[PATCH v4 0/2] tracing: fgraph: fix the quadratic shadow stack walk

[PATCH v4 0/2] tracing: fgraph: fix the quadratic shadow stack walk

[Vineet Gupta]

alloc_retstack_tasklist() assigns a shadow stack to every thread when
fgraph is enabled. It did so FTRACE_RETSTACK_ALLOC_SIZE (32) tasks at a
time, and since for_each_process_thread() has no cursor, every sweep
restarted from init_task and re-walked the tasks already served, making
the ftrace_graph_active 0 -> 1 transition quadratic in the thread count.
On a 60-core Sapphire Rapids machine with 400000 idle threads it took
227 s, inside a single bpf() syscall for the kprobe_multi case.

 - 1/2 allocates inline with GFP_NOWAIT, which removes the batching:
   a single sweep serves every task, 227 s -> 0.092 s. It keeps the
   open-coded rcu_read_lock() for ease of backport.

 - 2/2 is the cleanup: convert that same walk to scoped_guard(rcu).
   No functional change, and not needed by stable.

Both were tested on bare metal and in a VM, including forcing the
GFP_NOWAIT failure path with failslab injection.

v1: https://lore.kernel.org/all/20260922225526.1554758-1-vineet.gupta@linux.dev/
v2: https://lore.kernel.org/all/20260929005411.4105448-1-vineet.gupta@linux.dev/
v3: https://lore.kernel.org/all/20261001102909.1152572-1-vineet.gupta@linux.dev/

Changes since v3:
 - split into two patches: the fix keeps goto-based unwinding so it
   backports cleanly, and scoped_guard(rcu) moves to a separate
   cleanup, per Masami and Steve:
   https://lore.kernel.org/all/20261013194542.7e036b03@226.227.38.69.naturalwireless.com/
 - in the cleanup, -EAGAIN now exits via the existing free label
   rather than returning from inside the guard, so there is no longer
   a path that skips the free loop

Changes since v2:
 - update the stale alloc_retstack_tasklist() comment, which still
   described the old batch-of-32 behaviour (bot+bpf-ci)

Changes since v1:
 - replace both v1 patches with Peter's inline GFP_NOWAIT approach,
   which removes the quadratic behaviour rather than dividing it by a
   constant. With a single sweep there is no retry loop left, so the
   v1 cond_resched() patch has nothing to attach to and is dropped.

Vineet Gupta (2):
  tracing: fgraph: allocate shadow stacks inline with GFP_NOWAIT
  tracing: fgraph: use scoped_guard() for the task list walk

 kernel/trace/fgraph.c | 38 +++++++++++++++++++++++---------------
 1 file changed, 23 insertions(+), 15 deletions(-)


base-commit: c03c1a14507dfaf8b856c2608f9ac216021fc222
-- 
2.53.0-Meta

[PATCH v4 1/2] tracing: fgraph: allocate shadow stacks inline with GFP_NOWAIT

[Vineet Gupta]

When ftrace graphing is turned on, all tasks in the system missing
return stack page are assigned one. This is done in a simplistic
multi-sweep loop of FTRACE_RETSTACK_ALLOC_SIZE (currently 32) tasks
at a time as follows:

   start_graph_tracing()
	do {
		alloc_retstack_tasklist
	} while (-EAGAIN);

   alloc_retstack_tasklist()
	alloc x32           # GFP_KERNEL, may sleep
	rcu_read_lock()     # preempt off
		for_each_process_thread    walk N_total, no cond_resched
                     t->ret_stack = new_page
	rcu_read_unlock()   # preempt enable but no explicit yield

Each successive iteration of loop invokes for_each_process_thread()
which doesn't support cursor based resume and always restarts from the
init_task. Thus each successive loop needs to skip the tasks assigned
ret_stack in prior sweeps and thus take longer and longer to find the
candidate 32 tasks. And while the loop end calls rcu unlock,
and re-enables preemption briefly, there is no explicit yield.

The total number of iterations turn out to be

	N_total * N_null / (2 * FTRACE_RETSTACK_ALLOC_SIZE)

where N_total is the number of threads on the system and N_null the number
of those whose ret_stack is still NULL. After boot with no fgraph user that
is every thread.

This is not a tracing-only path. Since commit 4346ba160409 ("fprobe:
Rewrite fprobe on function-graph tracer") fprobe is built on fgraph,
so an ordinary kprobe_multi BPF attach that flips ftrace_graph_active
from 0 to 1 pays the whole cost inside one bpf() syscall. Commit
2c67dc457bc6 ("tracing: fprobe: optimization for entry only case")
narrows that to return and session probes but does not remove it.

