kernel/linux.git/drivers/mtd/mtd_blkdevs.c, branch v5.15.7

Merge tag 'mtd/for-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux

2021-09-05T17:50:12+00:00

Pull MTD updates from Miquel Raynal: "MTD changes: - blkdevs: - Simplify the refcounting in blktrans_{open, release} - Simplify blktrans_getgeo - Remove blktrans_ref_mutex - Simplify blktrans_dev_get - Use lockdep_assert_held - Don't hold del_mtd_blktrans_dev in blktrans_{open, release} - ftl: - Don't cast away the type when calling add_mtd_blktrans_dev - Don't cast away the type when calling add_mtd_blktrans_dev - Use container_of() rather than cast - Fix use-after-free - Add discard support - Allow use of MTD_RAM for testing purposes - concat: - Check _read, _write callbacks existence before assignment - Judge callback existence based on the master - maps: - Maps: remove dead MTD map driver for PMC-Sierra MSP boards - mtdblock: - Warn if added for a NAND device - Add comment about UBI block devices - Update old JFFS2 mention in Kconfig - partitions: - Redboot: convert to YAML NAND core changes: - Repair Miquel Raynal's email address in MAINTAINERS - Fix a couple of spelling mistakes in Kconfig - bbt: Skip bad blocks when searching for the BBT in NAND - Remove never changed ret variable Raw NAND changes: - cafe: Fix a resource leak in the error handling path of 'cafe_nand_probe()' - intel: Fix error handling in probe - omap: Fix kernel doc warning on 'calcuate' typo - gpmc: Fix the ECC bytes vs. OOB bytes equation SPI-NAND core changes: - Properly fill the OOB area. - Fix comment SPI-NAND drivers changes: - macronix: Add Quad support for serial NAND flash" * tag 'mtd/for-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux: (30 commits) mtd: rawnand: cafe: Fix a resource leak in the error handling path of 'cafe_nand_probe()' mtd_blkdevs: simplify the refcounting in blktrans_{open, release} mtd_blkdevs: simplify blktrans_getgeo mtd_blkdevs: remove blktrans_ref_mutex mtd_blkdevs: simplify blktrans_dev_get mtd/rfd_ftl: don't cast away the type when calling add_mtd_blktrans_dev mtd/ftl: don't cast away the type when calling add_mtd_blktrans_dev mtd_blkdevs: use lockdep_assert_held mtd_blkdevs: don't hold del_mtd_blktrans_dev in blktrans_{open, release} mtd: rawnand: intel: Fix error handling in probe mtd: mtdconcat: Check _read, _write callbacks existence before assignment mtd: mtdconcat: Judge callback existence based on the master mtd: maps: remove dead MTD map driver for PMC-Sierra MSP boards mtd: rfd_ftl: use container_of() rather than cast mtd: rfd_ftl: fix use-after-free mtd: rfd_ftl: add discard support mtd: rfd_ftl: allow use of MTD_RAM for testing purposes mtdblock: Warn if added for a NAND device mtd: spinand: macronix: Add Quad support for serial NAND flash mtdblock: Add comment about UBI block devices ...

mtd_blkdevs: simplify the refcounting in blktrans_{open, release}

2021-08-23T08:01:16+00:00

Always grab a reference to the mtd_blktrans_dev in ->open instead of just on the first open, and do away with the additional temporary references in ->open and ->release. Signed-off-by: Christoph Hellwig Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-9-hch@lst.de

mtd_blkdevs: simplify blktrans_getgeo

2021-08-23T08:01:13+00:00

No need to grab a mtd_blktrans_dev given that ->open already holds one and ->getgeo can only be called on an open disk. Signed-off-by: Christoph Hellwig Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-8-hch@lst.de

mtd_blkdevs: remove blktrans_ref_mutex

2021-08-23T08:01:11+00:00

blktrans_ref_mutex is not actually needed. The kref is serialized internally, and devnum assignment in add_mtd_blktrans_dev happens before the disk is added and thus any of the block_device_operations methods otherwise using it are called. It is also already serialized by the global mtd_table_mutex. Signed-off-by: Christoph Hellwig Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-7-hch@lst.de

