linux-kernel - [PATCH v2] x86/traps: Don't force in_interrupt() to return true in IST handlers

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 for Android: free password hash cracker in your pocket

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

Message-Id: <cdc215f94d118d691d73df35275022331156fb45.1464130360.git.luto@kernel.org>
Date:	Tue, 24 May 2016 15:54:04 -0700
From:	Andy Lutomirski <luto@...nel.org>
To:	x86@...nel.org
Cc:	linux-kernel@...r.kernel.org, Borislav Petkov <bp@...en8.de>,
	Andi Kleen <andi@...stfloor.org>,
	Oleg Nesterov <oleg@...hat.com>,
	Andy Lutomirski <luto@...nel.org>, stable@...r.kernel.org
Subject: [PATCH v2] x86/traps: Don't force in_interrupt() to return true in IST handlers

Forcing in_interrupt() to return true if we're not in a bona fide
interrupt confuses the softirq code.  This fixes warnings like:

NOHZ: local_softirq_pending 282

that can happen when running things like selftests/x86.

This will change perf's static percpu buffer usage in IST context.
I think this is okay, and it's changing the behavior to match
historical (pre-4.0) behavior.

Cc: stable@...r.kernel.org
Fixes: 959274753857 ("x86, traps: Track entry into and exit from IST context")
Signed-off-by: Andy Lutomirski <luto@...nel.org>
Message-Id: <ce1cd1d47b486c142d57a9a8189470a1c3809a9c.1464062136.git.luto@...nel.org>
---

This is intended for x86/urgent.

Changes from v1:
 - Get rid of the 3* crap (Linus)
 - Improve changelog (PeterZ)

arch/x86/kernel/traps.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 06cbe25861f1..87bd6b6bf5bd 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -95,6 +95,12 @@ static inline void cond_local_irq_disable(struct pt_regs *regs)
 		local_irq_disable();
 }
 
+/*
+ * In IST context, we explicitly disable preemption.  This serves two
+ * purposes: it makes it much less likely that we would accidentally
+ * schedule in IST context and it will force a warning if we somehow
+ * manage to schedule by accident.
+ */
 void ist_enter(struct pt_regs *regs)
 {
 	if (user_mode(regs)) {
@@ -109,13 +115,7 @@ void ist_enter(struct pt_regs *regs)
 		rcu_nmi_enter();
 	}
 
-	/*
-	 * We are atomic because we're on the IST stack; or we're on
-	 * x86_32, in which case we still shouldn't schedule; or we're
-	 * on x86_64 and entered from user mode, in which case we're
-	 * still atomic unless ist_begin_non_atomic is called.
-	 */
-	preempt_count_add(HARDIRQ_OFFSET);
+	preempt_disable();
 
 	/* This code is a bit fragile.  Test it. */
 	RCU_LOCKDEP_WARN(!rcu_is_watching(), "ist_enter didn't work");
@@ -123,7 +123,7 @@ void ist_enter(struct pt_regs *regs)
 
 void ist_exit(struct pt_regs *regs)
 {
-	preempt_count_sub(HARDIRQ_OFFSET);
+	preempt_enable_no_resched();
 
 	if (!user_mode(regs))
 		rcu_nmi_exit();
@@ -154,7 +154,7 @@ void ist_begin_non_atomic(struct pt_regs *regs)
 	BUG_ON((unsigned long)(current_top_of_stack() -
 			       current_stack_pointer()) >= THREAD_SIZE);
 
-	preempt_count_sub(HARDIRQ_OFFSET);
+	preempt_enable_no_resched();
 }
 
 /**
@@ -164,7 +164,7 @@ void ist_begin_non_atomic(struct pt_regs *regs)
  */
 void ist_end_non_atomic(void)
 {
-	preempt_count_add(HARDIRQ_OFFSET);
+	preempt_disable();
 }
 
 static nokprobe_inline int
-- 
2.5.5