm sata_svw serverworks libphy libata ide_core k10temp usbcore hwmon microcode ptp pcspkr pps_core e100 skge mii usb_common i2c_piix4 floppy evdev rtc_cmos i2c_core processor but! ton unix CPU: 7 PID: 2735 Comm: rmmod Tainted: G W 3.10.15-devel #15 Hardware name: empty empty/S3992-E, BIOS 'V1.06 ' 06/09/2009 task: ffff88043d38e780 ti: ffff88043d21e000 task.ti: ffff88043d21e000 RIP: 0010:[<ffffffff812057c9>] [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 RSP: 0018:ffff88043d21fe10 EFLAGS: 00010282 RAX: ffffffffa05102e0 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff88043ea82800 RDI: 0000000000000000 RBP: ffff88043d21fe48 R08: 0000000000000000 R09: 0000000000000001 R10: 0000000000000001 R11: 0000000000000000 R12: 00000000000000ff R13: 0000000000000080 R14: 0000000000000000 R15: ffff88043ea82800 FS: 00007ff646534700(0000) GS:ffff880447000000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000380 CR3: 000000043e9bf000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: ffffffff8100aba4 0000000000000092 ffff88043d21fe48 ffff88043ea82800 00000000000000ff ffff88043d21fe98 0000000000000000 ffff88043d21fe60 ffffffffa05102b4 0000000000000000 ffff88043d21fe70 ffffffffa05102ec Call Trace: [<ffffffff8100aba4>] ? native_sched_clock+0x24/0x80 [<ffffffffa05102b4>] loop_remove+0x14/0x40 [loop] [<ffffffffa05102ec>] loop_exit_cb+0xc/0x10 [loop] [<ffffffff81217b74>] idr_for_each+0x104/0x190 [<ffffffffa05102e0>] ? loop_remove+0x40/0x40 [loop] [<ffffffff8109adc5>] ? trace_hardirqs_on_caller+0x105/0x1d0 [<ffffffffa05135dc>] loop_exit+0x34/0xa58 [loop] [<ffffffff810a98ea>] SyS_delete_module+0x13a/0x260 [<ffffffff81221d5e>] ? trace_hardirqs_on_thunk+0x3a/0x3f [<ffffffff813cff16>] system_call_fastpath+0x1a/0x1f Code: f0 4c 8b 6d f8 c9 c3 66 66 2e 0f 1f 84 00 00 00 00 00 55 48 89 e5 41 56 41 55 4c 8d af 80 00 00 00 41 54 53 48 89 fb 48 83 ec 18 <48> 83 bf 80 03 00 00 00 74 4d e8 98 fe ff ff 31 f6 48 c7 c7 20 RIP [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 RSP <ffff88043d21fe10> CR2: 0000000000000380 ---[ end trace 64ec069ec70f1309 ]--- Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: stable@kernel.org # 3.1+ Signed-off-by: Jens Axboe <axboe@kernel.dk>
2013-10-14 23:12:24 +07:00
goto out_free_idr;
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
lo->lo_queue = blk_mq_init_queue(&lo->tag_set);
if (IS_ERR_OR_NULL(lo->lo_queue)) {
err = PTR_ERR(lo->lo_queue);
goto out_cleanup_tags;
}
loop: fix crash when using unassigned loop device When the loop module is loaded, it creates 8 loop devices /dev/loop[0-7]. The devices have no request routine and thus, when they are used without being assigned, a crash happens. For example, these commands cause crash (assuming there are no used loop devices): Kernel Fault: Code=26 regs=000000007f420980 (Addr=0000000000000010) CPU: 1 PID: 50 Comm: kworker/1:1 Not tainted 3.11.0 #1 Workqueue: ksnaphd do_metadata [dm_snapshot] task: 000000007fcf4078 ti: 000000007f420000 task.ti: 000000007f420000 [ 116.319988] YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI PSW: 00001000000001001111111100001111 Not tainted r00-03 000000ff0804ff0f 00000000408bf5d0 00000000402d8204 000000007b7ff6c0 r04-07 00000000408a95d0 000000007f420950 000000007b7ff6c0 000000007d06c930 r08-11 000000007f4205c0 0000000000000001 000000007f4205c0 000000007f4204b8 r12-15 0000000000000010 0000000000000000 0000000000000000 0000000000000000 r16-19 000000001108dd48 000000004061cd7c 000000007d859800 000000000800000f r20-23 0000000000000000 0000000000000008 0000000000000000 0000000000000000 r24-27 00000000ffffffff 000000007b7ff6c0 000000007d859800 00000000408a95d0 r28-31 0000000000000000 000000007f420950 000000007f420980 000000007f4208e8 sr00-03 0000000000000000 0000000000000000 0000000000000000 0000000000303000 sr04-07 0000000000000000 0000000000000000 0000000000000000 0000000000000000 [ 117.549988] IASQ: 0000000000000000 0000000000000000 IAOQ: 00000000402d82fc 00000000402d8300 IIR: 53820020 ISR: 0000000000000000 IOR: 0000000000000010 CPU: 1 CR30: 000000007f420000 CR31: ffffffffffffffff ORIG_R28: 0000000000000001 IAOQ[0]: generic_make_request+0x11c/0x1a0 IAOQ[1]: generic_make_request+0x120/0x1a0 RP(r2): generic_make_request+0x24/0x1a0 Backtrace: [<00000000402d83f0>] submit_bio+0x70/0x140 [<0000000011087c4c>] dispatch_io+0x234/0x478 [dm_mod] [<0000000011087f44>] sync_io+0xb4/0x190 [dm_mod] [<00000000110883bc>] dm_io+0x2c4/0x310 [dm_mod] [<00000000110bfcd0>] do_metadata+0x28/0xb0 [dm_snapshot] [<00000000401591d8>] process_one_work+0x160/0x460 [<0000000040159bc0>] worker_thread+0x300/0x478 [<0000000040161a70>] kthread+0x118/0x128 [<0000000040104020>] end_fault_vector+0x20/0x28 [<0000000040177220>] task_tick_fair+0x420/0x4d0 [<00000000401aa048>] invoke_rcu_core+0x50/0x60 [<00000000401ad5b8>] rcu_check_callbacks+0x210/0x8d8 [<000000004014aaa0>] update_process_times+0xa8/0xc0 [<00000000401ab86c>] rcu_process_callbacks+0x4b4/0x598 [<0000000040142408>] __do_softirq+0x250/0x2c0 [<00000000401789d0>] find_busiest_group+0x3c0/0xc70 [ 119.379988] Kernel panic - not syncing: Kernel Fault Rebooting in 1 seconds.. Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Cc: stable@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2013-10-16 03:14:38 +07:00
lo->lo_queue->queuedata = lo;
/*
* It doesn't make sense to enable merge because the I/O
* submitted to backing file is handled page by page.
*/
queue_flag_set_unlocked(QUEUE_FLAG_NOMERGES, lo->lo_queue);
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
disk = lo->lo_disk = alloc_disk(1 << part_shift);
if (!disk)
goto out_free_queue;
2011-08-24 01:12:04 +07:00
/*
* Disable partition scanning by default. The in-kernel partition
* scanning can be requested individually per-device during its
* setup. Userspace can always add and remove partitions from all
* devices. The needed partition minors are allocated from the
* extended minor space, the main loop device numbers will continue
* to match the loop minors, regardless of the number of partitions
* used.
*
* If max_part is given, partition scanning is globally enabled for
* all loop devices. The minors for the main loop devices will be
* multiples of max_part.
*
* Note: Global-for-all-devices, set-only-at-init, read-only module
* parameteters like 'max_loop' and 'max_part' make things needlessly
* complicated, are too static, inflexible and may surprise
* userspace tools. Parameters like this in general should be avoided.
*/
if (!part_shift)
disk->flags |= GENHD_FL_NO_PART_SCAN;
disk->flags |= GENHD_FL_EXT_DEVT;
mutex_init(&lo->lo_ctl_mutex);
atomic_set(&lo->lo_refcnt, 0);
lo->lo_number = i;
spin_lock_init(&lo->lo_lock);
disk->major = LOOP_MAJOR;
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
disk->first_minor = i << part_shift;
disk->fops = &lo_fops;
disk->private_data = lo;
disk->queue = lo->lo_queue;
sprintf(disk->disk_name, "loop%d", i);
add_disk(disk);
*l = lo;
return lo->lo_number;
out_free_queue:
blk_cleanup_queue(lo->lo_queue);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
out_cleanup_tags:
blk_mq_free_tag_set(&lo->tag_set);
loop: fix crash if blk_alloc_queue fails loop: fix crash if blk_alloc_queue fails If blk_alloc_queue fails, loop_add cleans up, but it doesn't clean up the identifier allocated with idr_alloc. That causes crash on module unload in idr_for_each(&loop_index_idr, &loop_exit_cb, NULL); where we attempt to remove non-existed device with that id. BUG: unable to handle kernel NULL pointer dereference at 0000000000000380 IP: [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 PGD 43d399067 PUD 43d0ad067 PMD 0 Oops: 0000 [#1] PREEMPT SMP Modules linked in: loop(-) dm_snapshot dm_zero dm_mirror dm_region_hash dm_log dm_loop dm_mod ip6table_filter ip6_tables uvesafb cfbcopyarea cfbimgblt cfbfillrect fbcon font bitblit fbcon_rotate fbcon_cw fbcon_ud fbcon_ccw softcursor fb fbdev msr ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 xt_state ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp llc tun ipv6 cpufreq_userspace cpufreq_stats cpufreq_ondemand cpufreq_conservative cpufreq_powersave spadfs fuse hid_generic usbhid hid raid0 md_mod dmi_sysfs nf_nat_ftp nf_nat nf_conntrack_ftp nf_conntrack snd_usb_audio snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd_page_alloc lm85 hwmon_vid snd_hwdep snd_usbmidi_lib snd_rawmidi snd soundcore acpi_cpufreq ohci_hcd freq_table tg3 ehci_pci mperf ehci_hcd kvm_amd kvm sata_svw serverworks libphy libata ide_core k10temp usbcore hwmon microcode ptp pcspkr pps_core e100 skge mii usb_common i2c_piix4 floppy evdev rtc_cmos i2c_core processor but! ton unix CPU: 7 PID: 2735 Comm: rmmod Tainted: G W 3.10.15-devel #15 Hardware name: empty empty/S3992-E, BIOS 'V1.06 ' 06/09/2009 task: ffff88043d38e780 ti: ffff88043d21e000 task.ti: ffff88043d21e000 