Skip to contentSkip to conversation

signal: improve warning about using SI_TKILL in rt_[tg]sigqueueinfo

PATCH In review Vladimir Davydov 2 participants 4 messages Mar 19, 2015

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 signal: improve warning about using SI_TKILL in rt_[tg]sigqueueinfo

Sending SI_TKILL from rt_[tg]sigqueueinfo was deprecated, so now we issue a warning on the first attempt of doing it. We use WARN_ON_ONCE, which is not informative and, what is worse, taints the kernel, making the trinity syscall fuzzer complain false-positively from time to time.

This patch therefore substitutes the WARN_ON_ONCE with a pr_warn_once.

---
 kernel/signal.c |   16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
kernel/signal.c +11 -5
index a390499943e4..3cbcd94457af 100644
@@ -2993,8 +2993,11 @@ static int do_rt_sigqueueinfo(pid_t pid, int sig, siginfo_t *info)
 	 */
 	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
 	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+		/* We used to allow SI_TKILL */
+		if (info->si_code == SI_TKILL)
+			pr_warn_once("%s (%d): Sending SI_TKILL from "
+				     "rt_sigqueueinfo is deprecated\n",
+				     current->comm, current->pid);
 		return -EPERM;
 	}
 	info->si_signo = sig;
@@ -3041,10 +3044,13 @@ static int do_rt_tgsigqueueinfo(pid_t tgid, pid_t pid, int sig, siginfo_t *info)
 	/* Not even root can pretend to send signals from the kernel.
 	 * Nor can they impersonate a kill()/tgkill(), which adds source info.
 	 */
-	if (((info->si_code >= 0 || info->si_code == SI_TKILL)) &&
+	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
 	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+		/* We used to allow SI_TKILL */
+		if (info->si_code == SI_TKILL)
+			pr_warn_once("%s (%d): Sending SI_TKILL from "
+				     "rt_tgsigqueueinfo is deprecated\n",
+				     current->comm, current->pid);
 		return -EPERM;
 	}
 	info->si_signo = sig;
Signature · 1 line
1.7.10.4
#1 lore raw
Quoted 7 lines · Vladimir Davydov

Sending SI_TKILL from rt_[tg]sigqueueinfo was deprecated, so now we issue a warning on the first attempt of doing it. We use WARN_ON_ONCE, which is not informative and, what is worse, taints the kernel, making the trinity syscall fuzzer complain false-positively from time to time.

This patch therefore substitutes the WARN_ON_ONCE with a pr_warn_once.

Agreed.

but perhaps we can simply remove this warning at all?

Quoted 43 lines
---
 kernel/signal.c |   16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
kernel/signal.c +11 -5
index a390499943e4..3cbcd94457af 100644
@@ -2993,8 +2993,11 @@ static int do_rt_sigqueueinfo(pid_t pid, int sig, siginfo_t *info)
 	 */
 	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
 	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+		/* We used to allow SI_TKILL */
+		if (info->si_code == SI_TKILL)
+			pr_warn_once("%s (%d): Sending SI_TKILL from "
+				     "rt_sigqueueinfo is deprecated\n",
+				     current->comm, current->pid);
 		return -EPERM;
 	}
 	info->si_signo = sig;
@@ -3041,10 +3044,13 @@ static int do_rt_tgsigqueueinfo(pid_t tgid, pid_t pid, int sig, siginfo_t *info)
 	/* Not even root can pretend to send signals from the kernel.
 	 * Nor can they impersonate a kill()/tgkill(), which adds source info.
 	 */
-	if (((info->si_code >= 0 || info->si_code == SI_TKILL)) &&
+	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
 	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+		/* We used to allow SI_TKILL */
+		if (info->si_code == SI_TKILL)
+			pr_warn_once("%s (%d): Sending SI_TKILL from "
+				     "rt_tgsigqueueinfo is deprecated\n",
+				     current->comm, current->pid);
 		return -EPERM;
 	}
 	info->si_signo = sig;
Signature · 1 line
1.7.10.4
#2 lore raw
Quoted 7 lines · Oleg Nesterov
Quoted 6 lines · Vladimir Davydov

Sending SI_TKILL from rt_[tg]sigqueueinfo was deprecated, so now we issue a warning on the first attempt of doing it. We use WARN_ON_ONCE, which is not informative and, what is worse, taints the kernel, making the trinity syscall fuzzer complain false-positively from time to time.

Quoted 5 lines
Quoted 1 line

This patch therefore substitutes the WARN_ON_ONCE with a pr_warn_once.

Agreed.

but perhaps we can simply remove this warning at all?

We can, I suppose. Personally, I do not have any strong preference. The patch removing the warning is attached.
---
From: Vladimir Davydov <[email protected]>
Subject: [PATCH] signal: remove warning about using SI_TKILL in
rt_[tg]sigqueueinfo

Sending SI_TKILL from rt_[tg]sigqueueinfo was deprecated, so now we issue a warning on the first attempt of doing it. We use WARN_ON_ONCE, which is not informative and, what is worse, taints the kernel, making the trinity syscall fuzzer complain false-positively from time to time.

It does not look like we need this warning at all, because the behaviour changed quite a long time ago (2.6.39), and if an application relies on the old API, it gets EPERM anyway and can issue a warning by itself.

So let us zap the warning in kernel.

kernel/signal.c +5 -9
index a390499943e4..d51c5ddd855c 100644
@@ -2992,11 +2992,9 @@ static int do_rt_sigqueueinfo(pid_t pid, int sig, siginfo_t *info)
 	 * Nor can they impersonate a kill()/tgkill(), which adds source info.
 	 */
 	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
-	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+	    (task_pid_vnr(current) != pid))
 		return -EPERM;
-	}
+
 	info->si_signo = sig;
 
 	/* POSIX.1b doesn't mention process groups.  */
@@ -3041,12 +3039,10 @@ static int do_rt_tgsigqueueinfo(pid_t tgid, pid_t pid, int sig, siginfo_t *info)
 	/* Not even root can pretend to send signals from the kernel.
 	 * Nor can they impersonate a kill()/tgkill(), which adds source info.
 	 */
-	if (((info->si_code >= 0 || info->si_code == SI_TKILL)) &&
-	    (task_pid_vnr(current) != pid)) {
-		/* We used to allow any < 0 si_code */
-		WARN_ON_ONCE(info->si_code < 0);
+	if ((info->si_code >= 0 || info->si_code == SI_TKILL) &&
+	    (task_pid_vnr(current) != pid))
 		return -EPERM;
-	}
+
 	info->si_signo = sig;
 
 	return do_send_specific(tgid, pid, sig, info);
Acked-by
#3 lore raw
Quoted 8 lines · Vladimir Davydov
Quoted 4 lines · Oleg Nesterov

Agreed.

but perhaps we can simply remove this warning at all?

We can, I suppose. Personally, I do not have any strong preference.

Me too, but personally I like this version more ;)

Quoted 9 lines

From: Vladimir Davydov <[email protected]>
Subject: [PATCH] signal: remove warning about using SI_TKILL in
rt_[tg]sigqueueinfo

Sending SI_TKILL from rt_[tg]sigqueueinfo was deprecated, so now we issue a warning on the first attempt of doing it. We use WARN_ON_ONCE, which is not informative and, what is worse, taints the kernel, making the trinity syscall fuzzer complain false-positively from time to time.

Quoted 7 lines

It does not look like we need this warning at all, because the behaviour changed quite a long time ago (2.6.39), and if an application relies on the old API, it gets EPERM anyway and can issue a warning by itself.

So let us zap the warning in kernel.