linux-kernel - [repost PATCH] Fix SMP alternatives : use mutex instead of spinlock, text

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20080422214240.33e79a94@daedalus.pq.iki.fi>
Date:	Tue, 22 Apr 2008 21:42:40 +0300
From:	Pekka Paalanen <pq@....fi>
To:	Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>,
	Ingo Molnar <mingo@...e.hu>
Cc:	linux-kernel@...r.kernel.org, Steven Rostedt <rostedt@...dmis.org>
Subject: [repost PATCH] Fix SMP alternatives : use mutex instead of
 spinlock, text_poke is sleepable

>From 58390e93507669d860ef2d178b60ad447d5260c4 Mon Sep 17 00:00:00 2001
From: Pekka Paalanen <pq@....fi>
Date: Sat, 19 Apr 2008 23:22:11 +0300
Subject: [PATCH] x86: Fix SMP alternatives: use mutex instead of spinlock.

text_poke is sleepable.
The original fix by Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>.

Signed-off-by: Pekka Paalanen <pq@....fi>
---

> Mathieu,
> 
> thank you for the quick reply. Your patch did not apply on sched-devel/latest
> so I made the changes by hand, resulting patch here. The change was in the
> context lines, which made `patch' fail, e.g. __FUNCTION__ -> __func__.
> Now my enter_uniprocessor(), that disables all but the first cpu, works fine.
> 
> My next quest is, why attempting to re-enable the cpus makes the whole
> machine reboot after a short hang.

Ingo,

I see you applied the NX fix. This patch is the other one needed to make
CPU hotplug working. Could you apply this, unless Mathieu has something
against it?

After this one I can resubmit the mmiotrace CPU auto-disable patch
(modified a bit). Though it will be a couple of days before I can get
back to coding.


Thanks.

 arch/x86/kernel/alternative.c |   18 +++++++++---------
 1 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 5412fd7..7ce0939 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -1,6 +1,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
-#include <linux/spinlock.h>
+#include <linux/mutex.h>
 #include <linux/list.h>
 #include <linux/kprobes.h>
 #include <linux/mm.h>
@@ -279,7 +279,7 @@ struct smp_alt_module {
 	struct list_head next;
 };
 static LIST_HEAD(smp_alt_modules);
-static DEFINE_SPINLOCK(smp_alt);
+static DEFINE_MUTEX(smp_alt);
 static int smp_mode = 1;	/* protected by smp_alt */
 
 void alternatives_smp_module_add(struct module *mod, char *name,
@@ -312,12 +312,12 @@ void alternatives_smp_module_add(struct module *mod, char *name,
 		__func__, smp->locks, smp->locks_end,
 		smp->text, smp->text_end, smp->name);
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 	list_add_tail(&smp->next, &smp_alt_modules);
 	if (boot_cpu_has(X86_FEATURE_UP))
 		alternatives_smp_unlock(smp->locks, smp->locks_end,
 					smp->text, smp->text_end);
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 void alternatives_smp_module_del(struct module *mod)
@@ -327,17 +327,17 @@ void alternatives_smp_module_del(struct module *mod)
 	if (smp_alt_once || noreplace_smp)
 		return;
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 	list_for_each_entry(item, &smp_alt_modules, next) {
 		if (mod != item->mod)
 			continue;
 		list_del(&item->next);
-		spin_unlock(&smp_alt);
+		mutex_unlock(&smp_alt);
 		DPRINTK("%s: %s\n", __func__, item->name);
 		kfree(item);
 		return;
 	}
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 void alternatives_smp_switch(int smp)
@@ -359,7 +359,7 @@ void alternatives_smp_switch(int smp)
 		return;
 	BUG_ON(!smp && (num_online_cpus() > 1));
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 
 	/*
 	 * Avoid unnecessary switches because it forces JIT based VMs to
@@ -383,7 +383,7 @@ void alternatives_smp_switch(int smp)
 						mod->text, mod->text_end);
 	}
 	smp_mode = smp;
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 #endif
-- 
1.5.3.7

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/