Skip to contentSkip to conversation

timer: fix timer_migration to accept only 0 and 1

PATCH In review Myungho Jung 3 participants 6 messages Apr 19, 2017 – Apr 20, 2017

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 timer: fix timer_migration to accept only 0 and 1

Error is not shown by setting invalid value to timer_migration. Valid values for timer_migration should be restricted to 0 and 1. Testcase for this bug is ltp/runpwtests06.

---
 kernel/sysctl.c     | 2 ++
 kernel/time/timer.c | 2 +-
 2 files changed, 3 insertions(+), 1 deletion(-)
kernel/sysctl.c +2 -0
index 8c8714f..21343d1 100644
@@ -1176,6 +1176,8 @@ static struct ctl_table kern_table[] = {
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0644,
 		.proc_handler	= timer_migration_handler,
+		.extra1		= &zero,
+		.extra2		= &one,
 	},
 #endif
 #ifdef CONFIG_BPF_SYSCALL
kernel/time/timer.c +1 -1
index 1dc0256..cc6b6bd 100644
@@ -241,7 +241,7 @@ int timer_migration_handler(struct ctl_table *table, int write,
 	int ret;
 
 	mutex_lock(&mutex);
-	ret = proc_dointvec(table, write, buffer, lenp, ppos);
+	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 	if (!ret && write)
 		timers_update_migration(false);
 	mutex_unlock(&mutex);
Signature · 1 line
2.7.4
#1 lore raw
Quoted 3 lines · Myungho Jung

Error is not shown by setting invalid value to timer_migration. Valid values for timer_migration should be restricted to 0 and 1. Testcase for this bug is ltp/runpwtests06.

While I agree with the change, I disagree with the changelog. Where is the bug?

The timer code checks for timer_migration != 0 resp. == 0. So as long as the value is != 0 it's enabled, if it's 0 it is disabled.

It's a correctness issue that we treat a sysctl which is basically a boolean as such. Ideally we'd have: proc_doboolvec() for such cases and convert all similar sysctls over to that.

Thanks,

tglx

#2 lore raw
Quoted 11 lines · Thomas Gleixner
Quoted 3 lines · Myungho Jung

Error is not shown by setting invalid value to timer_migration. Valid values for timer_migration should be restricted to 0 and 1. Testcase for this bug is ltp/runpwtests06.

While I agree with the change, I disagree with the changelog. Where is the bug?

The timer code checks for timer_migration != 0 resp. == 0. So as long as the value is != 0 it's enabled, if it's 0 it is disabled.

Quoted 7 lines

It's a correctness issue that we treat a sysctl which is basically a boolean as such. Ideally we'd have: proc_doboolvec() for such cases and convert all similar sysctls over to that.

Thanks,

tglx

I understood. I'll resubmit the patch after fixing log.

Thanks,
Myungho

#3 lore raw
Quoted 1 line · Myungho Jung

I understood. I'll resubmit the patch after fixing log.

I recommend to check your inbox first.

Thanks,

tglx

#4 lore raw
Quoted 8 lines · Thomas Gleixner
Quoted 1 line · Myungho Jung

I understood. I'll resubmit the patch after fixing log.

I recommend to check your inbox first.

Thanks,

tglx

Yes, I confirmed a message from tip-bot. I'll refer to it.

Thanks,
Myungho

#5 lore raw

[tip:timers/core] timer/sysclt: Restrict timer migration sysctl values to 0 and 1

In reply to Myungho Jung
timer/sysclt: Restrict timer migration sysctl values to 0 and 1
Commit-ID:  b94bf594cf8ed67cdd0439e70fa939783471597a
Gitweb:     http://git.kernel.org/tip/b94bf594cf8ed67cdd0439e70fa939783471597a
Author:     Myungho Jung <[email protected]>
AuthorDate: Wed, 19 Apr 2017 15:24:50 -0700
Committer:  Thomas Gleixner <[email protected]>
CommitDate: Thu, 20 Apr 2017 14:56:59 +0200

timer/sysclt: Restrict timer migration sysctl values to 0 and 1

timer_migration sysctl acts as a boolean switch, so the allowed values should be restricted to 0 and 1.

Add the necessary extra fields to the sysctl table entry to enforce that.

[ tglx: Rewrote changelog ]

---
 kernel/sysctl.c     | 2 ++
 kernel/time/timer.c | 2 +-
 2 files changed, 3 insertions(+), 1 deletion(-)
kernel/sysctl.c +2 -0
index acf0a5a..0863769 100644
@@ -1176,6 +1176,8 @@ static struct ctl_table kern_table[] = {
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0644,
 		.proc_handler	= timer_migration_handler,
+		.extra1		= &zero,
+		.extra2		= &one,
 	},
 #endif
 #ifdef CONFIG_BPF_SYSCALL
kernel/time/timer.c +1 -1
index 1dc0256..cc6b6bd 100644
@@ -241,7 +241,7 @@ int timer_migration_handler(struct ctl_table *table, int write,
 	int ret;
 
 	mutex_lock(&mutex);
-	ret = proc_dointvec(table, write, buffer, lenp, ppos);
+	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 	if (!ret && write)
 		timers_update_migration(false);
 	mutex_unlock(&mutex);