RIP: 0010:[<ffffffff812057c9>] [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 RSP: 0018:ffff88043d21fe10 EFLAGS: 00010282 RAX: ffffffffa05102e0 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff88043ea82800 RDI: 0000000000000000 RBP: ffff88043d21fe48 R08: 0000000000000000 R09: 0000000000000001 R10: 0000000000000001 R11: 0000000000000000 R12: 00000000000000ff R13: 0000000000000080 R14: 0000000000000000 R15: ffff88043ea82800 FS: 00007ff646534700(0000) GS:ffff880447000000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000380 CR3: 000000043e9bf000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: ffffffff8100aba4 0000000000000092 ffff88043d21fe48 ffff88043ea82800 00000000000000ff ffff88043d21fe98 0000000000000000 ffff88043d21fe60 ffffffffa05102b4 0000000000000000 ffff88043d21fe70 ffffffffa05102ec Call Trace: [<ffffffff8100aba4>] ? native_sched_clock+0x24/0x80 [<ffffffffa05102b4>] loop_remove+0x14/0x40 [loop] [<ffffffffa05102ec>] loop_exit_cb+0xc/0x10 [loop] [<ffffffff81217b74>] idr_for_each+0x104/0x190 [<ffffffffa05102e0>] ? loop_remove+0x40/0x40 [loop] [<ffffffff8109adc5>] ? trace_hardirqs_on_caller+0x105/0x1d0 [<ffffffffa05135dc>] loop_exit+0x34/0xa58 [loop] [<ffffffff810a98ea>] SyS_delete_module+0x13a/0x260 [<ffffffff81221d5e>] ? trace_hardirqs_on_thunk+0x3a/0x3f [<ffffffff813cff16>] system_call_fastpath+0x1a/0x1f Code: f0 4c 8b 6d f8 c9 c3 66 66 2e 0f 1f 84 00 00 00 00 00 55 48 89 e5 41 56 41 55 4c 8d af 80 00 00 00 41 54 53 48 89 fb 48 83 ec 18 <48> 83 bf 80 03 00 00 00 74 4d e8 98 fe ff ff 31 f6 48 c7 c7 20 RIP [<ffffffff812057c9>] del_gendisk+0x19/0x2d0 RSP <ffff88043d21fe10> CR2: 0000000000000380 ---[ end trace 64ec069ec70f1309 ]--- Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: stable@kernel.org # 3.1+ Signed-off-by: Jens Axboe <axboe@kernel.dk>
2013-10-14 23:12:24 +07:00
out_free_idr:
idr_remove(&loop_index_idr, i);
out_free_dev:
kfree(lo);
out:
return err;
}
static void loop_remove(struct loop_device *lo)
{
blk_cleanup_queue(lo->lo_queue);
block: destroy bdi before blockdev is unregistered. Because of the peculiar way that md devices are created (automatically when the device node is opened), a new device can be created and registered immediately after the blk_unregister_region(disk_devt(disk), disk->minors); call in del_gendisk(). Therefore it is important that all visible artifacts of the previous device are removed before this call. In particular, the 'bdi'. Since: commit c4db59d31e39ea067c32163ac961e9c80198fd37 Author: Christoph Hellwig <hch@lst.de> fs: don't reassign dirty inodes to default_backing_dev_info moved the device_unregister(bdi->dev); call from bdi_unregister() to bdi_destroy() it has been quite easy to lose a race and have a new (e.g.) "md127" be created after the blk_unregister_region() call and before bdi_destroy() is ultimately called by the final 'put_disk', which must come after del_gendisk(). The new device finds that the bdi name is already registered in sysfs and complains > [ 9627.630029] WARNING: CPU: 18 PID: 3330 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x5a/0x70() > [ 9627.630032] sysfs: cannot create duplicate filename '/devices/virtual/bdi/9:127' We can fix this by moving the bdi_destroy() call out of blk_release_queue() (which can happen very late when a refcount reaches zero) and into blk_cleanup_queue() - which happens exactly when the md device driver calls it. Then it is only necessary for md to call blk_cleanup_queue() before del_gendisk(). As loop.c devices are also created on demand by opening the device node, we make the same change there. Fixes: c4db59d31e39ea067c32163ac961e9c80198fd37 Reported-by: Azat Khuzhin <a3at.mail@gmail.com> Cc: Christoph Hellwig <hch@lst.de> Cc: stable@vger.kernel.org (v4.0) Signed-off-by: NeilBrown <neilb@suse.de> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-04-27 11:12:22 +07:00
del_gendisk(lo->lo_disk);
