Hello, Fix for this is already submitted @
https://patchwork.ozlabs.org/project/linux-ext4/patch/534f275016296996f54ecf65168bb3392b6f653d.1591699601.git.rite...@linux.ibm.com/ -ritesh On 6/10/20 7:19 PM, YueHaibing wrote:
BUG: using smp_processor_id() in preemptible [00000000] code: kworker/u16:3/2181 caller is ext4_mb_new_blocks+0x388/0xed0 CPU: 2 PID: 2181 Comm: kworker/u16:3 Not tainted 5.7.0+ #182 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Workqueue: writeback wb_workfn (flush-8:0) Call Trace: dump_stack+0xb9/0xfc debug_smp_processor_id+0xc8/0xd0 ext4_mb_new_blocks+0x388/0xed0 ext4_ext_map_blocks+0xa92/0xff0 ext4_map_blocks+0x34e/0x580 ext4_writepages+0xa28/0x11b0 do_writepages+0x46/0xe0 __writeback_single_inode+0x5f/0x6b0 writeback_sb_inodes+0x290/0x620 __writeback_inodes_wb+0x62/0xb0 wb_writeback+0x36c/0x520 wb_workfn+0x319/0x680 process_one_work+0x271/0x640 worker_thread+0x3a/0x3a0 kthread+0x14e/0x170 ret_from_fork+0x27/0x40 Disable preemption before accessing discard_pa_seq. Fixes: 07b5b8e1ac40 ("ext4: mballoc: introduce pcpu seqcnt for freeing PA to improve ENOSPC handling") Signed-off-by: YueHaibing <yuehaib...@huawei.com> --- fs/ext4/mballoc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index a9083113a8c0..30b3bfb1e06a 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -4708,7 +4708,8 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, }ac->ac_op = EXT4_MB_HISTORY_PREALLOC;- seq = *this_cpu_ptr(&discard_pa_seq); + seq = *get_cpu_ptr(&discard_pa_seq); + put_cpu_ptr(&discard_pa_seq); if (!ext4_mb_use_preallocated(ac)) { ac->ac_op = EXT4_MB_HISTORY_ALLOC; ext4_mb_normalize_request(ac, ar);