mtd_blkdevs: simplify blktrans_dev_get

2021-08-23T08:01:09+00:00

->private_data is set before the disk is added and never cleared, so don't bother trying to handle a NULL pointer there. Signed-off-by: Christoph Hellwig Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-6-hch@lst.de

mtd_blkdevs: use lockdep_assert_held

2021-08-23T08:01:01+00:00

Use lockdep_assert_held to ensure mtd_table_mutex is held instead of mutex_trylock games. Signed-off-by: Christoph Hellwig Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-3-hch@lst.de

mtd_blkdevs: don't hold del_mtd_blktrans_dev in blktrans_{open, release}

2021-08-23T08:00:59+00:00

There is nothing that this protects against except for slightly reducing the window when new opens can appear just before calling del_gendisk. Reported-by: Guenter Roeck Signed-off-by: Christoph Hellwig Tested-by: Guenter Roeck Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210823073359.705281-2-hch@lst.de

mtd: fix lock hierarchy in deregister_mtd_blktrans

2021-08-06T19:44:16+00:00

There is a lock hierarchy of major_names_lock --> mtd_table_mutex. One existing chain is as follows: 1. major_names_lock --> loop_ctl_mutex (when blk_request_module calls loop_probe) 2. loop_ctl_mutex --> bdev->bd_mutex (when loop_control_ioctl calls loop_remove, which then calls del_gendisk) 3. bdev->bd_mutex --> mtd_table_mutex (when blkdev_get_by_dev calls __blkdev_get, which then calls blktrans_open) Since unregister_blkdev grabs the major_names_lock, we need to call it outside the critical section for mtd_table_mutex, otherwise we invert the lock hierarchy. Reported-by: Hillf Danton Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210717100719.728829-1-desmondcheongzx@gmail.com

mtd: break circular locks in register_mtd_blktrans

2021-07-15T22:49:17+00:00

Syzbot reported a circular locking dependency: https://syzkaller.appspot.com/bug?id=7bd106c28e846d1023d4ca915718b1a0905444cb This happens because of the following lock dependencies: 1. loop_ctl_mutex -> bdev->bd_mutex (when loop_control_ioctl calls loop_remove, which then calls del_gendisk; this also happens in loop_exit which eventually calls loop_remove) 2. bdev->bd_mutex -> mtd_table_mutex (when blkdev_get_by_dev calls __blkdev_get, which then calls blktrans_open) 3. mtd_table_mutex -> major_names_lock (when register_mtd_blktrans calls __register_blkdev) 4. major_names_lock -> loop_ctl_mutex (when blk_request_module calls loop_probe) Hence there's an overall dependency of: loop_ctl_mutex ----------> bdev->bd_mutex ^ | | | | v major_names_lock <--------- mtd_table_mutex We can break this circular dependency by holding mtd_table_mutex only for the required critical section in register_mtd_blktrans. This avoids the mtd_table_mutex -> major_names_lock dependency. Reported-and-tested-by: syzbot+6a8a0d93c91e8fbf2e80@syzkaller.appspotmail.com Co-developed-by: Christoph Hellwig Signed-off-by: Christoph Hellwig Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210617160904.570111-1-desmondcheongzx@gmail.com

mtd: mtd_blkdevs: Initialize rq.limits.discard_granularity

2021-07-15T22:49:13+00:00

Since commit b35fd7422c2f8("block: check queue's limits.discard_granularity in __blkdev_issue_discard()") checks rq.limits.discard_granularity in __blkdev_issue_discard(), we may get following warnings on formatted ftl: WARNING: CPU: 2 PID: 7313 at block/blk-lib.c:51 __blkdev_issue_discard+0x2a7/0x390 Reproducer: 1. ftl_format /dev/mtd0 2. modprobe ftl 3. mkfs.vfat /dev/ftla 4. mount -odiscard /dev/ftla temp 5. dd if=/dev/zero of=temp/tst bs=1M count=10 oflag=direct 6. dd if=/dev/zero of=temp/tst bs=1M count=10 oflag=direct Fix it by initializing rq.limits.discard_granularity if device supports discard operation. Signed-off-by: Zhihao Cheng Signed-off-by: Miquel Raynal Link: https://lore.kernel.org/linux-mtd/20210615093905.3473709-1-chengzhihao1@huawei.com