[syzbot] [netfs?] KASAN: slab-use-after-free Write in io_submit_one

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 69b54314c975 Merge tag 'kbuild-fixes-v6.14' of git://git.k..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=147e83f8580000
kernel config: https://syzkaller.appspot.com/x/.config?x=a7ddf49cf33ba213
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=118998e4580000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=10fcbb18580000

Downloadable assets:
disk image (non-bootable): https://storage.googleapis.com/syzbot-assets/7feb34a89c2a/non_bootable_disk-69b54314.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/2d0a58d1d655/vmlinux-69b54314.xz
kernel image: https://storage.googleapis.com/syzbot-assets/b99949b40299/bzImage-69b54314.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+e1dc29...@syzkaller.appspotmail.com

netfs: Couldn't get user pages (rc=-14)
==================================================================
BUG: KASAN: slab-use-after-free in instrument_atomic_read_write include/linux/instrumented.h:96 [inline]
BUG: KASAN: slab-use-after-free in atomic_fetch_sub_release include/linux/atomic/atomic-instrumented.h:400 [inline]
BUG: KASAN: slab-use-after-free in __refcount_sub_and_test include/linux/refcount.h:264 [inline]
BUG: KASAN: slab-use-after-free in __refcount_dec_and_test include/linux/refcount.h:307 [inline]
BUG: KASAN: slab-use-after-free in refcount_dec_and_test include/linux/refcount.h:325 [inline]
BUG: KASAN: slab-use-after-free in iocb_put fs/aio.c:1208 [inline]
BUG: KASAN: slab-use-after-free in io_submit_one+0x4e5/0x1da0 fs/aio.c:2055
Write of size 4 at addr ffff8880317b3b08 by task syz-executor210/6000

CPU: 3 UID: 0 PID: 6000 Comm: syz-executor210 Not tainted 6.14.0-rc1-syzkaller-00276-g69b54314c975 #0
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:94 [inline]
dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120
print_address_description mm/kasan/report.c:378 [inline]
print_report+0xc3/0x620 mm/kasan/report.c:489
kasan_report+0xd9/0x110 mm/kasan/report.c:602
check_region_inline mm/kasan/generic.c:183 [inline]
kasan_check_range+0xef/0x1a0 mm/kasan/generic.c:189
instrument_atomic_read_write include/linux/instrumented.h:96 [inline]
atomic_fetch_sub_release include/linux/atomic/atomic-instrumented.h:400 [inline]
__refcount_sub_and_test include/linux/refcount.h:264 [inline]
__refcount_dec_and_test include/linux/refcount.h:307 [inline]
refcount_dec_and_test include/linux/refcount.h:325 [inline]
iocb_put fs/aio.c:1208 [inline]
io_submit_one+0x4e5/0x1da0 fs/aio.c:2055
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f9104587229
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 81 1d 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007f9104537168 EFLAGS: 00000246 ORIG_RAX: 00000000000000d1
RAX: ffffffffffffffda RBX: 00007f910460b408 RCX: 00007f9104587229
RDX: 00004000000002c0 RSI: 0000000000000001 RDI: 00007f9104516000
RBP: 00007f910460b400 R08: 00007f91045376c0 R09: 0000000000000000
R10: 00007f91045376c0 R11: 0000000000000246 R12: 00007f910460b40c
R13: 000000000000000b R14: 00007fff6ba87360 R15: 00007fff6ba87448
</TASK>

Allocated by task 6000:
kasan_save_stack+0x33/0x60 mm/kasan/common.c:47
kasan_save_track+0x14/0x30 mm/kasan/common.c:68
unpoison_slab_object mm/kasan/common.c:319 [inline]
__kasan_slab_alloc+0x89/0x90 mm/kasan/common.c:345
kasan_slab_alloc include/linux/kasan.h:250 [inline]
slab_post_alloc_hook mm/slub.c:4115 [inline]
slab_alloc_node mm/slub.c:4164 [inline]
kmem_cache_alloc_noprof+0x226/0x3d0 mm/slub.c:4171
aio_get_req fs/aio.c:1058 [inline]
io_submit_one+0x123/0x1da0 fs/aio.c:2048
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