block: loop: improve performance via blk-mq The conversion is a bit straightforward, and use work queue to dispatch requests of loop block, and one big change is that requests is submitted to backend file/device concurrently with work queue, so throughput may get improved much. Given write requests over same file are often run exclusively, so don't handle them concurrently for avoiding extra context switch cost, possible lock contention and work schedule cost. Also with blk-mq, there is opportunity to get loop I/O merged before submitting to backend file/device. In the following test: - base: v3.19-rc2-2041231 - loop over file in ext4 file system on SSD disk - bs: 4k, libaio, io depth: 64, O_DIRECT, num of jobs: 1 - throughput: IOPS ------------------------------------------------------ | | base | base with loop-mq | delta | ------------------------------------------------------ | randread | 1740 | 25318 | +1355%| ------------------------------------------------------ | read | 42196 | 51771 | +22.6%| ----------------------------------------------------- | randwrite | 35709 | 34624 | -3% | ----------------------------------------------------- | write | 39137 | 40326 | +3% | ----------------------------------------------------- So loop-mq can improve throughput for both read and randread, meantime, performance of write and randwrite isn't hurted basically. Another benefit is that loop driver code gets simplified much after blk-mq conversion, and the patch can be thought as cleanup too. Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2014-12-31 20:22:57 +07:00
blk_mq_free_tag_set(&lo->tag_set);
put_disk(lo->lo_disk);
kfree(lo);
}
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
static int find_free_cb(int id, void *ptr, void *data)
{
struct loop_device *lo = ptr;
struct loop_device **l = data;
if (lo->lo_state == Lo_unbound) {
*l = lo;
return 1;
}
return 0;
}
static int loop_lookup(struct loop_device **l, int i)
{
struct loop_device *lo;
int ret = -ENODEV;
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
if (i < 0) {
int err;
err = idr_for_each(&loop_index_idr, &find_free_cb, &lo);
if (err == 1) {
*l = lo;
ret = lo->lo_number;
}
goto out;
}
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
/* lookup and return a specific i */
lo = idr_find(&loop_index_idr, i);
if (lo) {
*l = lo;
ret = lo->lo_number;
}
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
out:
return ret;
}
static struct kobject *loop_probe(dev_t dev, int *part, void *data)
{
struct loop_device *lo;
struct kobject *kobj;
int err;
mutex_lock(&loop_index_mutex);
err = loop_lookup(&lo, MINOR(dev) >> part_shift);
if (err < 0)
err = loop_add(&lo, MINOR(dev) >> part_shift);
if (err < 0)
block: fix a probe argument to blk_register_region The probe function is supposed to return NULL on failure (as we can see in kobj_lookup: kobj = probe(dev, index, data); ... if (kobj) return kobj; However, in loop and brd, it returns negative error from ERR_PTR. This causes a crash if we simulate disk allocation failure and run less -f /dev/loop0 because the negative number is interpreted as a pointer: BUG: unable to handle kernel NULL pointer dereference at 00000000000002b4 IP: [<ffffffff8118b188>] __blkdev_get+0x28/0x450 PGD 23c677067 PUD 23d6d1067 PMD 0 Oops: 0000 [#1] PREEMPT SMP Modules linked in: loop hpfs nvidia(PO) ip6table_filter ip6_tables uvesafb cfbcopyarea cfbimgblt cfbfillrect fbcon font bitblit fbcon_rotate fbcon_cw fbcon_ud fbcon_ccw softcursor fb fbdev msr ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_conntrack_ipv4 nf_defrag_ipv4 xt_state ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp llc tun ipv6 cpufreq_stats cpufreq_ondemand cpufreq_userspace cpufreq_powersave cpufreq_conservative hid_generic spadfs usbhid hid fuse raid0 snd_usb_audio snd_pcm_oss snd_mixer_oss md_mod snd_pcm snd_timer snd_page_alloc snd_hwdep snd_usbmidi_lib dmi_sysfs snd_rawmidi nf_nat_ftp nf_nat nf_conntrack_ftp nf_conntrack snd soundcore lm85 hwmon_vid ohci_hcd ehci_pci ehci_hcd serverworks sata_svw libata acpi_cpufreq freq_table mperf ide_core usbcore kvm_amd kvm tg3 i2c_piix4 libphy microcode e100 usb_common ptp skge i2c_core pcspkr k10temp evdev floppy hwmon pps_core mii rtc_cmos button processor unix [last unloaded: nvidia] CPU: 1 PID: 6831 Comm: less Tainted: P W O 3.10.15-devel #18 Hardware name: empty empty/S3992-E, BIOS 'V1.06 ' 06/09/2009 task: ffff880203cc6bc0 ti: ffff88023e47c000 task.ti: ffff88023e47c000 RIP: 0010:[<ffffffff8118b188>] [<ffffffff8118b188>] __blkdev_get+0x28/0x450 RSP: 0018:ffff88023e47dbd8 EFLAGS: 00010286 RAX: ffffffffffffff74 RBX: ffffffffffffff74 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000001 RBP: ffff88023e47dc18 R08: 0000000000000002 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000000 R12: ffff88023f519658 R13: ffffffff8118c300 R14: 0000000000000000 R15: