Re: [PATCH 2/2] ftrace: Complete subops teardown after hash update failure

From: Weiming Shi

Date: Thu Nov 19 2026 - 11:20:17 EST


Steven Rostedt <rostedt@xxxxxxxxxxx> 于2026年11月18日周三 10:49写道:
>
> On Tue, 3 Nov 2026 09:58:31 -0800
> Weiming Shi <bestswngs@xxxxxxxxx> wrote:
>
> > When removing a graph operation while other operations remain,
> > ftrace_shutdown_subops() rebuilds the manager hashes. Memory allocations
> > in rebuild_hashes() or ftrace_update_ops() can fail with -ENOMEM.
> >
> > On a rebuild failure, the subop has been removed from the list but still
> > has FTRACE_OPS_FL_ENABLED and FTRACE_OPS_FL_SUBOP set. On an update
> > failure, it is added back to the list. Neither outcome matches
> > unregister_ftrace_graph(), which has already removed the graph slot.
> >
> > For example, enable a function_graph tracer with a nonempty filter and a
> > return fprobe. Fail an allocation in rebuild_hashes() while disabling the
> > fprobe, disable the tracer, then re-enable the fprobe. The stale SUBOP
> > state routes its filter update through the now-empty manager list.
> > rebuild_hashes() leaves its output pointers uninitialized, and ops_equal()
> > can crash when dereferencing them.
>
> Please show the example that can cause the above situation.
>
> >
> > Complete non-final subop removal on both error paths by clearing its
> > enabled/subop flags and manager pointer without adding it back to the
> > list. Initialize the rebuild outputs to EMPTY_HASH as well.
> >
> > A failed manager hash update may leave functions enabled that only the
> > removed operation traced. Keep per-subop filtering in that case, since
> > direct graph dispatch skips those checks. Restore direct dispatch after
> > a successful removal from two graph operations to one, or after a
> > successful first registration.
> >
> > This depends on the reader synchronization in the preceding patch.
>
> Every patch needs to be stand alone, and not mention other patches.
>
> > Clearing the enabled flag after a failure allows the caller to release the
> > subop filter, while an existing graph callback may still hold it. Backport
> > both patches together.
> >
> > Oops: general protection fault, probably for non-canonical address 0xdffffc0000000802: 0000 [#1] SMP KASAN NOPTI
> > KASAN: probably user-memory-access in range [0x0000000000004010-0x0000000000004017]
> > RIP: ops_equal (kernel/trace/trace.h:1011 kernel/trace/ftrace.c:3416)
> > Call Trace:
>
> What were the commands that caused the above?
>
> -- Steve
>
> > ftrace_update_ops (kernel/trace/ftrace.c:3464)
> > ftrace_hash_move_and_update_subops.isra.0 (kernel/trace/ftrace.c:3795)
> > ftrace_hash_move_and_update_ops (kernel/trace/ftrace.c:5095)
> > ftrace_set_hash (kernel/trace/ftrace.c:6006)
> > ftrace_set_filter_ips (kernel/trace/ftrace.c:6020 kernel/trace/ftrace.c:6769)
> > register_fprobe_ips (kernel/trace/fprobe.c:290 kernel/trace/fprobe.c:1020)
> > register_fprobe (kernel/trace/fprobe.c:979)
> > enable_trace_fprobe (kernel/trace/trace_fprobe.c:838 kernel/trace/trace_fprobe.c:1480)
> > fprobe_register (kernel/trace/trace_fprobe.c:1555)
> > __ftrace_event_enable_disable (kernel/trace/trace_events.c:913)
> > event_enable_write (kernel/trace/trace_events.c:943 kernel/trace/trace_events.c:1995)
> > vfs_write (fs/read_write.c:685)
> > ksys_write (fs/read_write.c:739)
> > do_syscall_64 (arch/x86/entry/syscall_64.c:61 arch/x86/entry/syscall_64.c:84)
> > entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
> > Kernel panic - not syncing: Fatal exception
> >
> > Cc: stable@xxxxxxxxxxxxxxx
> > Fixes: 0ae6b8ce200d ("ftrace: Fix accounting of subop hashes")
> > Reported-by: <co+bba0c6b49150bef3@xxxxxxx>
