From 6050fa4c84cc93ae509f5105f585a429dffc5633 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Wed, 24 Nov 2021 19:47:40 +0900 Subject: loop: don't hold lo_mutex during __loop_clr_fd() syzbot is reporting circular locking problem at __loop_clr_fd() [1], for commit 87579e9b7d8dc36e ("loop: use worker per cgroup instead of kworker") is calling destroy_workqueue() with lo->lo_mutex held. Since all functions where lo->lo_state matters are already checking lo->lo_state with lo->lo_mutex held (in order to avoid racing with e.g. ioctl(LOOP_CTL_REMOVE)), and __loop_clr_fd() can be called from either ioctl(LOOP_CLR_FD) xor close(), lo->lo_state == Lo_rundown is considered as an exclusive lock for __loop_clr_fd(). Therefore, hold lo->lo_mutex inside __loop_clr_fd() only when asserting/updating lo->lo_state. Since ioctl(LOOP_CLR_FD) depends on lo->lo_state == Lo_bound, a valid lo->lo_backing_file must have been assigned by ioctl(LOOP_SET_FD) or ioctl(LOOP_CONFIGURE). Thus, we can remove lo->lo_backing_file test, and convert __loop_clr_fd() into a void function. Link: https://syzkaller.appspot.com/bug?extid=63614029dfb79abd4383 [1] Reported-by: syzbot Signed-off-by: Tetsuo Handa Reviewed-by: Christoph Hellwig Link: https://lore.kernel.org/r/8ebe3b2e-8975-7f26-0620-7144a3b8b8cd@i-love.sakura.ne.jp Signed-off-by: Jens Axboe --- drivers/block/loop.c | 55 +++++++++++++++++++++------------------------------- 1 file changed, 22 insertions(+), 33 deletions(-) (limited to 'drivers/block/loop.c') diff --git a/drivers/block/loop.c b/drivers/block/loop.c index 0954ea8cf9e3..ba76319b5544 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -1082,13 +1082,10 @@ out_putf: return error; } -static int __loop_clr_fd(struct loop_device *lo, bool release) +static void __loop_clr_fd(struct loop_device *lo, bool release) { - struct file *filp = NULL; + struct file *filp; gfp_t gfp = lo->old_gfp_mask; - int err = 0; - bool partscan = false; - int lo_number; struct loop_worker *pos, *worker; /* @@ -1103,17 +1100,14 @@ static int __loop_clr_fd(struct loop_device *lo, bool release) * became visible. */ + /* + * Since this function is called upon "ioctl(LOOP_CLR_FD)" xor "close() + * after ioctl(LOOP_CLR_FD)", it is a sign of something going wrong if + * lo->lo_state has changed while waiting for lo->lo_mutex. + */ mutex_lock(&lo->lo_mutex); - if (WARN_ON_ONCE(lo->lo_state != Lo_rundown)) { - err = -ENXIO; - goto out_unlock; - } - - filp = lo->lo_backing_file; - if (filp == NULL) { - err = -EINVAL; - goto out_unlock; - } + BUG_ON(lo->lo_state != Lo_rundown); + mutex_unlock(&lo->lo_mutex); if (test_bit(QUEUE_FLAG_WC, &lo->lo_queue->queue_flags)) blk_queue_write_cache(lo->lo_queue, false, false); @@ -1134,6 +1128,7 @@ static int __loop_clr_fd(struct loop_device *lo, bool release) del_timer_sync(&lo->timer); spin_lock_irq(&lo->lo_lock); + filp = lo->lo_backing_file; lo->lo_backing_file = NULL; spin_unlock_irq(&lo->lo_lock); @@ -1153,12 +1148,11 @@ static int __loop_clr_fd(struct loop_device *lo, bool release) module_put(THIS_MODULE); blk_mq_unfreeze_queue(lo->lo_queue); - partscan = lo->lo_flags & LO_FLAGS_PARTSCAN; - lo_number = lo->lo_number; disk_force_media_change(lo->lo_disk, DISK_EVENT_MEDIA_CHANGE); -out_unlock: - mutex_unlock(&lo->lo_mutex); - if (partscan) { + + if (lo->lo_flags & LO_FLAGS_PARTSCAN) { + int err; + /* * open_mutex has been held already in release path, so don't * acquire it if this function is called in such case. @@ -1174,24 +1168,20 @@ out_unlock: mutex_unlock(&lo->lo_disk->open_mutex); if (err) pr_warn("%s: partition scan of loop%d failed (rc=%d)\n", - __func__, lo_number, err); + __func__, lo->lo_number, err); /* Device is gone, no point in returning error */ - err = 0; } /* * lo->lo_state is set to Lo_unbound here after above partscan has - * finished. - * - * There cannot be anybody else entering __loop_clr_fd() as - * lo->lo_backing_file is already cleared and Lo_rundown state - * protects us from all the other places trying to change the 'lo' - * device. + * finished. There cannot be anybody else entering __loop_clr_fd() as + * Lo_rundown state protects us from all the other places trying to + * change the 'lo' device. */ - mutex_lock(&lo->lo_mutex); lo->lo_flags = 0; if (!part_shift) lo->lo_disk->flags |= GENHD_FL_NO_PART; + mutex_lock(&lo->lo_mutex); lo->lo_state = Lo_unbound; mutex_unlock(&lo->lo_mutex); @@ -1200,9 +1190,7 @@ out_unlock: * lo_mutex triggers a circular lock dependency possibility warning as * fput can take open_mutex which is usually taken before lo_mutex. */ - if (filp) - fput(filp); - return err; + fput(filp); } static int loop_clr_fd(struct loop_device *lo) @@ -1234,7 +1222,8 @@ static int loop_clr_fd(struct loop_device *lo) lo->lo_state = Lo_rundown; mutex_unlock(&lo->lo_mutex); - return __loop_clr_fd(lo, false); + __loop_clr_fd(lo, false); + return 0; } static int -- cgit v1.2.3 From 322c4293ecc58110227b49d7e47ae37b9b03566f Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Mon, 13 Dec 2021 21:55:27 +0900 Subject: loop: make autoclear operation asynchronous syzbot is reporting circular locking problem at __loop_clr_fd() [1], for commit 87579e9b7d8dc36e ("loop: use worker per cgroup instead of kworker") is calling destroy_workqueue() with disk->open_mutex held. This circular dependency cannot be broken unless we call __loop_clr_fd() without holding disk->open_mutex. Therefore, defer __loop_clr_fd() from lo_release() to a WQ context. Link: https://syzkaller.appspot.com/bug?extid=643e4ce4b6ad1347d372 [1] Reported-by: syzbot Suggested-by: Christoph Hellwig Cc: Jan Kara Tested-by: syzbot+643e4ce4b6ad1347d372@syzkaller.appspotmail.com Signed-off-by: Tetsuo Handa Reviewed-by: Christoph Hellwig Link: https://lore.kernel.org/r/1ed7df28-ebd6-71fb-70e5-1c2972e05ddb@i-love.sakura.ne.jp Signed-off-by: Jens Axboe --- drivers/block/loop.c | 65 +++++++++++++++++++++++++++++----------------------- 1 file changed, 36 insertions(+), 29 deletions(-) (limited to 'drivers/block/loop.c') diff --git a/drivers/block/loop.c b/drivers/block/loop.c index ba76319b5544..7f4ea06534c2 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -1082,7 +1082,7 @@ out_putf: return error; } -static void __loop_clr_fd(struct loop_device *lo, bool release) +static void __loop_clr_fd(struct loop_device *lo) { struct file *filp; gfp_t gfp = lo->old_gfp_mask; @@ -1144,8 +1144,6 @@ static void __loop_clr_fd(struct loop_device *lo, bool release) /* let user-space know about this change */ kobject_uevent(&disk_to_dev(lo->lo_disk)->kobj, KOBJ_CHANGE); mapping_set_gfp_mask(filp->f_mapping, gfp); - /* This is safe: open() is still holding a reference. */ - module_put(THIS_MODULE); blk_mq_unfreeze_queue(lo->lo_queue); disk_force_media_change(lo->lo_disk, DISK_EVENT_MEDIA_CHANGE); @@ -1153,44 +1151,52 @@ static void __loop_clr_fd(struct loop_device *lo, bool release) if (lo->lo_flags & LO_FLAGS_PARTSCAN) { int err; - /* - * open_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 (!release) - mutex_lock(&lo->lo_disk->open_mutex); + mutex_lock(&lo->lo_disk->open_mutex); err = bdev_disk_changed(lo->lo_disk, false); - if (!release) - mutex_unlock(&lo->lo_disk->open_mutex); + mutex_unlock(&lo->lo_disk->open_mutex); if (err) pr_warn("%s: partition scan of loop%d failed (rc=%d)\n", __func__, lo->lo_number, err); /* Device is gone, no point in returning error */ } - /* - * lo->lo_state is set to Lo_unbound here after above partscan has - * finished. There cannot be anybody else entering __loop_clr_fd() as - * Lo_rundown state protects us from all the other places trying to - * change the 'lo' device. - */ lo->lo_flags = 0; if (!part_shift) lo->lo_disk->flags |= GENHD_FL_NO_PART; + + fput(filp); +} + +static void loop_rundown_completed(struct loop_device *lo) +{ mutex_lock(&lo->lo_mutex); lo->lo_state = Lo_unbound; mutex_unlock(&lo->lo_mutex); + module_put(THIS_MODULE); +} - /* - * Need not hold lo_mutex to fput backing file. Calling fput holding - * lo_mutex triggers a circular lock dependency possibility warning as - * fput can take open_mutex which is usually taken before lo_mutex. - */ - fput(filp); +static void loop_rundown_workfn(struct work_struct *work) +{ + struct loop_device *lo = container_of(work, struct loop_device, + rundown_work); + struct block_device *bdev = lo->lo_device; + struct gendisk *disk = lo->lo_disk; + + __loop_clr_fd(lo); + kobject_put(&bdev->bd_device.kobj); + module_put(disk->fops->owner); + loop_rundown_completed(lo); +} + +static void loop_schedule_rundown(struct loop_device *lo) +{ + struct block_device *bdev = lo->lo_device; + struct gendisk *disk = lo->lo_disk; + + __module_get(disk->fops->owner); + kobject_get(&bdev->bd_device.kobj); + INIT_WORK(&lo->rundown_work, loop_rundown_workfn); + queue_work(system_long_wq, &lo->rundown_work); } static int loop_clr_fd(struct loop_device *lo) @@ -1222,7 +1228,8 @@ static int loop_clr_fd(struct loop_device *lo) lo->lo_state = Lo_rundown; mutex_unlock(&lo->lo_mutex); - __loop_clr_fd(lo, false); + __loop_clr_fd(lo); + loop_rundown_completed(lo); return 0; } @@ -1747,7 +1754,7 @@ static void lo_release(struct gendisk *disk, fmode_t mode) * In autoclear mode, stop the loop thread * and remove configuration after last close. */ - __loop_clr_fd(lo, true); + loop_schedule_rundown(lo); return; } else if (lo->lo_state == Lo_bound) { /* -- cgit v1.2.3