ffff88023f519640 FS: 00007f2070bf7700(0000) GS:ffff880247400000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000000002b4 CR3: 000000023da1d000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: 0000000000000002 0000001d00000000 000000003e47dc50 ffff88023f519640 ffff88043d5bb668 ffffffff8118c300 ffff88023d683550 ffff88023e47de60 ffff88023e47dc98 ffffffff8118c10d 0000001d81605698 0000000000000292 Call Trace: [<ffffffff8118c300>] ? blkdev_get_by_dev+0x60/0x60 [<ffffffff8118c10d>] blkdev_get+0x1dd/0x370 [<ffffffff8118c300>] ? blkdev_get_by_dev+0x60/0x60 [<ffffffff813cea6c>] ? _raw_spin_unlock+0x2c/0x50 [<ffffffff8118c300>] ? blkdev_get_by_dev+0x60/0x60 [<ffffffff8118c365>] blkdev_open+0x65/0x80 [<ffffffff8114d12e>] do_dentry_open.isra.18+0x23e/0x2f0 [<ffffffff8114d214>] finish_open+0x34/0x50 [<ffffffff8115e122>] do_last.isra.62+0x2d2/0xc50 [<ffffffff8115eb58>] path_openat.isra.63+0xb8/0x4d0 [<ffffffff81115a8e>] ? might_fault+0x4e/0xa0 [<ffffffff8115f4f0>] do_filp_open+0x40/0x90 [<ffffffff813cea6c>] ? _raw_spin_unlock+0x2c/0x50 [<ffffffff8116db85>] ? __alloc_fd+0xa5/0x1f0 [<ffffffff8114e45f>] do_sys_open+0xef/0x1d0 [<ffffffff8114e559>] SyS_open+0x19/0x20 [<ffffffff813cff16>] system_call_fastpath+0x1a/0x1f Code: 44 00 00 55 48 89 e5 41 57 49 89 ff 41 56 41 89 d6 41 55 41 54 4c 8d 67 18 53 48 83 ec 18 89 75 cc e9 f2 00 00 00 0f 1f 44 00 00 <48> 8b 80 40 03 00 00 48 89 df 4c 8b 68 58 e8 d5 a4 07 00 44 89 RIP [<ffffffff8118b188>] __blkdev_get+0x28/0x450 RSP <ffff88023e47dbd8> CR2: 00000000000002b4 ---[ end trace bb7f32dbf02398dc ]--- The brd change should be backported to stable kernels starting with 2.6.25. The loop change should be backported to stable kernels starting with 2.6.22. Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: stable@kernel.org # 2.6.22+ Signed-off-by: Jens Axboe <axboe@kernel.dk>
2013-10-14 23:13:24 +07:00
kobj = NULL;
else
kobj = get_disk(lo->lo_disk);
mutex_unlock(&loop_index_mutex);
*part = 0;
return kobj;
}
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
static long loop_control_ioctl(struct file *file, unsigned int cmd,
unsigned long parm)
{
struct loop_device *lo;
int ret = -ENOSYS;
mutex_lock(&loop_index_mutex);
switch (cmd) {
case LOOP_CTL_ADD:
ret = loop_lookup(&lo, parm);
if (ret >= 0) {
ret = -EEXIST;
break;
}
ret = loop_add(&lo, parm);
break;
case LOOP_CTL_REMOVE:
ret = loop_lookup(&lo, parm);
if (ret < 0)
break;
mutex_lock(&lo->lo_ctl_mutex);
if (lo->lo_state != Lo_unbound) {
ret = -EBUSY;
mutex_unlock(&lo->lo_ctl_mutex);
break;
}
if (atomic_read(&lo->lo_refcnt) > 0) {
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
ret = -EBUSY;
mutex_unlock(&lo->lo_ctl_mutex);
break;
}
lo->lo_disk->private_data = NULL;
mutex_unlock(&lo->lo_ctl_mutex);
idr_remove(&loop_index_idr, lo->lo_number);
loop_remove(lo);
break;
case LOOP_CTL_GET_FREE:
ret = loop_lookup(&lo, -1);
if (ret >= 0)
break;
ret = loop_add(&lo, -1);
}
mutex_unlock(&loop_index_mutex);
return ret;
}
static const struct file_operations loop_ctl_fops = {
.open = nonseekable_open,
.unlocked_ioctl = loop_control_ioctl,
.compat_ioctl = loop_control_ioctl,
.owner = THIS_MODULE,
.llseek = noop_llseek,
};
static struct miscdevice loop_misc = {
.minor = LOOP_CTRL_MINOR,
.name = "loop-control",
.fops = &loop_ctl_fops,
};
MODULE_ALIAS_MISCDEV(LOOP_CTRL_MINOR);
MODULE_ALIAS("devname:loop-control");
static int __init loop_init(void)
{
int i, nr;
unsigned long range;
struct loop_device *lo;
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
int err;
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
err = misc_register(&loop_misc);
if (err < 0)
return err;
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
part_shift = 0;
if (max_part > 0) {
loop: manage partitions in disk image This patch allows to use loop device with partitionned disk image. Original behavior of loop is not modified. A new parameter is introduced to define how many partition we want to be able to manage per loop device. This parameter is "max_part". For instance, to manage 63 partitions / loop device, we will do: # modprobe loop max_part=63 # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 And to attach a raw partitionned disk image, the original losetup is used: # losetup -f etch.img # ls -l /dev/loop?