A host running hundreds of thousands of threads, with the current batch
size of 32, turns the ftrace_graph_active 0 -> 1 transition into a
multi-second, and eventually multi-minute, operation which showed up as
RCU stall and softlockup_panic on heavily loaded Meta Fleet machines
with preemption disabled.

    rcu: INFO: rcu_sched self-detected stall on CPU
    rcu: 0-....: (20718 ticks this GP) (t=21000 jiffies g=913 q=8 ncpus=8)
    CPU: 0 UID: 0 PID: 160141 Comm: bpftrace
     ___slab_alloc+0x549/0xa20
     kmem_cache_alloc_noprof+0x16c/0x340
     register_ftrace_graph+0x3d7/0x6c0
     register_fprobe_ips+0x2d5/0x310
     bpf_kprobe_multi_link_attach+0x218/0x7e0
     __sys_bpf+0x267a/0x27a0

Fix this by allocating inline with GFP_NOWAIT instead of pre-allocating
a fixed batch. GFP_NOWAIT does not sleep, so it is legal inside the RCU
read-side critical section.

A single sweep now serves every task, so the walk becomes O(N) rather
than O(N_total * N_null), and the retry loop is no longer the normal
path.

The pre-allocated array remains as the fallback when GFP_NOWAIT fails.
The runs below never needed it, so it was exercised with failslab
injection: at a 100% failure rate every task falls back and -EAGAIN
retries ~1560 times for 50000 threads, with no stall or splat.

The open-coded rcu_read_lock() and the goto unwind are kept as-is so
this stays easy to backport; converting them to scope-based cleanup is
left to a follow-up.

smp_store_release() is a drop-in replacement for the smp_wmb() and
plain store it replaces.

Measured on a 60-core Sapphire Rapids machine running a PREEMPT_LAZY
kernel (CONFIG_PREEMPT_DYNAMIC=y, mode lazy), timing the write that
drives the ftrace_graph_active 0 -> 1 transition. Threads are spawned
while fgraph is inactive so every one of them has ret_stack == NULL,
and the loop was instrumented to count sweeps:

	threads   before              after
	          sweeps   time       sweeps   time
	100000     3126     15.3 s      1       0.019 s
	200000     6253     59.1 s      1       0.047 s
	400000    12503    227.8 s      1       0.092 s

The sweep count is 1 throughout. Time now scales linearly with the
thread count at roughly 230 ns per task, which is the allocation and
initialisation of one shadow stack.

Fixes: f201ae2356c7 ("tracing/function-return-tracer: store return stack into task_struct and allocate it dynamically")
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Vineet Gupta <vineet.gupta@linux.dev>
Tested-by: Vineet Gupta <vineet.gupta@linux.dev>
Cc: stable@vger.kernel.org
---
 kernel/trace/fgraph.c | 39 ++++++++++++++++++++++++---------------
 1 file changed, 24 insertions(+), 15 deletions(-)

diff --git a/kernel/trace/fgraph.c b/kernel/trace/fgraph.c
index ed455b53513b..4c74f1a15ad4 100644
--- a/kernel/trace/fgraph.c
+++ b/kernel/trace/fgraph.c
@@ -1033,13 +1033,16 @@ void ftrace_stub_graph(struct ftrace_graph_ret *trace, struct fgraph_ops *gops,
 trace_func_graph_ret_t ftrace_graph_return = ftrace_stub_graph;
 trace_func_graph_ent_t ftrace_graph_entry = ftrace_graph_entry_stub;
 
-/* Try to assign a return stack array on FTRACE_RETSTACK_ALLOC_SIZE tasks. */
+/*
+ * Loop thru all the tasks in the system and assign ret_stack.
+ * The @ret_stack_list is a small pre-allocated fallback, in case
+ * the GFP_NOWAIT allocations here fail.
+ */
 static int alloc_retstack_tasklist(unsigned long **ret_stack_list)
 {
-	int i;
-	int ret = 0;
 	int start = 0, end = FTRACE_RETSTACK_ALLOC_SIZE;
 	struct task_struct *g, *t;
+	int i, ret = 0;
 
 	if (WARN_ON_ONCE(!fgraph_stack_cachep))
 		return -ENOMEM;
@@ -1056,20 +1059,26 @@ static int alloc_retstack_tasklist(unsigned long **ret_stack_list)
 
 	rcu_read_lock();
 	for_each_process_thread(g, t) {
-		if (start == end) {
-			ret = -EAGAIN;
-			goto unlock;
-		}
+		unsigned long *rs;
 