Freed by task 6000:
kasan_save_stack+0x33/0x60 mm/kasan/common.c:47
kasan_save_track+0x14/0x30 mm/kasan/common.c:68
kasan_save_free_info+0x3b/0x60 mm/kasan/generic.c:576
poison_slab_object mm/kasan/common.c:247 [inline]
__kasan_slab_free+0x51/0x70 mm/kasan/common.c:264
kasan_slab_free include/linux/kasan.h:233 [inline]
slab_free_hook mm/slub.c:2353 [inline]
slab_free mm/slub.c:4609 [inline]
kmem_cache_free+0x2e2/0x4d0 mm/slub.c:4711
iocb_destroy fs/aio.c:1110 [inline]
iocb_put fs/aio.c:1210 [inline]
iocb_put fs/aio.c:1206 [inline]
aio_complete_rw+0x3ec/0x7b0 fs/aio.c:1507
netfs_rreq_assess_dio fs/netfs/read_collect.c:375 [inline]
netfs_read_collection+0x30ae/0x3cb0 fs/netfs/read_collect.c:438
netfs_wait_for_pause+0x31c/0x3e0 fs/netfs/read_collect.c:689
netfs_dispatch_unbuffered_reads fs/netfs/direct_read.c:106 [inline]
netfs_unbuffered_read fs/netfs/direct_read.c:144 [inline]
netfs_unbuffered_read_iter_locked+0xb50/0x1610 fs/netfs/direct_read.c:229
netfs_unbuffered_read_iter+0xc5/0x100 fs/netfs/direct_read.c:264
v9fs_file_read_iter+0xbf/0x100 fs/9p/vfs_file.c:361
aio_read+0x313/0x4e0 fs/aio.c:1602
__io_submit_one fs/aio.c:2003 [inline]
io_submit_one+0x1580/0x1da0 fs/aio.c:2052
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

The buggy address belongs to the object at ffff8880317b3a40
which belongs to the cache aio_kiocb of size 216
The buggy address is located 200 bytes inside of
freed 216-byte region [ffff8880317b3a40, ffff8880317b3b18)

The buggy address belongs to the physical page:
page: refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x317b2
head: order:1 mapcount:0 entire_mapcount:0 nr_pages_mapped:0 pincount:0
flags: 0xfff00000000040(head|node=0|zone=1|lastcpupid=0x7ff)
page_type: f5(slab)
raw: 00fff00000000040 ffff888100e8e280 dead000000000122 0000000000000000
raw: 0000000000000000 0000000080190019 00000000f5000000 0000000000000000
head: 00fff00000000040 ffff888100e8e280 dead000000000122 0000000000000000
head: 0000000000000000 0000000080190019 00000000f5000000 0000000000000000
head: 00fff00000000001 ffffea0000c5ec81 ffffffffffffffff 0000000000000000
head: ffff888000000002 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as allocated
page last allocated via order 1, migratetype Unmovable, gfp_mask 0xd20c0(__GFP_IO|__GFP_FS|__GFP_NOWARN|__GFP_NORETRY|__GFP_COMP|__GFP_NOMEMALLOC), pid 5956, tgid 5955 (syz-executor210), ts 65172573555, free_ts 65103511107
set_page_owner include/linux/page_owner.h:32 [inline]
post_alloc_hook+0x181/0x1b0 mm/page_alloc.c:1551
prep_new_page mm/page_alloc.c:1559 [inline]
get_page_from_freelist+0xfce/0x2f80 mm/page_alloc.c:3477
__alloc_frozen_pages_noprof+0x221/0x2470 mm/page_alloc.c:4739
alloc_pages_mpol+0x1fc/0x540 mm/mempolicy.c:2270
alloc_slab_page mm/slub.c:2423 [inline]
allocate_slab mm/slub.c:2587 [inline]
new_slab+0x23d/0x330 mm/slub.c:2640
___slab_alloc+0xc5d/0x1720 mm/slub.c:3826
__slab_alloc.constprop.0+0x56/0xb0 mm/slub.c:3916
__slab_alloc_node mm/slub.c:3991 [inline]
slab_alloc_node mm/slub.c:4152 [inline]
kmem_cache_alloc_noprof+0xfa/0x3d0 mm/slub.c:4171
aio_get_req fs/aio.c:1058 [inline]
io_submit_one+0x123/0x1da0 fs/aio.c:2048
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f
page last free pid 0 tgid 0 stack trace:
reset_page_owner include/linux/page_owner.h:25 [inline]
free_pages_prepare mm/page_alloc.c:1127 [inline]
free_frozen_pages+0x6db/0xfb0 mm/page_alloc.c:2660
rcu_do_batch kernel/rcu/tree.c:2546 [inline]
rcu_core+0x79d/0x14d0 kernel/rcu/tree.c:2802
handle_softirqs+0x213/0x8f0 kernel/softirq.c:561
__do_softirq kernel/softirq.c:595 [inline]
invoke_softirq kernel/softirq.c:435 [inline]
__irq_exit_rcu+0x109/0x170 kernel/softirq.c:662
irq_exit_rcu+0x9/0x30 kernel/softirq.c:678
instr_sysvec_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1049 [inline]
sysvec_apic_timer_interrupt+0xa4/0xc0 arch/x86/kernel/apic/apic.c:1049
asm_sysvec_apic_timer_interrupt+0x1a/0x20 arch/x86/include/asm/idtentry.h:702

Memory state around the buggy address:
ffff8880317b3a00: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
ffff8880317b3a80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff8880317b3b00: fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff8880317b3b80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
ffff8880317b3c00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want syzbot to run the reproducer, reply with:
#syz test: git://repo/address.git branch-or-commit-hash
If you attach or paste a git patch, syzbot will apply it before testing.