* brw-rw---- 1 root disk 7, 0 2008-03-05 14:55 /dev/loop0 brw-rw---- 1 root disk 7, 1 2008-03-05 14:57 /dev/loop0p1 brw-rw---- 1 root disk 7, 2 2008-03-05 14:57 /dev/loop0p2 brw-rw---- 1 root disk 7, 5 2008-03-05 14:57 /dev/loop0p5 brw-rw---- 1 root disk 7, 64 2008-03-05 14:55 /dev/loop1 brw-rw---- 1 root disk 7, 128 2008-03-05 14:55 /dev/loop2 brw-rw---- 1 root disk 7, 192 2008-03-05 14:55 /dev/loop3 brw-rw---- 1 root disk 7, 256 2008-03-05 14:55 /dev/loop4 brw-rw---- 1 root disk 7, 320 2008-03-05 14:55 /dev/loop5 brw-rw---- 1 root disk 7, 384 2008-03-05 14:55 /dev/loop6 brw-rw---- 1 root disk 7, 448 2008-03-05 14:55 /dev/loop7 # mount /dev/loop0p1 /mnt # ls /mnt bench cdrom home lib mnt root srv usr bin dev initrd lost+found opt sbin sys var boot etc initrd.img media proc selinux tmp vmlinuz # umount /mnt # losetup -d /dev/loop0 Of course, the same behavior can be done using kpartx on a loop device, but modifying loop avoids to stack several layers of block device (loop + device mapper), this is a very light modification (40% of modifications are to manage the new parameter). Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2008-03-26 18:11:53 +07:00
part_shift = fls(max_part);
/*
* Adjust max_part according to part_shift as it is exported
* to user space so that user can decide correct minor number
* if [s]he want to create more devices.
*
* Note that -1 is required because partition 0 is reserved
* for the whole disk.
*/
max_part = (1UL << part_shift) - 1;
}
if ((1UL << part_shift) > DISK_MAX_PARTS) {
err = -EINVAL;
goto misc_out;
}
loop: limit 'max_part' module param to DISK_MAX_PARTS The 'max_part' parameter controls the number of maximum partition a loop block device can have. However if a user specifies very large value it would exceed the limitation of device minor number and can cause a kernel panic (or, at least, produce invalid device nodes in some cases). On my desktop system, following command kills the kernel. On qemu, it triggers similar oops but the kernel was alive: $ sudo modprobe loop max_part0000 ------------[ cut here ]------------ kernel BUG at /media/Linux_Data/project/linux/fs/sysfs/group.c:65! invalid opcode: 0000 [#1] SMP last sysfs file: CPU 0 Modules linked in: loop(+) Pid: 43, comm: insmod Tainted: G W 2.6.39-qemu+ #155 Bochs Bochs RIP: 0010:[<ffffffff8113ce61>] [<ffffffff8113ce61>] internal_create_group= +0x2a/0x170 RSP: 0018:ffff880007b3fde8 EFLAGS: 00000246 RAX: 00000000ffffffef RBX: ffff880007b3d878 RCX: 00000000000007b4 RDX: ffffffff8152da50 RSI: 0000000000000000 RDI: ffff880007b3d878 RBP: ffff880007b3fe38 R08: ffff880007b3fde8 R09: 0000000000000000 R10: ffff88000783b4a8 R11: ffff880007b3d878 R12: ffffffff8152da50 R13: ffff880007b3d868 R14: 0000000000000000 R15: ffff880007b3d800 FS: 0000000002137880(0063) GS:ffff880007c00000(0000) knlGS:00000000000000= 00 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000422680 CR3: 0000000007b50000 CR4: 00000000000006b0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 0000000000000000 DR7: 0000000000000000 Process insmod (pid: 43, threadinfo ffff880007b3e000, task ffff880007afb9c= 0) Stack: ffff880007b3fe58 ffffffff811e66dd ffff880007b3fe58 ffffffff811e570b 0000000000000010 ffff880007b3d800 ffff880007a7b390 ffff880007b3d868 0000000000400920 ffff880007b3d800 ffff880007b3fe48 ffffffff8113cfc8 Call Trace: [<ffffffff811e66dd>] ? device_add+0x4bc/0x5af [<ffffffff811e570b>] ? dev_set_name+0x3c/0x3e [<ffffffff8113cfc8>] sysfs_create_group+0xe/0x12 [<ffffffff810b420e>] blk_trace_init_sysfs+0x14/0x16 [<ffffffff8116a090>] blk_register_queue+0x47/0xf7 [<ffffffff8116f527>] add_disk+0xdf/0x290 [<ffffffffa00060eb>] loop_init+0xeb/0x1b8 [loop] [<ffffffffa0006000>] ? 0xffffffffa0005fff [<ffffffff8100020a>] do_one_initcall+0x7a/0x12e [<ffffffff81096804>] sys_init_module+0x9c/0x1e0 [<ffffffff813329bb>] system_call_fastpath+0x16/0x1b Code: c3 55 48 89 e5 41 57 41 56 41 89 f6 41 55 41 54 49 89 d4 53 48 89 fb= 48 83 ec 28 48 85 ff 74 0b 85 f6 75 0b 48 83 7f 30 00 75 14 <0f> 0b eb fe = 48 83 7f 30 00 b9 ea ff ff ff 0f 84 18 01 00 00 49 RIP [<ffffffff8113ce61>] internal_create_group+0x2a/0x170 RSP <ffff880007b3fde8> ---[ end trace a123eb592043acad ]--- Signed-off-by: Namhyung Kim <namhyung@gmail.com> Cc: Laurent Vivier <Laurent.Vivier@bull.net> Cc: stable@kernel.org Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-05-24 21:48:54 +07:00
if (max_loop > 1UL << (MINORBITS - part_shift)) {
err = -EINVAL;
goto misc_out;
}
/*
* If max_loop is specified, create that many devices upfront.
