kernel/linux.git/drivers/dax/super.c, branch v5.15.208

dax: make sure inodes are flushed before destroy cache

2022-04-08T12:23:31+00:00

[ Upstream commit a7e8de822e0b1979f08767c751f6c8a9c1d4ad86 ] A bug can be triggered by following command $ modprobe nd_pmem && modprobe -r nd_pmem [ 10.060014] BUG dax_cache (Not tainted): Objects remaining in dax_cache on __kmem_cache_shutdown() [ 10.060938] Slab 0x0000000085b729ac objects=9 used=1 fp=0x000000004f5ae469 flags=0x200000000010200(slab|head|node) [ 10.062433] Call Trace: [ 10.062673] dump_stack_lvl+0x34/0x44 [ 10.062865] slab_err+0x90/0xd0 [ 10.063619] __kmem_cache_shutdown+0x13b/0x2f0 [ 10.063848] kmem_cache_destroy+0x4a/0x110 [ 10.064058] __x64_sys_delete_module+0x265/0x300 This is caused by dax_fs_exit() not flushing inodes before destroy cache. To fix this issue, call rcu_barrier() before destroy cache. Signed-off-by: Tong Zhang Reviewed-by: Ira Weiny Reviewed-by: Christoph Hellwig Link: https://lore.kernel.org/r/20220212071111.148575-1-ztong0001@gmail.com Fixes: 7b6be8444e0f ("dax: refactor dax-fs into a generic provider of 'struct dax_device' instances") Signed-off-by: Dan Williams Signed-off-by: Sasha Levin

dax: remove bdev_dax_supported

2021-08-26T23:52:03+00:00

All callers already have a dax_device obtained from fs_dax_get_by_bdev at hand, so just pass that to dax_supported() insted of doing another lookup. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-10-hch@lst.de Signed-off-by: Dan Williams

dax: stub out dax_supported for !CONFIG_FS_DAX

2021-08-26T23:52:03+00:00

dax_supported calls into ->dax_supported which checks for fsdax support. Don't bother building it for !CONFIG_FS_DAX as it will always return false. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-8-hch@lst.de Signed-off-by: Dan Williams

dax: remove __generic_fsdax_supported

2021-08-26T23:52:03+00:00

Just implement generic_fsdax_supported directly out of line instead of adding a wrapper. Given that generic_fsdax_supported is only supplied for CONFIG_FS_DAX builds this also allows to not provide it at all for !CONFIG_FS_DAX builds. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-7-hch@lst.de Signed-off-by: Dan Williams

dax: move the dax_read_lock() locking into dax_supported

2021-08-26T23:52:03+00:00

Move the dax_read_lock/dax_read_unlock pair from the callers into dax_supported to make it a little easier to use. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-6-hch@lst.de Signed-off-by: Dan Williams

dax: mark dax_get_by_host static

2021-08-26T23:52:03+00:00

And move the code around a bit to avoid a forward declaration. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-5-hch@lst.de Signed-off-by: Dan Williams

dax: stop using bdevname

2021-08-26T23:52:02+00:00

Just use the %pg format specifier instead. Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Link: https://lore.kernel.org/r/20210826135510.6293-3-hch@lst.de Signed-off-by: Dan Williams

dax: Ensure errno is returned from dax_direct_access

2021-07-08T05:10:04+00:00

If the caller specifies a negative nr_pages that is an invalid parameter. Return -EINVAL to ensure callers get an errno if they want to check it. Reviewed-by: Dan Williams Signed-off-by: Ira Weiny Link: https://lore.kernel.org/r/20210525172428.3634316-4-ira.weiny@intel.com Signed-off-by: Dan Williams

whack-a-mole: don't open-code iminor/imajor

2021-02-23T15:25:29+00:00

several instances creeped back into the tree... Signed-off-by: Al Viro

device-dax/core: Fix memory leak when rmmod dax.ko

2020-12-16T06:03:52+00:00

When I repeatedly modprobe and rmmod dax.ko, kmemleak report a memory leak as follows: unreferenced object 0xffff9a5588c05088 (size 8): comm "modprobe", pid 261, jiffies 4294693644 (age 42.063s) ... backtrace: [<00000000e007ced0>] kstrdup+0x35/0x70 [<000000002ae73897>] kstrdup_const+0x3d/0x50 [<000000002b00c9c3>] kvasprintf_const+0xbc/0xf0 [<000000008023282f>] kobject_set_name_vargs+0x3b/0xd0 [<00000000d2cbaa4e>] kobject_set_name+0x62/0x90 [<00000000202e7a22>] bus_register+0x7f/0x2b0 [<000000000b77792c>] 0xffffffffc02840f7 [<000000002d5be5ac>] 0xffffffffc02840b4 [<00000000dcafb7cd>] do_one_initcall+0x58/0x240 [<00000000049fe480>] do_init_module+0x56/0x1e2 [<0000000022671491>] load_module+0x2517/0x2840 [<000000001a2201cb>] __do_sys_finit_module+0x9c/0xe0 [<000000003eb304e7>] do_syscall_64+0x33/0x40 [<0000000051c5fd06>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 When rmmod dax is executed, dax_bus_exit() is missing. This patch can fix this bug. Fixes: 9567da0b408a ("device-dax: Introduce bus + driver model") Cc: Reported-by: Hulk Robot Signed-off-by: Wang Hai Link: https://lore.kernel.org/r/20201201135929.66530-1-wanghai38@huawei.com Signed-off-by: Dan Williams