This is a note to let you know that I've just added the patch titled
tracing: Resize tgid_map to pid_max, not PID_MAX_DEFAULT
to the 5.12-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
tracing-resize-tgid_map-to-pid_max-not-pid_max_default.patch
and it can be found in the queue-5.12 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From 4030a6e6a6a4a42ff8c18414c9e0c93e24cc70b8 Mon Sep 17 00:00:00 2001
From: Paul Burton <paulburton@xxxxxxxxxx>
Date: Thu, 1 Jul 2021 10:24:07 -0700
Subject: tracing: Resize tgid_map to pid_max, not PID_MAX_DEFAULT
From: Paul Burton <paulburton@xxxxxxxxxx>
commit 4030a6e6a6a4a42ff8c18414c9e0c93e24cc70b8 upstream.
Currently tgid_map is sized at PID_MAX_DEFAULT entries, which means that
on systems where pid_max is configured higher than PID_MAX_DEFAULT the
ftrace record-tgid option doesn't work so well. Any tasks with PIDs
higher than PID_MAX_DEFAULT are simply not recorded in tgid_map, and
don't show up in the saved_tgids file.
In particular since systemd v243 & above configure pid_max to its
highest possible 1<<22 value by default on 64 bit systems this renders
the record-tgids option of little use.
Increase the size of tgid_map to the configured pid_max instead,
allowing it to cover the full range of PIDs up to the maximum value of
PID_MAX_LIMIT if the system is configured that way.
On 64 bit systems with pid_max == PID_MAX_LIMIT this will increase the
size of tgid_map from 256KiB to 16MiB. Whilst this 64x increase in
memory overhead sounds significant 64 bit systems are presumably best
placed to accommodate it, and since tgid_map is only allocated when the
record-tgid option is actually used presumably the user would rather it
spends sufficient memory to actually record the tgids they expect.
The size of tgid_map could also increase for CONFIG_BASE_SMALL=y
configurations, but these seem unlikely to be systems upon which people
are both configuring a large pid_max and running ftrace with record-tgid
anyway.
Of note is that we only allocate tgid_map once, the first time that the
record-tgid option is enabled. Therefore its size is only set once, to
the value of pid_max at the time the record-tgid option is first
enabled. If a user increases pid_max after that point, the saved_tgids
file will not contain entries for any tasks with pids beyond the earlier
value of pid_max.
Link: https://lkml.kernel.org/r/20210701172407.889626-2-paulburton@xxxxxxxxxx
Fixes: d914ba37d714 ("tracing: Add support for recording tgid of tasks")
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Joel Fernandes <joelaf@xxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Paul Burton <paulburton@xxxxxxxxxx>
[ Fixed comment coding style ]
Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
kernel/trace/trace.c | 63 ++++++++++++++++++++++++++++++++++++++-------------
1 file changed, 47 insertions(+), 16 deletions(-)
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -2184,8 +2184,15 @@ void tracing_reset_all_online_cpus(void)
}
}
+/*
+ * The tgid_map array maps from pid to tgid; i.e. the value stored at index i
+ * is the tgid last observed corresponding to pid=i.
+ */
static int *tgid_map;
+/* The maximum valid index into tgid_map. */
+static size_t tgid_map_max;
+
#define SAVED_CMDLINES_DEFAULT 128
#define NO_CMDLINE_MAP UINT_MAX
static arch_spinlock_t trace_cmdline_lock = __ARCH_SPIN_LOCK_UNLOCKED;
@@ -2458,24 +2465,41 @@ void trace_find_cmdline(int pid, char co
preempt_enable();
}
+static int *trace_find_tgid_ptr(int pid)
+{
+ /*
+ * Pairs with the smp_store_release in set_tracer_flag() to ensure that
+ * if we observe a non-NULL tgid_map then we also observe the correct
+ * tgid_map_max.
+ */
+ int *map = smp_load_acquire(&tgid_map);
+
+ if (unlikely(!map || pid > tgid_map_max))
+ return NULL;
+
+ return &map[pid];
+}
+
int trace_find_tgid(int pid)
{
- if (unlikely(!tgid_map || !pid || pid > PID_MAX_DEFAULT))
- return 0;
+ int *ptr = trace_find_tgid_ptr(pid);
- return tgid_map[pid];
+ return ptr ? *ptr : 0;
}
static int trace_save_tgid(struct task_struct *tsk)
{
+ int *ptr;
+
/* treat recording of idle task as a success */
if (!tsk->pid)
return 1;
- if (unlikely(!tgid_map || tsk->pid > PID_MAX_DEFAULT))
+ ptr = trace_find_tgid_ptr(tsk->pid);
+ if (!ptr)
return 0;
- tgid_map[tsk->pid] = tsk->tgid;
+ *ptr = tsk->tgid;
return 1;
}
@@ -4915,6 +4939,8 @@ int trace_keep_overwrite(struct tracer *
int set_tracer_flag(struct trace_array *tr, unsigned int mask, int enabled)
{
+ int *map;
+
if ((mask == TRACE_ITER_RECORD_TGID) ||
(mask == TRACE_ITER_RECORD_CMD))
lockdep_assert_held(&event_mutex);
@@ -4937,10 +4963,19 @@ int set_tracer_flag(struct trace_array *
trace_event_enable_cmd_record(enabled);
if (mask == TRACE_ITER_RECORD_TGID) {
- if (!tgid_map)
- tgid_map = kvcalloc(PID_MAX_DEFAULT + 1,
- sizeof(*tgid_map),
- GFP_KERNEL);
+ if (!tgid_map) {
+ tgid_map_max = pid_max;
+ map = kvcalloc(tgid_map_max + 1, sizeof(*tgid_map),
+ GFP_KERNEL);
+
+ /*
+ * Pairs with smp_load_acquire() in
+ * trace_find_tgid_ptr() to ensure that if it observes
+ * the tgid_map we just allocated then it also observes
+ * the corresponding tgid_map_max value.
+ */
+ smp_store_release(&tgid_map, map);
+ }
if (!tgid_map) {
tr->trace_flags &= ~TRACE_ITER_RECORD_TGID;
return -ENOMEM;
@@ -5354,18 +5389,14 @@ static void *saved_tgids_next(struct seq
{
int pid = ++(*pos);
- if (pid > PID_MAX_DEFAULT)
- return NULL;
-
- return &tgid_map[pid];
+ return trace_find_tgid_ptr(pid);
}
static void *saved_tgids_start(struct seq_file *m, loff_t *pos)
{
- if (!tgid_map || *pos > PID_MAX_DEFAULT)
- return NULL;
+ int pid = *pos;
- return &tgid_map[*pos];
+ return trace_find_tgid_ptr(pid);
}
static void saved_tgids_stop(struct seq_file *m, void *v)
Patches currently in stable-queue which might be from paulburton@xxxxxxxxxx are
queue-5.12/tracing-simplify-fix-saved_tgids-logic.patch
queue-5.12/tracing-resize-tgid_map-to-pid_max-not-pid_max_default.patch