* This also becomes a hard limit. If max_loop is not specified,
* create CONFIG_BLK_DEV_LOOP_MIN_COUNT loop devices at module
* init time. Loop devices can be requested on-demand with the
* /dev/loop-control interface, or be instantiated by accessing
* a 'dead' device node.
*/
if (max_loop) {
nr = max_loop;
loop: handle on-demand devices correctly When finding or allocating a loop device, loop_probe() did not take partition numbers into account so that it can result to a different device. Consider following example: $ sudo modprobe loop max_part=15 $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 $ sudo mknod /dev/loop8 b 7 128 $ sudo losetup /dev/loop8 ~/temp/disk-with-3-parts.img $ sudo losetup -a /dev/loop128: [0805]:278201 (/home/namhyung/temp/disk-with-3-parts.img) $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 2048 2011-05-24 22:18 /dev/loop128 brw-rw---- 1 root disk 7, 2049 2011-05-24 22:18 /dev/loop128p1 brw-rw---- 1 root disk 7, 2050 2011-05-24 22:18 /dev/loop128p2 brw-rw---- 1 root disk 7, 2051 2011-05-24 22:18 /dev/loop128p3 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 brw-r--r-- 1 root root 7, 128 2011-05-24 22:17 /dev/loop8 After this patch, /dev/loop8 - instead of /dev/loop128 - was accessed correctly. In addition, 'range' passed to blk_register_region() should include all range of dev_t that LOOP_MAJOR can address. It does not need to be limited by partition numbers unless 'max_loop' param was specified. Signed-off-by: Namhyung Kim <namhyung@gmail.com> Cc: Laurent Vivier <Laurent.Vivier@bull.net> Cc: stable@kernel.org Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-05-24 21:48:55 +07:00
range = max_loop << part_shift;
} else {
nr = CONFIG_BLK_DEV_LOOP_MIN_COUNT;
loop: handle on-demand devices correctly When finding or allocating a loop device, loop_probe() did not take partition numbers into account so that it can result to a different device. Consider following example: $ sudo modprobe loop max_part=15 $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 $ sudo mknod /dev/loop8 b 7 128 $ sudo losetup /dev/loop8 ~/temp/disk-with-3-parts.img $ sudo losetup -a /dev/loop128: [0805]:278201 (/home/namhyung/temp/disk-with-3-parts.img) $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 2048 2011-05-24 22:18 /dev/loop128 brw-rw---- 1 root disk 7, 2049 2011-05-24 22:18 /dev/loop128p1 brw-rw---- 1 root disk 7, 2050 2011-05-24 22:18 /dev/loop128p2 brw-rw---- 1 root disk 7, 2051 2011-05-24 22:18 /dev/loop128p3 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 brw-r--r-- 1 root root 7, 128 2011-05-24 22:17 /dev/loop8 After this patch, /dev/loop8 - instead of /dev/loop128 - was accessed correctly. In addition, 'range' passed to blk_register_region() should include all range of dev_t that LOOP_MAJOR can address. It does not need to be limited by partition numbers unless 'max_loop' param was specified. Signed-off-by: Namhyung Kim <namhyung@gmail.com> Cc: Laurent Vivier <Laurent.Vivier@bull.net> Cc: stable@kernel.org Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-05-24 21:48:55 +07:00
range = 1UL << MINORBITS;
}
if (register_blkdev(LOOP_MAJOR, "loop")) {
err = -EIO;
goto misc_out;
}
blk_register_region(MKDEV(LOOP_MAJOR, 0), range,
THIS_MODULE, loop_probe, NULL, NULL);
/* pre-create number of devices given by config or max_loop */
mutex_lock(&loop_index_mutex);
for (i = 0; i < nr; i++)
loop_add(&lo, i);
mutex_unlock(&loop_index_mutex);
printk(KERN_INFO "loop: module loaded\n");
return 0;
misc_out:
misc_deregister(&loop_misc);
return err;
}
static int loop_exit_cb(int id, void *ptr, void *data)
{
struct loop_device *lo = ptr;
loop_remove(lo);
return 0;
}
static void __exit loop_exit(void)
{
unsigned long range;
loop: handle on-demand devices correctly When finding or allocating a loop device, loop_probe() did not take partition numbers into account so that it can result to a different device. Consider following example: $ sudo modprobe loop max_part=15 $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 $ sudo mknod /dev/loop8 b 7 128 $ sudo losetup /dev/loop8 ~/temp/disk-with-3-parts.img $ sudo losetup -a /dev/loop128: [0805]:278201 (/home/namhyung/temp/disk-with-3-parts.img) $ ls -l /dev/loop* brw-rw---- 1 root disk 7, 0 2011-05-24 22:16 /dev/loop0 brw-rw---- 1 root disk 7, 16 2011-05-24 22:16 /dev/loop1 brw-rw---- 1 root disk 7, 2048 2011-05-24 22:18 /dev/loop128 brw-rw---- 1 root disk 7, 2049 2011-05-24 22:18 /dev/loop128p1 brw-rw---- 1 root disk 7, 2050 2011-05-24 22:18 /dev/loop128p2 brw-rw---- 1 root disk 7, 2051 2011-05-24 22:18 /dev/loop128p3 brw-rw---- 1 root disk 7, 32 2011-05-24 22:16 /dev/loop2 brw-rw---- 1 root disk 7, 48 2011-05-24 22:16 /dev/loop3 brw-rw---- 1 root disk 7, 64 2011-05-24 22:16 /dev/loop4 brw-rw---- 1 root disk 7, 80 2011-05-24 22:16 /dev/loop5 brw-rw---- 1 root disk 7, 96 2011-05-24 22:16 /dev/loop6 brw-rw---- 1 root disk 7, 112 2011-05-24 22:16 /dev/loop7 brw-r--r-- 1 root root 7, 128 2011-05-24 22:17 /dev/loop8 After this patch, /dev/loop8 - instead of /dev/loop128 - was accessed correctly. In addition, 'range' passed to blk_register_region() should include all range of dev_t that LOOP_MAJOR can address. It does not need to be limited by partition numbers unless 'max_loop' param was specified. Signed-off-by: Namhyung Kim <namhyung@gmail.com> Cc: Laurent Vivier <Laurent.Vivier@bull.net> Cc: stable@kernel.org Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-05-24 21:48:55 +07:00
range = max_loop ? max_loop << part_shift : 1UL << MINORBITS;
idr_for_each(&loop_index_idr, &loop_exit_cb, NULL);
idr_destroy(&loop_index_idr);
blk_unregister_region(MKDEV(LOOP_MAJOR, 0), range);
unregister_blkdev(LOOP_MAJOR, "loop");
loop: add management interface for on-demand device allocation Loop devices today have a fixed pre-allocated number of usually 8. The number can only be changed at module init time. To find a free device to use, /dev/loop%i needs to be scanned, and all devices need to be opened until a free one is possibly found. This adds a new /dev/loop-control device node, that allows to dynamically find or allocate a free device, and to add and remove loop devices from the running system: LOOP_CTL_ADD adds a specific device. Arg is the number of the device. It returns the device i or a negative error code. LOOP_CTL_REMOVE removes a specific device, Arg is the number the device. It returns the device i or a negative error code. LOOP_CTL_GET_FREE finds the next unbound device or allocates a new one. No arg is given. It returns the device i or a negative error code. The loop kernel module gets automatically loaded when /dev/loop-control is accessed the first time. The alias specified in the module, instructs udev to create this 'dead' device node, even when the module is not loaded. Example: cfd = open("/dev/loop-control", O_RDWR); # add a new specific loop device err = ioctl(cfd, LOOP_CTL_ADD, devnr); # remove a specific loop device err = ioctl(cfd, LOOP_CTL_REMOVE, devnr); # find or allocate a free loop device to use devnr = ioctl(cfd, LOOP_CTL_GET_FREE); sprintf(loopname, "/dev/loop%i", devnr); ffd = open("backing-file", O_RDWR); lfd = open(loopname, O_RDWR); err = ioctl(lfd, LOOP_SET_FD, ffd); Cc: Tejun Heo <tj@kernel.org> Cc: Karel Zak <kzak@redhat.com> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2011-08-01 03:08:04 +07:00
misc_deregister(&loop_misc);
}
module_init(loop_init);
module_exit(loop_exit);
#ifndef MODULE
static int __init max_loop_setup(char *str)
{
max_loop = simple_strtol(str, NULL, 0);
return 1;
}
__setup("max_loop=", max_loop_setup);
#endif