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c
index f65affa5a9e4..40c408705f1a 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -371,9 +371,13 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)

if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;
- if (rreq->iocb->ki_complete)
+ if (rreq->iocb->ki_complete) {
+ if (rreq->transferred < rreq->len)
+ refcount_add_not_zero(&rreq->iocb->ki_refcnt);
+
rreq->iocb->ki_complete(
rreq->iocb, rreq->error ? rreq->error : rreq->transferred);
+ }
}
if (rreq->netfs_ops->done)
rreq->netfs_ops->done(rreq);

[syzbot]

Hello,

syzbot tried to test the proposed patch but the build/boot failed:

fs/netfs/read_collect.c:376:66: error: 'struct kiocb' has no member named 'ki_refcnt'
fs/netfs/read_collect.c:376:33: error: too few arguments to function 'refcount_add_not_zero'


Tested on:

commit: 128c8f96 Merge tag 'drm-fixes-2025-02-14' of https://g..
git tree: upstream

kernel config: https://syzkaller.appspot.com/x/.config?x=a7ddf49cf33ba213
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=16c14ce4580000

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c

index f65affa5a9e4..7854d55120b8 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -371,9 +371,11 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)

if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;
- if (rreq->iocb->ki_complete)
+ if (rreq->iocb->ki_complete) {

rreq->iocb->ki_complete(
rreq->iocb, rreq->error ? rreq->error : rreq->transferred);

+ rreq->iocb->ki_complete = NULL;

[syzbot]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
KASAN: slab-use-after-free Write in netfs_read_collection

==================================================================
BUG: KASAN: slab-use-after-free in netfs_rreq_assess_dio fs/netfs/read_collect.c:377 [inline]
BUG: KASAN: slab-use-after-free in netfs_read_collection+0x3baa/0x3d10 fs/netfs/read_collect.c:440
Write of size 8 at addr ffff8880310c6650 by task kworker/u32:2/45

CPU: 0 UID: 0 PID: 45 Comm: kworker/u32:2 Not tainted 6.14.0-rc2-syzkaller-g78a632a2086c-dirty #0

Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014

Workqueue: events_unbound netfs_read_collection_worker

Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:94 [inline]
dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120
print_address_description mm/kasan/report.c:378 [inline]
print_report+0xc3/0x620 mm/kasan/report.c:489
kasan_report+0xd9/0x110 mm/kasan/report.c:602

netfs_rreq_assess_dio fs/netfs/read_collect.c:377 [inline]
netfs_read_collection+0x3baa/0x3d10 fs/netfs/read_collect.c:440
netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:466
process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK>

Allocated by task 6706:

kasan_save_stack+0x33/0x60 mm/kasan/common.c:47
kasan_save_track+0x14/0x30 mm/kasan/common.c:68
unpoison_slab_object mm/kasan/common.c:319 [inline]
__kasan_slab_alloc+0x89/0x90 mm/kasan/common.c:345
kasan_slab_alloc include/linux/kasan.h:250 [inline]
slab_post_alloc_hook mm/slub.c:4115 [inline]
slab_alloc_node mm/slub.c:4164 [inline]
kmem_cache_alloc_noprof+0x226/0x3d0 mm/slub.c:4171
aio_get_req fs/aio.c:1058 [inline]
io_submit_one+0x123/0x1da0 fs/aio.c:2048
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

Freed by task 45:

kasan_save_stack+0x33/0x60 mm/kasan/common.c:47
kasan_save_track+0x14/0x30 mm/kasan/common.c:68
kasan_save_free_info+0x3b/0x60 mm/kasan/generic.c:576
poison_slab_object mm/kasan/common.c:247 [inline]
__kasan_slab_free+0x51/0x70 mm/kasan/common.c:264
kasan_slab_free include/linux/kasan.h:233 [inline]
slab_free_hook mm/slub.c:2353 [inline]
slab_free mm/slub.c:4609 [inline]
kmem_cache_free+0x2e2/0x4d0 mm/slub.c:4711
iocb_destroy fs/aio.c:1110 [inline]
iocb_put fs/aio.c:1210 [inline]
iocb_put fs/aio.c:1206 [inline]
aio_complete_rw+0x3ec/0x7b0 fs/aio.c:1507
netfs_rreq_assess_dio fs/netfs/read_collect.c:375 [inline]

netfs_read_collection+0x30b3/0x3d10 fs/netfs/read_collect.c:440
netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:466
process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244

The buggy address belongs to the object at ffff8880310c6640

which belongs to the cache aio_kiocb of size 216

The buggy address is located 16 bytes inside of
freed 216-byte region [ffff8880310c6640, ffff8880310c6718)

The buggy address belongs to the physical page:

page: refcount:0 mapcount:0 mapping:0000000000000000 index:0xffff8880310c7e00 pfn:0x310c6

head: order:1 mapcount:0 entire_mapcount:0 nr_pages_mapped:0 pincount:0

anon flags: 0xfff00000000040(head|node=0|zone=1|lastcpupid=0x7ff)
page_type: f5(slab)
raw: 00fff00000000040 ffff88801bb3bcc0 0000000000000000 0000000000000001
raw: ffff8880310c7e00 0000000080190015 00000000f5000000 0000000000000000
head: 00fff00000000040 ffff88801bb3bcc0 0000000000000000 0000000000000001
head: ffff8880310c7e00 0000000080190015 00000000f5000000 0000000000000000
head: 00fff00000000001 ffffea0000c43181 ffffffffffffffff 0000000000000000

head: ffff888000000002 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as allocated

page last allocated via order 1, migratetype Unmovable, gfp_mask 0xd20c0(__GFP_IO|__GFP_FS|__GFP_NOWARN|__GFP_NORETRY|__GFP_COMP|__GFP_NOMEMALLOC), pid 6549, tgid 6548 (syz.0.17), ts 88470255044, free_ts 87864364394

set_page_owner include/linux/page_owner.h:32 [inline]
post_alloc_hook+0x181/0x1b0 mm/page_alloc.c:1551
prep_new_page mm/page_alloc.c:1559 [inline]
get_page_from_freelist+0xfce/0x2f80 mm/page_alloc.c:3477
__alloc_frozen_pages_noprof+0x221/0x2470 mm/page_alloc.c:4739
alloc_pages_mpol+0x1fc/0x540 mm/mempolicy.c:2270
alloc_slab_page mm/slub.c:2423 [inline]
allocate_slab mm/slub.c:2587 [inline]
new_slab+0x23d/0x330 mm/slub.c:2640
___slab_alloc+0xc5d/0x1720 mm/slub.c:3826
__slab_alloc.constprop.0+0x56/0xb0 mm/slub.c:3916
__slab_alloc_node mm/slub.c:3991 [inline]
slab_alloc_node mm/slub.c:4152 [inline]
kmem_cache_alloc_noprof+0xfa/0x3d0 mm/slub.c:4171
aio_get_req fs/aio.c:1058 [inline]
io_submit_one+0x123/0x1da0 fs/aio.c:2048
__do_sys_io_submit fs/aio.c:2111 [inline]
__se_sys_io_submit fs/aio.c:2081 [inline]
__x64_sys_io_submit+0x1b2/0x340 fs/aio.c:2081
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

page last free pid 6535 tgid 6535 stack trace:

reset_page_owner include/linux/page_owner.h:25 [inline]
free_pages_prepare mm/page_alloc.c:1127 [inline]
free_frozen_pages+0x6db/0xfb0 mm/page_alloc.c:2660

mm_free_pgd kernel/fork.c:815 [inline]
__mmdrop+0xd5/0x460 kernel/fork.c:931
mmdrop include/linux/sched/mm.h:55 [inline]
__mmput+0x36c/0x410 kernel/fork.c:1367
mmput+0x62/0x70 kernel/fork.c:1378
exec_mmap fs/exec.c:1011 [inline]
begin_new_exec+0x152b/0x3800 fs/exec.c:1267
load_elf_binary+0x886/0x4fc0 fs/binfmt_elf.c:1002
search_binary_handler fs/exec.c:1775 [inline]
exec_binprm fs/exec.c:1807 [inline]
bprm_execve fs/exec.c:1859 [inline]
bprm_execve+0x8dd/0x16d0 fs/exec.c:1835
do_execveat_common.isra.0+0x4a2/0x610 fs/exec.c:1966
do_execve fs/exec.c:2040 [inline]
__do_sys_execve fs/exec.c:2116 [inline]
__se_sys_execve fs/exec.c:2111 [inline]
__x64_sys_execve+0x8c/0xb0 fs/exec.c:2111

do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f

Memory state around the buggy address:

ffff8880310c6500: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff8880310c6580: fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc fc
>ffff8880310c6600: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
^
ffff8880310c6680: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff8880310c6700: fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================


Tested on:

commit: 78a632a2 Merge tag 'pci-v6.14-fixes-3' of git://git.ke..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=159267df980000
kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3

dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=10178bf8580000

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c

index f65affa5a9e4..93eb315fffea 100644

--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -371,9 +371,11 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)

if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;
- if (rreq->iocb->ki_complete)
+ if (rreq->iocb->ki_complete) {
rreq->iocb->ki_complete(
rreq->iocb, rreq->error ? rreq->error : rreq->transferred);

+ rreq->iocb = NULL;

[syzbot]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:

general protection fault in netfs_read_collection

Oops: general protection fault, probably for non-canonical address 0xdffffc0000000002: 0000 [#1] PREEMPT SMP KASAN NOPTI
KASAN: null-ptr-deref in range [0x0000000000000010-0x0000000000000017]
CPU: 1 UID: 0 PID: 103 Comm: kworker/u32:5 Not tainted 6.14.0-rc2-syzkaller-g78a632a2086c-dirty #0

Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: events_unbound netfs_read_collection_worker

