linux-kernel - [PATCH 2/2] tracing/irqsoff: Balance prolog on fgraph failure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251006175848.1906912-2-sashal@kernel.org>
Date: Mon,  6 Oct 2025 13:58:48 -0400
From: Sasha Levin <sashal@...nel.org>
To: rostedt@...dmis.org,
	mhiramat@...nel.org
Cc: mathieu.desnoyers@...icios.com,
	linux-kernel@...r.kernel.org,
	linux-trace-kernel@...r.kernel.org,
	Sasha Levin <sashal@...nel.org>
Subject: [PATCH 2/2] tracing/irqsoff: Balance prolog on fgraph failure

Commit a485ea9e3ef3 ("tracing: Fix irqsoff and wakeup latency tracers when
using function graph") added fgraph_reserve_data()/fgraph_retrieve_data() to
the irqsoff tracer's function-graph hooks.

When func_prolog_dec() succeeded but those calls failed, the code returned
without undoing the prolog (no local_dec(&data->disabled)). While
func_prolog_dec() does not disable preemption, this left the tracer recursion
guard (data->disabled) incremented on this CPU, suppressing subsequent tracing
and leaving the tracer state inconsistent.

Balance the prolog on these failure paths in both entry and return hooks.

Fixes: a485ea9e3ef3 ("tracing: Fix irqsoff and wakeup latency tracers when using function graph")
Assisted-by: gpt-5-codex
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 kernel/trace/trace_irqsoff.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c
index 5496758b6c760..39152ef3d2432 100644
--- a/kernel/trace/trace_irqsoff.c
+++ b/kernel/trace/trace_irqsoff.c
@@ -202,8 +202,10 @@ static int irqsoff_graph_entry(struct ftrace_graph_ent *trace,
 		return 0;
 
 	calltime = fgraph_reserve_data(gops->idx, sizeof(*calltime));
-	if (!calltime)
+	if (!calltime) {
+		local_dec(&data->disabled);
 		return 0;
+	}
 
 	*calltime = trace_clock_local();
 
@@ -233,8 +235,10 @@ static void irqsoff_graph_return(struct ftrace_graph_ret *trace,
 
 	rettime = trace_clock_local();
 	calltime = fgraph_retrieve_data(gops->idx, &size);
-	if (!calltime)
+	if (!calltime) {
+		local_dec(&data->disabled);
 		return;
+	}
 
 	trace_ctx = tracing_gen_ctx_flags(flags);
 	__trace_graph_return(tr, trace, trace_ctx, *calltime, rettime);
-- 
2.51.0