-		if (t->ret_stack == NULL) {
-			atomic_set(&t->trace_overrun, 0);
-			ret_stack_init_task_vars(ret_stack_list[start]);
-			t->curr_ret_stack = 0;
-			t->curr_ret_depth = -1;
-			/* Make sure the tasks see the 0 first: */
-			smp_wmb();
-			t->ret_stack = ret_stack_list[start++];
+		if (t->ret_stack)
+			continue;
+
+		rs = kmem_cache_alloc(fgraph_stack_cachep, GFP_NOWAIT);
+		if (!rs) {
+			if (start == end) {
+				ret = -EAGAIN;
+				goto unlock;
+			}
+			rs = ret_stack_list[start++];
 		}
+
+		atomic_set(&t->trace_overrun, 0);
+		ret_stack_init_task_vars(rs);
+		t->curr_ret_stack = 0;
+		t->curr_ret_depth = -1;
+		/* Make sure the tasks see the 0 first: */
+		smp_store_release(&t->ret_stack, rs);
 	}
 
 unlock:

base-commit: c03c1a14507dfaf8b856c2608f9ac216021fc222
-- 
2.53.0-Meta

[PATCH v4 2/2] tracing: fgraph: use scoped_guard() for the task list walk

[Vineet Gupta]

alloc_retstack_tasklist() open-codes rcu_read_lock() and unwinds it
through an unlock label. Use scoped_guard(rcu) instead and let the
compiler emit the unlock on every exit path.

The -EAGAIN case now jumps straight to the free label rather than to
unlock; leaving the guard's scope releases the RCU read lock on the
way, so the unlock label has no remaining users and is removed.

No functional change: objdump of fgraph.o shows identical call targets
and counts, including the three __rcu_read_unlock() sites the compiler
emits for the exits from the read-side section. A PROVE_RCU +
DEBUG_ATOMIC_SLEEP build with the -EAGAIN path forced several thousand
times is clean.

The preceding patch deliberately left this alone so that it stays
straightforward to backport. This is the follow-up.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Vineet Gupta <vineet.gupta@linux.dev>
Tested-by: Vineet Gupta <vineet.gupta@linux.dev>
---
 kernel/trace/fgraph.c | 39 +++++++++++++++++++--------------------
 1 file changed, 19 insertions(+), 20 deletions(-)

diff --git a/kernel/trace/fgraph.c b/kernel/trace/fgraph.c
index 4c74f1a15ad4..bbf6e4e2da2d 100644
--- a/kernel/trace/fgraph.c
+++ b/kernel/trace/fgraph.c
@@ -1057,32 +1057,31 @@ static int alloc_retstack_tasklist(unsigned long **ret_stack_list)
 		}
 	}
 
-	rcu_read_lock();
-	for_each_process_thread(g, t) {
-		unsigned long *rs;
+	scoped_guard (rcu) {
+		for_each_process_thread(g, t) {
+			unsigned long *rs;
 
-		if (t->ret_stack)
-			continue;
+			if (t->ret_stack)
+				continue;
 
-		rs = kmem_cache_alloc(fgraph_stack_cachep, GFP_NOWAIT);
-		if (!rs) {
-			if (start == end) {
-				ret = -EAGAIN;
-				goto unlock;
+			rs = kmem_cache_alloc(fgraph_stack_cachep, GFP_NOWAIT);
+			if (!rs) {
+				if (start == end) {
+					ret = -EAGAIN;
+					goto free;
+				}
+				rs = ret_stack_list[start++];
 			}
-			rs = ret_stack_list[start++];
-		}
 
-		atomic_set(&t->trace_overrun, 0);
-		ret_stack_init_task_vars(rs);
-		t->curr_ret_stack = 0;
-		t->curr_ret_depth = -1;
-		/* Make sure the tasks see the 0 first: */
-		smp_store_release(&t->ret_stack, rs);
+			atomic_set(&t->trace_overrun, 0);
+			ret_stack_init_task_vars(rs);
+			t->curr_ret_stack = 0;
+			t->curr_ret_depth = -1;
+			/* Make sure the tasks see the 0 first: */
+			smp_store_release(&t->ret_stack, rs);
+		}
 	}
 
-unlock:
-	rcu_read_unlock();
 free:
 	for (i = start; i < end; i++)
 		kmem_cache_free(fgraph_stack_cachep, ret_stack_list[i]);
-- 
2.53.0-Meta

Re: [PATCH v4 0/2] tracing: fgraph: fix the quadratic shadow stack walk

[Alexei Starovoitov]

On Wed, Oct 21, 2026 at 10:44 AM Vineet Gupta <vineet.gupta@linux.dev> wrote:
> On a 60-core Sapphire Rapids machine with 400000 idle threads it took
> 227 s, inside a single bpf() syscall for the kprobe_multi case.

Acked-by: Alexei Starovoitov <ast@kernel.org>