On Tue, Apr 25, 2023 at 11:27 AM Ming Lei <ming.lei@xxxxxxxxxx> wrote: > > On Tue, Apr 25, 2023 at 10:37:05AM +0800, Changhui Zhong wrote: > > Hello, > > > > Below issue was triggered in my test,it caused system panic ,please > > help check it. > > branch: for-6.4/block > > https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git > > > > mdadm -CR /dev/md0 -l 1 -n 2 /dev/sda /dev/sdb -e 1.0 > > sgdisk -n 0:0:+100MiB /dev/md0 > > cat /proc/partitions > > mdadm -S /dev/md0 > > mdadm -A /dev/md0 /dev/sda /dev/sdb > > cat /proc/partitions > > > > > > [ 34.219123] BUG: kernel NULL pointer dereference, address: 00000000000000fc > > [ 34.219507] #PF: supervisor read access in kernel mode > > [ 34.219784] #PF: error_code(0x0000) - not-present page > > [ 34.220039] PGD 0 P4D 0 > > [ 34.220176] Oops: 0000 [#1] PREEMPT SMP PTI > > [ 34.220374] CPU: 8 PID: 1956 Comm: systemd-udevd Kdump: loaded Not > > tainted 6.3.0-rc2+ #1 > > [ 34.220787] Hardware name: HP ProLiant DL360 Gen9/ProLiant DL360 > > Gen9, BIOS P89 05/21/2018 > > [ 34.221188] RIP: 0010:blk_mq_sched_bio_merge+0x6d/0xf0 > > Hi Changhui, > > Please try the following fix: > > diff --git a/block/bdev.c b/block/bdev.c > index 850852fe4b78..fa2838ca2e6d 100644 > --- a/block/bdev.c > +++ b/block/bdev.c > @@ -419,7 +419,11 @@ struct block_device *bdev_alloc(struct gendisk *disk, u8 partno) > bdev->bd_inode = inode; > bdev->bd_queue = disk->queue; > bdev->bd_stats = alloc_percpu(struct disk_stats); > - bdev->bd_has_submit_bio = false; > + > + if (partno) > + bdev->bd_has_submit_bio = disk->part0->bd_has_submit_bio; > + else > + bdev->bd_has_submit_bio = false; > if (!bdev->bd_stats) { > iput(inode); > return NULL; > > Fixes: 9f4107b07b17 ("block: store bdev->bd_disk->fops->submit_bio state in bdev") > Hi,Ming I retest the latest updated branch for-6.4/block (https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git/log/?h=for-6.4/block), which contain your fix patch "block: sync part's ->bd_has_submit_bio with disk's". the kernel panic no longer happen, but the test will failed, the system will reread partition table on exclusively open device, :: [ 01:50:05 ] :: [ BEGIN ] :: Running 'mdadm -S /dev/md0' mdadm: stopped /dev/md0 :: [ 01:50:06 ] :: [ PASS ] :: Command 'mdadm -S /dev/md0' (Expected 0, got 0) :: [ 01:50:06 ] :: [ BEGIN ] :: Running 'mdadm -A /dev/md0 /dev/"$dev0" /dev/"$dev1"' mdadm: /dev/md0 has been started with 2 drives. :: [ 01:50:06 ] :: [ PASS ] :: Command 'mdadm -A /dev/md0 /dev/"$dev0" /dev/"$dev1"' (Expected 0, got 0) :: [ 01:50:09 ] :: [ BEGIN ] :: Running 'lsblk' NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 1 447.1G 0 disk ├─sda1 8:1 1 1G 0 part /boot └─sda2 8:2 1 446.1G 0 part ├─rhel_storageqe--104-root 253:0 0 70G 0 lvm / ├─rhel_storageqe--104-swap 253:1 0 7.7G 0 lvm [SWAP] └─rhel_storageqe--104-home 253:2 0 368.4G 0 lvm /home sdb 8:16 1 447.1G 0 disk ├─sdb1 8:17 1 100M 0 part └─md0 9:0 0 447.1G 0 raid1 └─md0p1 259:0 0 100M 0 part sdc 8:32 1 447.1G 0 disk ├─sdc1 8:33 1 100M 0 part └─md0 9:0 0 447.1G 0 raid1 └─md0p1 259:0 0 100M 0 part sdd 8:48 1 447.1G 0 disk :: [ 01:50:09 ] :: [ PASS ] :: Command 'lsblk' (Expected 0, got 0) :: [ 01:50:09 ] :: [ BEGIN ] :: Running 'cat /proc/partitions' major minor #blocks name 8 0 468851544 sda 8 1 1048576 sda1 8 2 467801088 sda2 8 48 468851544 sdd 8 32 468851544 sdc 8 33 102400 sdc1 8 16 468851544 sdb 8 17 102400 sdb1 253 0 73400320 dm-0 253 1 8060928 dm-1 253 2 386338816 dm-2 9 0 468851392 md0 259 0 102400 md0p1 Thanks，
