Skip to contentSkip to conversation

tracing: Fixes for 6.11

PATCH New Steven Rostedt 1 participant 3 messages Aug 16, 2024

Subsystems (3)
lore
Share & tools
lore

Steven Rostedt

author maintainer
#0 lore raw

cover tracing: Fixes for 6.11

A couple of fixes for tracing:

- Prevent a NULL pointer dereference in the error path of rtla

- Fix an infinite loop bug when reading from the ring buffer when closed. If there's a thread trying to read the ring buffer and it gets closed by another thread, the one reading will go into an infinite loop when the buffer is empty instead of exiting back to user space.

git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace.git
trace/urgent

Head SHA1: 90574d2a675947858b47008df8d07f75ea50d0d0

Dan Carpenter (1):
rtla/osnoise: Prevent NULL dereference in error handling

Steven Rostedt (1):
tracing: Return from tracing_buffers_read() if the file has been closed

----

 kernel/trace/trace.c                 |  2 +-
 tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
 2 files changed, 5 insertions(+), 8 deletions(-)
author maintainer
#1 lore raw

PATCH 1/2 tracing: Return from tracing_buffers_read() if the file has been closed

From: Steven Rostedt <[email protected]>

When running the following:

 # cd /sys/kernel/tracing/
 # echo 1 > events/sched/sched_waking/enable
 # echo 1 > events/sched/sched_switch/enable
 # echo 0 > tracing_on
 # dd if=per_cpu/cpu0/trace_pipe_raw of=/tmp/raw0.dat

The dd task would get stuck in an infinite loop in the kernel. What would happen is the following:

When ring_buffer_read_page() returns -1 (no data) then a check is made to see if the buffer is empty (as happens when the page is not full), it will call wait_on_pipe() to wait until the ring buffer has data. When it is it will try again to read data (unless O_NONBLOCK is set).

The issue happens when there's a reader and the file descriptor is closed. The wait_on_pipe() will return when that is the case. But this loop will continue to try again and wait_on_pipe() will again return immediately and the loop will continue and never stop.

Simply check if the file was closed before looping and exit out if it is.

Cc Masami Hiramatsu <[email protected]>
Cc Mathieu Desnoyers <[email protected]>
Fixes 2aa043a55b9a7 ("tracing/ring-buffer: Fix wait_on_pipe() race")
---
 kernel/trace/trace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/trace/trace.c +1 -1
index 10cd38bce2f1..ebe7ce2f5f4a 100644
@@ -7956,7 +7956,7 @@ tracing_buffers_read(struct file *filp, char __user *ubuf,
 	trace_access_unlock(iter->cpu_file);
 
 	if (ret < 0) {
-		if (trace_empty(iter)) {
+		if (trace_empty(iter) && !iter->closed) {
 			if ((filp->f_flags & O_NONBLOCK))
 				return -EAGAIN;
 
Signature · 1 line
2.43.0
author maintainer
#2 lore raw

PATCH 2/2 rtla/osnoise: Prevent NULL dereference in error handling

In reply to Steven Rostedt

From: Dan Carpenter <[email protected]>

If the "tool->data" allocation fails then there is no need to call osnoise_free_top() and, in fact, doing so will lead to a NULL dereference.

Cc John Kacur <[email protected]>
Cc "Luis Claudio R. Goncalves" <[email protected]>
Cc Clark Williams <[email protected]>
Fixes 1eceb2fc2ca5 ("rtla/osnoise: Add osnoise top mode")
---
 tools/tracing/rtla/src/osnoise_top.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
tools/tracing/rtla/src/osnoise_top.c +4 -7
index f594a44df840..2f756628613d 100644
@@ -651,8 +651,10 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
 		return NULL;
 
 	tool->data = osnoise_alloc_top(nr_cpus);
-	if (!tool->data)
-		goto out_err;
+	if (!tool->data) {
+		osnoise_destroy_tool(tool);
+		return NULL;
+	}
 
 	tool->params = params;
 
@@ -660,11 +662,6 @@ struct osnoise_tool *osnoise_init_top(struct osnoise_top_params *params)
 				   osnoise_top_handler, NULL);
 
 	return tool;
-
-out_err:
-	osnoise_free_top(tool->data);
-	osnoise_destroy_tool(tool);
-	return NULL;
 }
 
 static int stop_tracing;
Signature · 1 line
2.43.0