RIP: 0010:netfs_rreq_assess_dio fs/netfs/read_collect.c:374 [inline]
RIP: 0010:netfs_read_collection+0x3045/0x3ce0 fs/netfs/read_collect.c:440
Code: 0f 85 16 0b 00 00 4d 03 a6 88 02 00 00 48 b8 00 00 00 00 00 fc ff df 4c 89 65 08 49 8b 6e 58 48 8d 7d 10 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 de 0a 00 00 4c 8b 65 10 4d 85 e4 74 7d e8 e3 55
RSP: 0018:ffffc9000171fb10 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: ffff88804a5f6dd8 RCX: ffffffff82668821
RDX: 0000000000000002 RSI: ffffffff826680c3 RDI: 0000000000000010
RBP: 0000000000000000 R08: 0000000000000004 R09: 0000000000000001
R10: 0000000000000001 R11: 0000000000000001 R12: 00000000ff010000
R13: ffff88804a5f7008 R14: ffff88804a5f6d80 R15: ffff88804a5f6f98
FS: 0000000000000000(0000) GS:ffff88806a700000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001b3185ffff CR3: 000000002a4a6000 CR4: 0000000000352ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>

netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:466
process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK>

Modules linked in:
---[ end trace 0000000000000000 ]---
RIP: 0010:netfs_rreq_assess_dio fs/netfs/read_collect.c:374 [inline]
RIP: 0010:netfs_read_collection+0x3045/0x3ce0 fs/netfs/read_collect.c:440
Code: 0f 85 16 0b 00 00 4d 03 a6 88 02 00 00 48 b8 00 00 00 00 00 fc ff df 4c 89 65 08 49 8b 6e 58 48 8d 7d 10 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 de 0a 00 00 4c 8b 65 10 4d 85 e4 74 7d e8 e3 55
RSP: 0018:ffffc9000171fb10 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: ffff88804a5f6dd8 RCX: ffffffff82668821
RDX: 0000000000000002 RSI: ffffffff826680c3 RDI: 0000000000000010
RBP: 0000000000000000 R08: 0000000000000004 R09: 0000000000000001
R10: 0000000000000001 R11: 0000000000000001 R12: 00000000ff010000
R13: ffff88804a5f7008 R14: ffff88804a5f6d80 R15: ffff88804a5f6f98
FS: 0000000000000000(0000) GS:ffff88806a700000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001b3185ffff CR3: 000000002a4a6000 CR4: 0000000000352ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
----------------
Code disassembly (best guess):
0: 0f 85 16 0b 00 00 jne 0xb1c
6: 4d 03 a6 88 02 00 00 add 0x288(%r14),%r12
d: 48 b8 00 00 00 00 00 movabs $0xdffffc0000000000,%rax
14: fc ff df
17: 4c 89 65 08 mov %r12,0x8(%rbp)
1b: 49 8b 6e 58 mov 0x58(%r14),%rbp
1f: 48 8d 7d 10 lea 0x10(%rbp),%rdi
23: 48 89 fa mov %rdi,%rdx
26: 48 c1 ea 03 shr $0x3,%rdx
* 2a: 80 3c 02 00 cmpb $0x0,(%rdx,%rax,1) <-- trapping instruction
2e: 0f 85 de 0a 00 00 jne 0xb12
34: 4c 8b 65 10 mov 0x10(%rbp),%r12
38: 4d 85 e4 test %r12,%r12
3b: 74 7d je 0xba
3d: e8 .byte 0xe8
3e: e3 55 jrcxz 0x95

Tested on:

commit: 78a632a2 Merge tag 'pci-v6.14-fixes-3' of git://git.ke..
git tree: upstream

console output: https://syzkaller.appspot.com/x/log.txt?x=109b19a4580000

kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=17df8bf8580000

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c

index f65affa5a9e4..73fea10d9a78 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -369,12 +369,17 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)
}
}

+ spin_lock(&rreq->lock);

if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;
- if (rreq->iocb->ki_complete)
+ if (rreq->iocb->ki_complete) {
rreq->iocb->ki_complete(
rreq->iocb, rreq->error ? rreq->error : rreq->transferred);
+ rreq->iocb = NULL;
+ }
}

+ spin_unlock(&rreq->lock);

+
if (rreq->netfs_ops->done)
rreq->netfs_ops->done(rreq);

if (rreq->origin == NETFS_DIO_READ)

[syzbot]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:

WARNING: refcount bug in netfs_put_subrequest

------------[ cut here ]------------
refcount_t: underflow; use-after-free.
WARNING: CPU: 1 PID: 81 at lib/refcount.c:28 refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28
Modules linked in:
CPU: 1 UID: 0 PID: 81 Comm: kworker/u32:4 Not tainted 6.14.0-rc2-syzkaller-g78a632a2086c-dirty #0

Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: events_unbound netfs_read_collection_worker

RIP: 0010:refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28
Code: ff 89 de e8 c8 5a f5 fc 84 db 0f 85 66 ff ff ff e8 1b 60 f5 fc c6 05 d7 5c 86 0b 01 90 48 c7 c7 c0 00 d3 8b e8 77 99 b5 fc 90 <0f> 0b 90 90 e9 43 ff ff ff e8 f8 5f f5 fc 0f b6 1d b2 5c 86 0b 31
RSP: 0018:ffffc9000162fab8 EFLAGS: 00010282
RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffffffff817a1229
RDX: ffff888020af2440 RSI: ffffffff817a1236 RDI: 0000000000000001
RBP: ffff8880305354a0 R08: 0000000000000001 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000002 R12: 0000000000000005
R13: 00000000000006ed R14: 0000000000000001 R15: ffff8880305354a0

