Hi Steven,

On 2026/7/14 3:36, Steven Rostedt wrote:

On Mon, 13 Jul 2026 13:46:40 +0000
Tengda Wu [off-list ref] wrote:

quoted

The trace_parser structure is allocated and initialized when a trace
file is opened, and is subsequently used in the write handler to parse
user input. If userspace opens a trace file descriptor and shares it
across multiple threads, concurrent write calls will race on the
parser's internal state, specifically the idx, cont, and buffer fields,
leading to corrupted input or undefined behavior.

Fix this by embedding a mutex directly in struct trace_parser. The mutex
is initialized in trace_parser_get_init() and destroyed in
trace_parser_put(). All write-side users that access parser state
(trace_get_user() followed by checking trace_parser_loaded() /
trace_parser_cont() against the buffer) now hold the mutex across the
full critical section, avoiding any TOCTOU gap between the parse and the
subsequent consumption of parser->buffer.

Affected write paths:
  - ftrace_graph_write / ftrace_graph_release
  - ftrace_regex_write / ftrace_regex_release

Fixes: e704eff3ff51 ("ftrace: Have set_graph_function handle multiple functions in one write")
Fixes: 689fd8b65d66 ("tracing: trace parser support for function and graph")
Cc: stable@vger.kernel.org
Signed-off-by: Tengda Wu <redacted>
---
 kernel/trace/ftrace.c | 7 +++++++
 kernel/trace/trace.c  | 2 ++
 kernel/trace/trace.h  | 1 +
 3 files changed, 10 insertions(+)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index f93e34dd2328..ef47e5659283 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c

@@ -5842,6 +5842,8 @@ ftrace_regex_write(struct file *file, const char __user *ubuf,
 	/* iter->hash is a local copy, so we don't need regex_lock */
 
 	parser = &iter->parser;
+
+	guard(mutex)(&parser->lock);
 	read = trace_get_user(parser, ubuf, cnt, ppos);

Why are the other users of trace_get_user() not a problem? If
anything, trace_get_user() should have a lockdep assert to make sure
the lock is held.

I think we need to add a lockdep assertion in all the callers that use
the parser and we need to make sure it's taken by every user.

-- Steve

Other places that use trace_get_user():

ftrace_event_write
    trace_parser_get_init
    trace_get_user

trace_pid_write
    trace_parser_get_init
    trace_get_user

In both of these cases, the parser is allocated via trace_parser_get_init()
before use and freed immediately after, with no multi-threaded sharing.

However, ftrace_graph_write() and ftrace_regex_write() are different.
The parser used by these two functions is allocated at open() time and
retrieved from struct file at write() time. Userspace can have multiple
threads invoking write() concurrently, e.g.:

    r0 = openat(AT_FDCWD, path, O_WRONLY | O_CREAT | O_TRUNC |
                O_NOCTTY | O_NONBLOCK, 0);
    r1 = dup(r0);
    write(r1, data1, len1);        /* thread 1 */
    pwrite64(r1, data2, len2, offset); /* thread 2 */

Without proper synchronization in trace_get_user(), the parser state
becomes undefined.

We have locally reproduced a slab-out-of-bounds issue with syzkaller [1],
which appears to be caused by this race.

Regarding the fix, adding a lockdep assertion to trace_get_user() does
not seem necessary. As mentioned above, ftrace_event_write() and
trace_pid_write() do not involve concurrency, so they do not require
locking, and thus do not need a lockdep assertion.

[1] slab-out-of-bounds call trace:

