user/sven/linux.git/block, branch v3.3.4

Merge branch 'for-linus' of git://git.kernel.dk/linux-block

2012-03-15T00:16:45Z

Pull block fixes from Jens Axboe: "Been sitting on this for a while, but lets get this out the door. This fixes various important bugs for 3.3 final, along with a few more trivial ones. Please pull!" * 'for-linus' of git://git.kernel.dk/linux-block: block: fix ioc leak in put_io_context block, sx8: fix pointer math issue getting fw version Block: use a freezable workqueue for disk-event polling drivers/block/DAC960: fix -Wuninitialized warning drivers/block/DAC960: fix DAC960_V2_IOCTL_Opcode_T -Wenum-compare warning block: fix __blkdev_get and add_disk race condition block: Fix setting bio flags in drivers (sd_dif/floppy) block: Fix NULL pointer dereference in sd_revalidate_disk block: exit_io_context() should call elevator_exit_icq_fn() block: simplify ioc_release_fn() block: replace icq->changed with icq->flags

block: fix ioc leak in put_io_context

2012-03-14T14:34:48Z

When put_io_context is called, if ioc->icq_list is empty and refcount is 1, kernel will not free the ioc. This is caught by following kmemleak: unreferenced object 0xffff880036349fe0 (size 216): comm "sh", pid 2137, jiffies 4294931140 (age 290579.412s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 01 00 01 00 ad 4e ad de ff ff ff ff 00 00 00 00 .....N.......... backtrace: [] kmemleak_alloc+0x26/0x50 [] kmem_cache_alloc_node+0x1cc/0x2a0 [] create_io_context_slowpath+0x27/0x130 [] get_task_io_context+0xbb/0xf0 [] copy_process+0x188e/0x18b0 [] do_fork+0x11b/0x420 [] sys_clone+0x28/0x30 [] stub_clone+0x13/0x20 [] 0xffffffffffffffff ioc should be freed if ioc->icq_list is empty. Signed-off-by: Xiaotian Feng Acked-by: Vivek Goyal Acked-by: Tejun Heo Signed-off-by: Jens Axboe

Block: use a freezable workqueue for disk-event polling

2012-03-02T09:51:00Z

This patch (as1519) fixes a bug in the block layer's disk-events polling. The polling is done by a work routine queued on the system_nrt_wq workqueue. Since that workqueue isn't freezable, the polling continues even in the middle of a system sleep transition. Obviously, polling a suspended drive for media changes and such isn't a good thing to do; in the case of USB mass-storage devices it can lead to real problems requiring device resets and even re-enumeration. The patch fixes things by creating a new system-wide, non-reentrant, freezable workqueue and using it for disk-events polling. Signed-off-by: Alan Stern CC: Acked-by: Tejun Heo Acked-by: Rafael J. Wysocki Signed-off-by: Jens Axboe

block: fix __blkdev_get and add_disk race condition

2012-03-02T09:44:17Z

The following situation might occur: __blkdev_get: add_disk: register_disk() get_gendisk() disk_block_events() disk->ev == NULL disk_add_events() __disk_unblock_events() disk->ev != NULL --ev->block Then we unblock events, when they are suppose to be blocked. This can trigger events related block/genhd.c warnings, but also can crash in sd_check_events() or other places. I'm able to reproduce crashes with the following scripts (with connected usb dongle as sdb disk). DEV=/dev/sdb ENABLE=/sys/bus/usb/devices/1-2/bConfigurationValue function stop_me() { for i in `jobs -p` ; do kill $i 2> /dev/null ; done exit } trap stop_me SIGHUP SIGINT SIGTERM for ((i = 0; i < 10; i++)) ; do while true; do fdisk -l $DEV 2>&1 > /dev/null ; done & done while true ; do echo 1 > $ENABLE sleep 1 echo 0 > $ENABLE done I use the script to verify patch fixing oops in sd_revalidate_disk http://marc.info/?l=linux-scsi&m=132935572512352&w=2 Without Jun'ichi Nomura patch titled "Fix NULL pointer dereference in sd_revalidate_disk" or this one, script easily crash kernel within a few seconds. With both patches applied I do not observe crash. Unfortunately after some time (dozen of minutes), script will hung in: [ 1563.906432] [] schedule_timeout_uninterruptible+0x15/0x20 [ 1563.906437] [] msleep+0x15/0x20 [ 1563.906443] [] blk_drain_queue+0x32/0xd0 [ 1563.906447] [] blk_cleanup_queue+0xd0/0x170 [ 1563.906454] [] scsi_free_queue+0x3f/0x60 [ 1563.906459] [] __scsi_remove_device+0x6e/0xb0 [ 1563.906463] [] scsi_forget_host+0x4f/0x60 [ 1563.906468] [] scsi_remove_host+0x5a/0xf0 [ 1563.906482] [] quiesce_and_remove_host+0x5b/0xa0 [usb_storage] [ 1563.906490] [] usb_stor_disconnect+0x13/0x20 [usb_storage] Anyway I think this patch is some step forward. As drawback, I do not teardown on sysfs file create error, because I do not know how to nullify disk->ev (since it can be used). However add_disk error handling practically does not exist too, and things will work without this sysfs file, except events will not be exported to user space. Signed-off-by: Stanislaw Gruszka Acked-by: Tejun Heo Cc: stable@kernel.org Signed-off-by: Jens Axboe