> > Assisted-by: LLM
> > Signed-off-by: Weiming Shi <bestswngs@xxxxxxxxx>
> > ---
> > kernel/trace/fgraph.c | 10 +++++++---
> > kernel/trace/ftrace.c | 22 ++++++++++++++--------
> > 2 files changed, 21 insertions(+), 11 deletions(-)
> >
> > diff --git a/kernel/trace/fgraph.c b/kernel/trace/fgraph.c
> > index 14733f519..48521bc33 100644
> > --- a/kernel/trace/fgraph.c
> > +++ b/kernel/trace/fgraph.c
> > @@ -1406,8 +1406,11 @@ int register_ftrace_graph(struct fgraph_ops *gops)
> > gops->ops.flags |= FTRACE_OPS_FL_GRAPH;
> >
> > ret = ftrace_startup_subops(&graph_ops, &gops->ops, command);
> > - if (!ret)
> > + if (!ret) {
> > fgraph_array[i] = gops;
> > + if (ftrace_graph_active == 1)
> > + ftrace_graph_enable_direct(true, gops);
> > + }
> >
> > error:
> > if (ret) {
> > @@ -1435,6 +1438,7 @@ int register_ftrace_graph(struct fgraph_ops *gops)
> > bool unregister_ftrace_graph_start(struct fgraph_ops *gops)
> > {
> > int command = 0;
> > + int ret = -EINVAL;
> > bool removed = false;
> >
> > if (WARN_ONCE(!(gops->ops.flags & FTRACE_OPS_FL_GRAPH),
> > @@ -1461,9 +1465,9 @@ bool unregister_ftrace_graph_start(struct fgraph_ops *gops)
> > if (!ftrace_graph_active)
> > command = FTRACE_STOP_FUNC_RET;
> >
> > - ftrace_shutdown_subops(&graph_ops, &gops->ops, command);
> > + ret = ftrace_shutdown_subops(&graph_ops, &gops->ops, command);
> >
> > - if (ftrace_graph_active == 1)
> > + if (ftrace_graph_active == 1 && !ret)
> > ftrace_graph_enable_direct(true, NULL);
> > else if (!ftrace_graph_active)
> > ftrace_graph_disable_direct(false);
> > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> > index c7d74b33d..7c712d99f 100644
> > --- a/kernel/trace/ftrace.c
> > +++ b/kernel/trace/ftrace.c
> > @@ -3672,6 +3672,8 @@ static int rebuild_hashes(struct ftrace_hash **filter_hash, struct ftrace_hash *
> > bool first = true;
> > int ret;
> >
> > + *filter_hash = EMPTY_HASH;
> > + *notrace_hash = EMPTY_HASH;
> > temp_hash.filter_hash = EMPTY_HASH;
> > temp_hash.notrace_hash = EMPTY_HASH;
> >
> > @@ -3713,7 +3715,13 @@ static int rebuild_hashes(struct ftrace_hash **filter_hash, struct ftrace_hash *
> > * will not affect functions that are being traced by other subops that
> > * still exist in @ops.
> > *
> > - * If the last subops is removed from @ops, then @ops is shutdown normally.
> > + * If @subops is the last subop, @ops is shut down normally. Otherwise,
> > + * @subops is detached even if the manager hash update fails. The manager
> > + * hashes may be partially updated, but still cover all functions required by
> > + * the remaining subops.
> > + *
> > + * Return: 0 on success, or a negative error code on shutdown or hash update
> > + * failure.
> > */
> > int ftrace_shutdown_subops(struct ftrace_ops *ops, struct ftrace_ops *subops, int command)
> > {
> > @@ -3753,15 +3761,13 @@ int ftrace_shutdown_subops(struct ftrace_ops *ops, struct ftrace_ops *subops, in
> > /* Rebuild the hashes without subops */
> > ret = rebuild_hashes(&filter_hash, &notrace_hash, ops);
> > if (ret < 0)
> > - return ret;
> > + goto out;
> >
> > ret = ftrace_update_ops(ops, filter_hash, notrace_hash);
> > - if (ret < 0) {
> > - list_add(&subops->list, &ops->subop_list);
> > - } else {
> > - subops->flags &= ~(FTRACE_OPS_FL_ENABLED | FTRACE_OPS_FL_SUBOP);
> > - subops->managed = NULL;
> > - }
> > +
> > +out:
> > + subops->flags &= ~(FTRACE_OPS_FL_ENABLED | FTRACE_OPS_FL_SUBOP);
> > + subops->managed = NULL;
> > free_ftrace_hash(filter_hash);
> > free_ftrace_hash(notrace_hash);
> > return ret;
>
Hi Steven,

Thanks for the feedback. I will address this in v2 and include the
reproduction PoC to show how the issue occurs.

Best regards,
Weiming