[   29.629421][ T1269] ==================================================================
[   29.630348][ T1269] BUG: KASAN: slab-out-of-bounds in strstr+0x44/0x190
[   29.631098][ T1269] Read of size 1 at addr ffff00000ffa95f3 by task syz.3.4/1269
[   29.632120][ T1269] 
[   29.632390][ T1269] CPU: 2 PID: 1269 Comm: syz.3.4 Not tainted 5.10.0-27349-g9a396a20c0b1-dirty #1
[   29.634205][ T1269] Hardware name: linux,dummy-virt (DT)
[   29.635862][ T1269] Call trace:
[   29.636949][ T1269]  dump_backtrace+0x0/0x510
[   29.637954][ T1269]  show_stack+0x34/0x44
[   29.638438][ T1269]  dump_stack+0x1d0/0x248
[   29.639006][ T1269]  print_address_description.constprop.0+0x2c/0x1fc
[   29.639776][ T1269]  __kasan_report+0xfc/0x160
[   29.640325][ T1269]  kasan_report+0x44/0x60
[   29.640904][ T1269]  __asan_load1+0x70/0xa0
[   29.641544][ T1269]  strstr+0x44/0x190
[   29.642074][ T1269]  ftrace_match+0x130/0x200
[   29.642662][ T1269]  ftrace_match_record.isra.0+0x16c/0x210
[   29.643317][ T1269]  ftrace_graph_set_hash+0x20c/0x440
[   29.643959][ T1269]  ftrace_graph_write+0x170/0x200
[   29.644534][ T1269]  vfs_write+0x17c/0x66c
[   29.645018][ T1269]  ksys_write+0x108/0x220
[   29.645510][ T1269]  __arm64_sys_write+0x54/0x70
[   29.646058][ T1269]  el0_svc_common.constprop.0+0xd8/0x37c
[   29.646683][ T1269]  do_el0_svc+0x50/0x130
[   29.647191][ T1269]  el0_svc+0x24/0x34
[   29.647648][ T1269]  el0_sync_handler+0x180/0x18c
[   29.648232][ T1269]  el0_sync+0x160/0x180
[   29.648712][ T1269] 
[   29.648994][ T1269] Allocated by task 1269:
[   29.649505][ T1269]  kasan_save_stack+0x28/0x60
[   29.650069][ T1269]  __kasan_kmalloc.constprop.0+0xd8/0x140
[   29.650794][ T1269]  kasan_kmalloc+0x10/0x20
[   29.651365][ T1269]  __kmalloc+0xd4/0x7c0
[   29.651902][ T1269]  trace_parser_get_init+0x3c/0xa0
[   29.652547][ T1269]  __ftrace_graph_open.part.0+0xc4/0x250
[   29.653178][ T1269]  ftrace_graph_open+0xd8/0x154
[   29.653715][ T1269]  do_dentry_open+0x400/0x950
[   29.654231][ T1269]  vfs_open+0x60/0x70
[   29.654681][ T1269]  do_open+0x4f0/0x6f0
[   29.655223][ T1269]  path_openat+0x268/0x3a4
[   29.655707][ T1269]  do_filp_open+0x10c/0x23c
[   29.656234][ T1269]  do_sys_openat2+0x4e4/0x700
[   29.656749][ T1269]  __arm64_sys_openat+0x11c/0x1a0
[   29.657335][ T1269]  el0_svc_common.constprop.0+0xd8/0x37c
[   29.657342][ T1269]  do_el0_svc+0x50/0x130
[   29.657350][ T1269]  el0_svc+0x24/0x34
[   29.657359][ T1269]  el0_sync_handler+0x180/0x18c
[   29.657366][ T1269]  el0_sync+0x160/0x180
[   29.657368][ T1269] 
[   29.657375][ T1269] The buggy address belongs to the object at ffff00000ffa9500
[   29.657375][ T1269]  which belongs to the cache kmalloc-256 of size 256
[   29.657383][ T1269] The buggy address is located 243 bytes inside of
[   29.657383][ T1269]  256-byte region [ffff00000ffa9500, ffff00000ffa9600)
[   29.657386][ T1269] The buggy address belongs to the page:
[   29.657398][ T1269] page:fffffe00001fea00 refcount:1 mapcount:0 mapping:0000000000000000 index:0xffff00000ffa8500 pfn:0x4ffa8
[   29.657405][ T1269] head:fffffe00001fea00 order:2 compound_mapcount:0 compound_pincount:0
[   29.657417][ T1269] flags: 0x7ffff800010200(slab|head|node=0|zone=0|lastcpupid=0xfffff)
[   29.657430][ T1269] raw: 007ffff800010200 ffff0000c0000820 ffff0000c0000820 ffff0000c000cd00
[   29.657441][ T1269] raw: ffff00000ffa8500 0000000000100008 00000001ffffffff 0000000000000000
[   29.657445][ T1269] page dumped because: kasan: bad access detected
[   29.657447][ T1269] 
[   29.657451][ T1269] Memory state around the buggy address:
[   29.657458][ T1269]  ffff00000ffa9480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   29.657465][ T1269]  ffff00000ffa9500: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[   29.657473][ T1269] >ffff00000ffa9580: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 03 fc
[   29.657477][ T1269]                                                              ^
[   29.657484][ T1269]  ffff00000ffa9600: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   29.657491][ T1269]  ffff00000ffa9680: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   29.657494][ T1269] ==================================================================
[   29.657498][ T1269] Disabling lock debugging due to kernel taint
[   29.657656][ T1269] Kernel panic - not syncing: KASAN: panic_on_warn set ...
[   29.657668][ T1269] CPU: 2 PID: 1269 Comm: syz.3.4 Tainted: G    B             5.10.0-27349-g9a396a20c0b1-dirty #1
[   29.657672][ T1269] Hardware name: linux,dummy-virt (DT)
[   29.657676][ T1269] Call trace:
[   29.657685][ T1269]  dump_backtrace+0x0/0x510
[   29.657692][ T1269]  show_stack+0x34/0x44
[   29.657700][ T1269]  dump_stack+0x1d0/0x248
[   29.657711][ T1269]  panic+0x390/0x728
[   29.657719][ T1269]  check_panic_on_warn+0x108/0x12c
[   29.657727][ T1269]  end_report+0x5c/0xe0
[   29.657748][ T1269]  __kasan_report+0x118/0x160
[   29.657756][ T1269]  kasan_report+0x44/0x60
[   29.657764][ T1269]  __asan_load1+0x70/0xa0
[   29.657771][ T1269]  strstr+0x44/0x190
[   29.657780][ T1269]  ftrace_match+0x130/0x200
[   29.657790][ T1269]  ftrace_match_record.isra.0+0x16c/0x210
[   29.657799][ T1269]  ftrace_graph_set_hash+0x20c/0x440
[   29.657808][ T1269]  ftrace_graph_write+0x170/0x200
[   29.657815][ T1269]  vfs_write+0x17c/0x66c
[   29.657821][ T1269]  ksys_write+0x108/0x220
[   29.657828][ T1269]  __arm64_sys_write+0x54/0x70
[   29.657836][ T1269]  el0_svc_common.constprop.0+0xd8/0x37c
[   29.657842][ T1269]  do_el0_svc+0x50/0x130
[   29.657850][ T1269]  el0_svc+0x24/0x34
[   29.657858][ T1269]  el0_sync_handler+0x180/0x18c
[   29.657865][ T1269]  el0_sync+0x160/0x180
[   29.657876][ T1269] SMP: stopping secondary CPUs
[   29.657896][ T1269] Kernel Offset: disabled
[   29.690695][ T1269] HKRR parameters: disabled
[   29.691167][ T1269] CPU features: 0x0004,88040002,6a208038
[   29.691749][ T1269] Memory Limit: none
[   29.692186][ T1269] ---[ end Kernel panic - not syncing: KASAN: panic_on_warn set ... ]---

Thanks,
Tengda