block: Fix NULL pointer dereference in sd_revalidate_disk

2012-03-02T09:38:33Z

Since 2.6.39 (1196f8b), when a driver returns -ENOMEDIUM for open(), __blkdev_get() calls rescan_partitions() to remove in-kernel partition structures and raise KOBJ_CHANGE uevent. However it ends up calling driver's revalidate_disk without open and could cause oops. In the case of SCSI: process A process B ---------------------------------------------- sys_open __blkdev_get sd_open returns -ENOMEDIUM scsi_remove_device rescan_partitions sd_revalidate_disk Oopses are reported here: http://marc.info/?l=linux-scsi&m=132388619710052 This patch separates the partition invalidation from rescan_partitions() and use it for -ENOMEDIUM case. Reported-by: Huajun Li Signed-off-by: Jun'ichi Nomura Acked-by: Tejun Heo Cc: stable@kernel.org Signed-off-by: Jens Axboe

LDM: Fix reassembly of extended VBLKs.

2012-02-24T09:37:42Z

From: Ben Hutchings Extended VBLKs (those larger than the preset VBLK size) are divided into fragments, each with its own VBLK header. Our LDM implementation generally assumes that each VBLK is contiguous in memory, so these fragments must be assembled before further processing. Currently the reassembly seems to be done quite wrongly - no VBLK header is copied into the contiguous buffer, and the length of the header is subtracted twice from each fragment. Also the total length of the reassembled VBLK is calculated incorrectly. Signed-off-by: Ben Hutchings Signed-off-by: Anton Altaparmakov

block: exit_io_context() should call elevator_exit_icq_fn()

2012-02-15T08:45:53Z

While updating locking, b2efa05265 "block, cfq: unlink cfq_io_context's immediately" moved elevator_exit_icq_fn() invocation from exit_io_context() to the final ioc put. While this doesn't cause catastrophic failure, it effectively removes task exit notification to elevator and cause noticeable IO performance degradation with CFQ. On task exit, CFQ used to immediately expire the slice if it was being used by the exiting task as no more IO would be issued by the task; however, after b2efa05265, the notification is lost and disk could sit idle needlessly, leading to noticeable IO performance degradation for certain workloads. This patch renames ioc_exit_icq() to ioc_destroy_icq(), separates elevator_exit_icq_fn() invocation into ioc_exit_icq() and invokes it from exit_io_context(). ICQ_EXITED flag is added to avoid invoking the callback more than once for the same icq. Walking icq_list from ioc side and invoking elevator callback requires reverse double locking. This may be better implemented using RCU; unfortunately, using RCU isn't trivial. e.g. RCU protection would need to cover request_queue and queue_lock switch on cleanup makes grabbing queue_lock from RCU unsafe. Reverse double locking should do, at least for now. Signed-off-by: Tejun Heo Reported-and-bisected-by: Shaohua Li LKML-Reference: Tested-by: Shaohua Li Signed-off-by: Jens Axboe

block: simplify ioc_release_fn()

2012-02-15T08:45:52Z

Reverse double lock dancing in ioc_release_fn() can be simplified by just using trylock on the queue_lock and back out from ioc lock on trylock failure. Simplify it. Signed-off-by: Tejun Heo Tested-by: Shaohua Li Signed-off-by: Jens Axboe

block: replace icq->changed with icq->flags

2012-02-15T08:45:49Z

icq->changed was used for ICQ_*_CHANGED bits. Rename it to flags and access it under ioc->lock instead of using atomic bitops. ioc_get_changed() is added so that the changed part can be fetched and cleared as before. icq->flags will be used to carry other flags. Signed-off-by: Tejun Heo Tested-by: Shaohua Li Signed-off-by: Jens Axboe

block: fix lockdep warning on io_context release put_io_context()

2012-02-11T11:37:25Z

11a3122f6c "block: strip out locking optimization in put_io_context()" removed ioc_lock depth lockdep annoation along with locking optimization; however, while recursing from put_io_context() is no longer possible, ioc_release_fn() may still end up putting the last reference of another ioc through elevator, which wlil grab ioc->lock triggering spurious (as the ioc is always different one) A-A deadlock warning. As this can only happen one time from ioc_release_fn(), using non-zero subclass from ioc_release_fn() is enough. Use subclass 1. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe