[next-20261111] ext4: Oops in ext4_writepages() during fast commit, replay (generic/483, 64K pagesize, ppc64le)

[Venkat Rao Bagalkote <venkat88@linux.ibm.com>]

Hi,

While running xfstests on linux-next next-20261111 on a Power9 pSeries
LPAR with 64K PAGE_SIZE, generic/483 on ext4 (64K block size, fast_commit
enabled) triggers a kernel oops.


Reproducer
----------
Kernel:   7.4.0-rc2-next-20261111 (ppc64le, PAGE_SIZE=64K, Hash MMU)
Hardware: IBM 8375-42A POWER9 (pSeries LPAR)

xfstests local.config:
   export FSTYP=ext4
   export TEST_DEV=/dev/loop0
   export TEST_DIR=/mnt/test
   export SCRATCH_DEV=/dev/loop1
   export SCRATCH_MNT=/mnt/scratch
   export MKFS_OPTIONS="-b 65536 -O fast_commit"

   ./check -ext4 generic/483



Oops
----
[14844.411091] EXT4-fs (dm-0): mounted filesystem 7ef13de7-... r/w with 
ordered data mode. Quota mode: none.
[14847.236154] EXT4-fs (dm-0): unmounting filesystem 7ef13de7-...
[14847.275380] Kernel attempted to read user page (1174770000) - exploit 
attempt? (uid: 0)
[14847.275406] BUG: Unable to handle kernel data access on read at 
0x1174770000
[14847.275414] Faulting instruction address: 0xc00800000ed42f74
[14847.275423] Oops: Kernel access of bad area, sig: 11 [#1]
[14847.275430] LE PAGE_SIZE=64K MMU=Hash  SMP NR_CPUS=8192 NUMA pSeries
[14847.275575] CPU: 26 UID: 0 PID: 53042 Comm: mount Kdump: loaded Not 
tainted 7.4.0-rc2-next-20261111 #1 PREEMPT(lazy)
[14847.275587] Hardware name: IBM,8375-42A POWER9 (architected) 0x4e0202 
0xf000005 of:IBM,FW950.80 (VL950_131) hv:phyp pSeries
[14847.275597] NIP:  c00800000ed42f74 LR: c00800000ed42f70 CTR: 
c0000000009e2a60
[14847.275647] CFAR: c00000000000db4c DAR: 0000001174770000 DSISR: 
40000000 IRQMASK: 1
[14847.275771] NIP [c00800000ed42f74] ext4_writepages+0x12c/0x3a0 [ext4]
[14847.275908] LR [c00800000ed42f70] ext4_writepages+0x128/0x3a0 [ext4]
[14847.276038] Call Trace:
[14847.276058]  do_writepages+0x190/0x330
[14847.276073]  __writeback_single_inode+0x9c/0xa30
[14847.276087]  writeback_single_inode+0x18c/0x2a0
[14847.276101]  write_inode_now+0xc0/0x100
[14847.276114]  iput_final+0x10c/0x300
[14847.276128]  ext4_fc_replay_add_range+0x2f0/0x5a0 [ext4]
[14847.276262]  ext4_fc_replay+0x564/0xdd0 [ext4]
[14847.276394]  do_one_pass+0x9bc/0x1370 [jbd2]
[14847.276439]  jbd2_journal_recover+0x358/0x400 [jbd2]
[14847.276481]  jbd2_journal_load+0x28/0x110 [jbd2]
[14847.276530]  ext4_load_journal+0x7cc/0xc60 [ext4]
[14847.276664]  ext4_load_and_init_journal+0x6c/0x680 [ext4]
[14847.276801]  __ext4_fill_super+0x1454/0x2360 [ext4]
[14847.276937]  ext4_fill_super+0x188/0x4d0 [ext4]
[14847.277072]  get_tree_bdev_flags+0x250/0x350
[14847.277086]  ext4_get_tree+0x24/0x40 [ext4]
[14847.277221]  vfs_get_tree+0x78/0x210
[14847.277234]  vfs_cmd_create+0x8c/0x1a0
[14847.277246]  __do_sys_fsconfig+0x58c/0x880
[14847.277259]  system_call_exception+0x3c4/0x600
[14847.277273]  system_call_vectored_common+0x15c/0x2ec
[14847.277477] Code: 387f0740 480b8b11 e8410018 ebff0740 7dbe6b78 
386d0030 480b8afd e8410018 ebde0030 7c7ff214 480b8b65 e8410018 
<7d3ff02e> 7f83e378 39290001 7d3ff12e

If you happen to fix this, Please add below tag.

Reported-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>

Regards,

Venkat.