FS: 0000000000000000(0000) GS:ffff88806a700000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033

CR2: 00007fb7675a2f98 CR3: 000000000df80000 CR4: 0000000000352ef0

DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>

__refcount_sub_and_test include/linux/refcount.h:275 [inline]
__refcount_dec_and_test include/linux/refcount.h:307 [inline]
netfs_put_subrequest+0x2c1/0x4d0 fs/netfs/objects.c:230
netfs_collect_read_results fs/netfs/read_collect.c:300 [inline]
netfs_read_collection+0x25af/0x3d00 fs/netfs/read_collect.c:422
netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:469

process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK>

Tested on:

commit: 78a632a2 Merge tag 'pci-v6.14-fixes-3' of git://git.ke..
git tree: upstream

console output: https://syzkaller.appspot.com/x/log.txt?x=123719a4580000

kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=16a667df980000

[Edward Adam Davis]

@@ -301,6 +301,8 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
notes & ABANDON_SREQ ?
netfs_sreq_trace_put_abandon :
netfs_sreq_trace_put_done);
+ if (front == remove)
+ break;
}

trace_netfs_collect_stream(rreq, stream);

[syzbot]

Hello,

syzbot tried to test the proposed patch but the build/boot failed:

failed to apply patch:
checking file fs/netfs/read_collect.c
misordered hunks! output would be garbled
Hunk #2 FAILED at 306.
1 out of 2 hunks FAILED

Tested on:

commit: 78a632a2 Merge tag 'pci-v6.14-fixes-3' of git://git.ke..
git tree: upstream

kernel config: https://syzkaller.appspot.com/x/.config?x=a7ddf49cf33ba213

dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler:

patch: https://syzkaller.appspot.com/x/patch.diff?x=132099a4580000

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c

index f65affa5a9e4..72b08c1709fa 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c

@@ -301,6 +301,8 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
notes & ABANDON_SREQ ?
netfs_sreq_trace_put_abandon :
netfs_sreq_trace_put_done);
+ if (front == remove)
+ break;
}

trace_netfs_collect_stream(rreq, stream);

@@ -369,12 +371,17 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)

[syzbot]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
WARNING: refcount bug in netfs_put_subrequest

------------[ cut here ]------------
refcount_t: underflow; use-after-free.

WARNING: CPU: 1 PID: 85 at lib/refcount.c:28 refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28
Modules linked in:
CPU: 1 UID: 0 PID: 85 Comm: kworker/u32:4 Not tainted 6.14.0-rc2-syzkaller-g7ff71e6d9239-dirty #0

Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: events_unbound netfs_read_collection_worker
RIP: 0010:refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28

Code: ff 89 de e8 08 5b f5 fc 84 db 0f 85 66 ff ff ff e8 5b 60 f5 fc c6 05 17 5d 86 0b 01 90 48 c7 c7 c0 00 d3 8b e8 b7 99 b5 fc 90 <0f> 0b 90 90 e9 43 ff ff ff e8 38 60 f5 fc 0f b6 1d f2 5c 86 0b 31
RSP: 0018:ffffc90000fdfac0 EFLAGS: 00010286

RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffffffff817a1229

RDX: ffff888021798000 RSI: ffffffff817a1236 RDI: 0000000000000001
RBP: ffff888030d38e60 R08: 0000000000000001 R09: 0000000000000000

R10: 0000000000000000 R11: 0000000000000002 R12: 0000000000000005

R13: 0000000000000135 R14: 0000000000000001 R15: ffff888030d38e60

FS: 0000000000000000(0000) GS:ffff88806a700000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033

CR2: 000000c0022d90a0 CR3: 0000000023ac0000 CR4: 0000000000352ef0

DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
__refcount_sub_and_test include/linux/refcount.h:275 [inline]
__refcount_dec_and_test include/linux/refcount.h:307 [inline]
netfs_put_subrequest+0x2c1/0x4d0 fs/netfs/objects.c:230
netfs_collect_read_results fs/netfs/read_collect.c:300 [inline]

netfs_read_collection+0x1f09/0x3cc0 fs/netfs/read_collect.c:424
netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:471

process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK>


Tested on:

commit: 7ff71e6d Merge tag 'alpha-fixes-v6.14-rc2' of git://gi..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=142967df980000
kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130

compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=143167df980000

[Edward Adam Davis]

#syz test

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c

index f65affa5a9e4..6f3c0404f4b8 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -207,6 +207,7 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
* in progress. The issuer thread may be adding stuff to the tail
* whilst we're doing this.
*/
+ spin_lock(&rreq->lock);
front = READ_ONCE(stream->front);
while (front) {
size_t transferred;
@@ -288,7 +289,6 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)

