Skip to contentSkip to conversation

PM: sleep: Avoid using pr_cont() in tasks freezing code

PATCH In review Rafael J. Wysocki 2 participants 4 messages Dec 1, 2022 – Dec 6, 2022

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover PM: sleep: Avoid using pr_cont() in tasks freezing code

Hi All,

The first patch in this series rearranges the tasks freezing code used during system-wide suspend and hibernation to avoid printing continuation messages with pr_cont() which is problematic.

The second patch refines one more message in that code on top of the first one.

Thanks!

#1 lore raw

PATCH 1/2 PM: sleep: Avoid using pr_cont() in the tasks freezing code

From: Rafael J. Wysocki <[email protected]>

Using pr_cont() in the tasks freezing code related to system-wide suspend and hibernation is problematic, because the continuation messages printed there are susceptible to interspersing with other unrelated messages which results in output that is hard to understand.

Address this issue by modifying try_to_freeze_tasks() to print messages that don't require continuations and adjusting its callers accordingly.

---
 kernel/power/process.c |   21 ++++++++-------------
 1 file changed, 8 insertions(+), 13 deletions(-)
kernel/power/process.c +8 -13
Index: linux-pm/kernel/power/process.c
===================================================================
@@ -27,6 +27,8 @@ unsigned int __read_mostly freeze_timeou
 
 static int try_to_freeze_tasks(bool user_only)
 {
+	const char *what = user_only ? "user space processes" :
+					"remaining freezable tasks";
 	struct task_struct *g, *p;
 	unsigned long end_time;
 	unsigned int todo;
@@ -36,6 +38,8 @@ static int try_to_freeze_tasks(bool user
 	bool wakeup = false;
 	int sleep_usecs = USEC_PER_MSEC;
 
+	pr_info("Freezing %s\n", what);
+
 	start = ktime_get_boottime();
 
 	end_time = jiffies + msecs_to_jiffies(freeze_timeout_msecs);
@@ -82,7 +86,6 @@ static int try_to_freeze_tasks(bool user
 	elapsed_msecs = ktime_to_ms(elapsed);
 
 	if (todo) {
-		pr_cont("\n");
 		pr_err("Freezing of tasks %s after %d.%03d seconds "
 		       "(%d tasks refusing to freeze, wq_busy=%d):\n",
 		       wakeup ? "aborted" : "failed",
@@ -101,8 +104,8 @@ static int try_to_freeze_tasks(bool user
 			read_unlock(&tasklist_lock);
 		}
 	} else {
-		pr_cont("(elapsed %d.%03d seconds) ", elapsed_msecs / 1000,
-			elapsed_msecs % 1000);
+		pr_info("Freezing %s completed (elapsed %d.%03d seconds)\n",
+			what, elapsed_msecs / 1000, elapsed_msecs % 1000);
 	}
 
 	return todo ? -EBUSY : 0;
@@ -130,14 +133,11 @@ int freeze_processes(void)
 		static_branch_inc(&freezer_active);
 
 	pm_wakeup_clear(0);
-	pr_info("Freezing user space processes ... ");
 	pm_freezing = true;
 	error = try_to_freeze_tasks(true);
-	if (!error) {
+	if (!error)
 		__usermodehelper_set_disable_depth(UMH_DISABLED);
-		pr_cont("done.");
-	}
-	pr_cont("\n");
+
 	BUG_ON(in_atomic());
 
 	/*
@@ -166,14 +166,9 @@ int freeze_kernel_threads(void)
 {
 	int error;
 
-	pr_info("Freezing remaining freezable tasks ... ");
-
 	pm_nosig_freezing = true;
 	error = try_to_freeze_tasks(false);
-	if (!error)
-		pr_cont("done.");
 
-	pr_cont("\n");
 	BUG_ON(in_atomic());
 
 	if (error)
#2 lore raw

PATCH 2/2 PM: sleep: Refine error message in try_to_freeze_tasks()

In reply to Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

A previous change amended try_to_freeze_tasks() with the "what" variable pointing to a string describing the group of tasks subject to the freezing which may be used in the error message in there too, so make that happen.

Accordingly, update sleepgraph.py to catch the modified error message as appropriate.

---
 kernel/power/process.c             |    4 ++--
 tools/power/pm-graph/sleepgraph.py |    2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)
kernel/power/process.c +2 -2
Index: linux-pm/kernel/power/process.c
===================================================================
@@ -86,8 +86,8 @@ static int try_to_freeze_tasks(bool user
 	elapsed_msecs = ktime_to_ms(elapsed);
 
 	if (todo) {
-		pr_err("Freezing of tasks %s after %d.%03d seconds "
-		       "(%d tasks refusing to freeze, wq_busy=%d):\n",
+		pr_err("Freezing %s %s after %d.%03d seconds "
+		       "(%d tasks refusing to freeze, wq_busy=%d):\n", what,
 		       wakeup ? "aborted" : "failed",
 		       elapsed_msecs / 1000, elapsed_msecs % 1000,
 		       todo - wq_busy, wq_busy);
tools/power/pm-graph/sleepgraph.py +1 -1
Index: linux-pm/tools/power/pm-graph/sleepgraph.py
===================================================================
@@ -1462,7 +1462,7 @@ class Data:
 		'TIMEOUT' : r'(?i).*\bTIMEOUT\b.*',
 		'ABORT'   : r'(?i).*\bABORT\b.*',
 		'IRQ'     : r'.*\bgenirq: .*',
-		'TASKFAIL': r'.*Freezing of tasks *.*',
+		'TASKFAIL': r'.*Freezing .*after *.*',
 		'ACPI'    : r'.*\bACPI *(?P<b>[A-Za-z]*) *Error[: ].*',
 		'DISKFULL': r'.*\bNo space left on device.*',
 		'USBERR'  : r'.*usb .*device .*, error [0-9-]*',
Reviewed-by on the v1 series
#3 lore raw
In reply to Rafael J. Wysocki
Quoted 7 lines · Rafael J. Wysocki

Hi All,

The first patch in this series rearranges the tasks freezing code used during system-wide suspend and hibernation to avoid printing continuation messages with pr_cont() which is problematic.

The second patch refines one more message in that code on top of the first one.

Both patches seems to do what they are supposed to do. Feel free to use:

That said, I am not familiar with the freezer code. For example, I was not aware of linux-pm/tools/power/pm-graph/sleepgraph.py. Anyway, both changes make perfect sense and I do not see any problem.

Best Regards,
Petr