/* Remove if completely consumed. */
stream->source = front->source;
- spin_lock(&rreq->lock);

remove = front;
trace_netfs_sreq(front, netfs_sreq_trace_discard);
@@ -296,12 +296,12 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
front = list_first_entry_or_null(&stream->subrequests,
struct netfs_io_subrequest, rreq_link);
stream->front = front;
- spin_unlock(&rreq->lock);
netfs_put_subrequest(remove, false,

notes & ABANDON_SREQ ?
netfs_sreq_trace_put_abandon :
netfs_sreq_trace_put_done);
}

+ spin_unlock(&rreq->lock);

trace_netfs_collect_stream(rreq, stream);
trace_netfs_collect_state(rreq, rreq->collected_to, notes);
@@ -369,12 +369,17 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)

[syzbot]

Hello,

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-by: syzbot+e1dc29...@syzkaller.appspotmail.com
Tested-by: syzbot+e1dc29...@syzkaller.appspotmail.com

Tested on:

commit: 7ff71e6d Merge tag 'alpha-fixes-v6.14-rc2' of git://gi..
git tree: upstream

console output: https://syzkaller.appspot.com/x/log.txt?x=13abd9b0580000

kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=11924bf8580000

Note: testing is done by a robot and is best-effort only.

[Hillf Danton]

On Thu, 13 Feb 2025 21:55:19 -0800

> syzbot found the following issue on:
>
> HEAD commit: 69b54314c975 Merge tag 'kbuild-fixes-v6.14' of git://git.k..
> git tree: upstream

> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=10fcbb18580000

#syz test upstream master

--- x/fs/netfs/read_collect.c
+++ y/fs/netfs/read_collect.c
@@ -289,6 +289,10 @@ reassess:

/* Remove if completely consumed. */
stream->source = front->source;

spin_lock(&rreq->lock);
+ if (front != stream->front) {
+ spin_unlock(&rreq->lock);
+ goto reassess;
+ }

remove = front;
trace_netfs_sreq(front, netfs_sreq_trace_discard);

@@ -345,6 +349,7 @@ static void netfs_rreq_assess_dio(struct
struct netfs_io_subrequest *subreq;
struct netfs_io_stream *stream = &rreq->io_streams[0];
unsigned int i;
+ struct kiocb *iocb = NULL;

/* Collect unbuffered reads and direct reads, adding up the transfer
* sizes until we find the first short or failed subrequest.
@@ -369,12 +374,16 @@ static void netfs_rreq_assess_dio(struct

}
}

+ spin_lock(&rreq->lock);
if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;

if (rreq->iocb->ki_complete)
- rreq->iocb->ki_complete(
- rreq->iocb, rreq->error ? rreq->error : rreq->transferred);
+ iocb = rreq->iocb;

+ rreq->iocb = NULL;
}

+ spin_unlock(&rreq->lock);
+ if (iocb)
+ iocb->ki_complete(iocb, rreq->error ? rreq->error : rreq->transferred);

if (rreq->netfs_ops->done)
rreq->netfs_ops->done(rreq);
if (rreq->origin == NETFS_DIO_READ)

--

[syzbot]

Hello,

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-by: syzbot+e1dc29...@syzkaller.appspotmail.com
Tested-by: syzbot+e1dc29...@syzkaller.appspotmail.com

Tested on:

commit: 7ff71e6d Merge tag 'alpha-fixes-v6.14-rc2' of git://gi..
git tree: upstream

console output: https://syzkaller.appspot.com/x/log.txt?x=107d6098580000

kernel config: https://syzkaller.appspot.com/x/.config?x=c09dc55ba7f798e3
dashboard link: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
compiler: gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=126e99a4580000

[Edward Adam Davis]

syzbot reported a slab-use-after-free Write in io_submit_one. [1]

netfs_read_collection() can be reached by two paths: netfs read collection
worker and aio read.
In aio_get_req(), ki_refcnt is set to 2.

CPU1 CPU2
============== ==================
io_submit_one
->io_get_req
->__io_submit_one
-->aio_read
-->v9fs_file_read_iter
-->netfs_unbuffered_read_iter
-->netfs_unbuffered_read process_one_work
-->netfs_dispatch_unbuffered_reads netfs_read_collection_worker
-->netfs_wait_for_pause netfs_read_collection
-->netfs_read_collection netfs_rreq_assess_dio
-->netfs_rreq_assess_dio aio_complete_rw
-->aio_complete_rw iocb_put
-->iocb_put
->iocb_put

The netfs read collection worker calls ki_complete() once, plus aio_read calls
ki_complete() once, causing iocb to be released before iocb_put is executed
in io_submit_one(), which triggers [1].

The aio read and netfs read collection worker are synchronized by locking
"rreq->lock" in netfs_rreq_assess_dio(), and iocb is set to NULL to prevent
ki_complete from being executed repeatedly by aio_read and worker.

[2] This is also caused by the race condition between aio_read and netfs
read collection worker mentioned above. We deal with it here by expanding
the lock range of "rreq->lock" for both to ensure that stream->front will
not be modified by the other party before the remove is executed.

[1]

[2]
WARNING: CPU: 1 PID: 81 at lib/refcount.c:28 refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28
Modules linked in:
CPU: 1 UID: 0 PID: 81 Comm: kworker/u32:4 Not tainted 6.14.0-rc2-syzkaller-g78a632a2086c-dirty #0

Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014

Workqueue: events_unbound netfs_read_collection_worker
RIP: 0010:refcount_warn_saturate+0x14a/0x210 lib/refcount.c:28

Code: ff 89 de e8 c8 5a f5 fc 84 db 0f 85 66 ff ff ff e8 1b 60 f5 fc c6 05 d7 5c 86 0b 01 90 48 c7 c7 c0 00 d3 8b e8 77 99 b5 fc 90 <0f> 0b 90 90 e9 43 ff ff ff e8 f8 5f f5 fc 0f b6 1d b2 5c 86 0b 31
RSP: 0018:ffffc9000162fab8 EFLAGS: 00010282

RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffffffff817a1229

RDX: ffff888020af2440 RSI: ffffffff817a1236 RDI: 0000000000000001
RBP: ffff8880305354a0 R08: 0000000000000001 R09: 0000000000000000

R10: 0000000000000000 R11: 0000000000000002 R12: 0000000000000005

R13: 00000000000006ed R14: 0000000000000001 R15: ffff8880305354a0

FS: 0000000000000000(0000) GS:ffff88806a700000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033

CR2: 00007fb7675a2f98 CR3: 000000000df80000 CR4: 0000000000352ef0

DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>

__refcount_sub_and_test include/linux/refcount.h:275 [inline]
__refcount_dec_and_test include/linux/refcount.h:307 [inline]

netfs_put_subrequest+0x2c1/0x4d0 fs/netfs/objects.c:230
netfs_collect_read_results fs/netfs/read_collect.c:300 [inline]

netfs_read_collection+0x25af/0x3d00 fs/netfs/read_collect.c:422
netfs_read_collection_worker+0x285/0x350 fs/netfs/read_collect.c:469

process_one_work+0x9c5/0x1ba0 kernel/workqueue.c:3236
process_scheduled_works kernel/workqueue.c:3317 [inline]
worker_thread+0x6c8/0xf00 kernel/workqueue.c:3398
kthread+0x3af/0x750 kernel/kthread.c:464
ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:148
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK>

Fixes: ee4cdf7ba857 ("netfs: Speed up buffered reading")
Reported-by: syzbot+e1dc29...@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=e1dc29a4daf3f8051130
Tested-by: syzbot+e1dc29...@syzkaller.appspotmail.com
Signed-off-by: Edward Adam Davis <ead...@qq.com>
---
fs/netfs/read_collect.c | 11 ++++++++---
1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/fs/netfs/read_collect.c b/fs/netfs/read_collect.c
index f65affa5a9e4..6f3c0404f4b8 100644
--- a/fs/netfs/read_collect.c
+++ b/fs/netfs/read_collect.c
@@ -207,6 +207,7 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
* in progress. The issuer thread may be adding stuff to the tail
* whilst we're doing this.
*/
+ spin_lock(&rreq->lock);
front = READ_ONCE(stream->front);
while (front) {
size_t transferred;
@@ -288,7 +289,6 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)

/* Remove if completely consumed. */
stream->source = front->source;

- spin_lock(&rreq->lock);

remove = front;
trace_netfs_sreq(front, netfs_sreq_trace_discard);

@@ -296,12 +296,12 @@ static void netfs_collect_read_results(struct netfs_io_request *rreq)
front = list_first_entry_or_null(&stream->subrequests,
struct netfs_io_subrequest, rreq_link);
stream->front = front;
- spin_unlock(&rreq->lock);
netfs_put_subrequest(remove, false,
notes & ABANDON_SREQ ?
netfs_sreq_trace_put_abandon :
netfs_sreq_trace_put_done);
}
+ spin_unlock(&rreq->lock);

trace_netfs_collect_stream(rreq, stream);
trace_netfs_collect_state(rreq, rreq->collected_to, notes);
@@ -369,12 +369,17 @@ static void netfs_rreq_assess_dio(struct netfs_io_request *rreq)
}
}

+ spin_lock(&rreq->lock);
if (rreq->iocb) {
rreq->iocb->ki_pos += rreq->transferred;

- if (rreq->iocb->ki_complete)
+ if (rreq->iocb->ki_complete) {
rreq->iocb->ki_complete(

rreq->iocb, rreq->error ? rreq->error : rreq->transferred);

+ rreq->iocb = NULL;
+ }
}
+ spin_unlock(&rreq->lock);
+

if (rreq->netfs_ops->done)
rreq->netfs_ops->done(rreq);
if (rreq->origin == NETFS_DIO_READ)
--

2.43.0

[syzbot]

Auto-closing this bug as obsolete.
No recent activity, existing reproducers are no longer triggering the issue.