linux-kernel - [PATCH] sched: Work around undefined behavior in sched class checking

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-next>] [day] [month] [year] [list]

Date:   Tue,  4 May 2021 20:39:45 -0700
From:   Andi Kleen <ak@...ux.intel.com>
To:     peterz@...radead.org
Cc:     linux-kernel@...r.kernel.org, Andi Kleen <andi@...stfloor.org>
Subject: [PATCH] sched: Work around undefined behavior in sched class checking

From: Andi Kleen <andi@...stfloor.org>

The scheduler initialization code checks that the scheduling
classes are consecutive in memory by comparing the end
addresses with the next address.

Technically in ISO C comparing symbol addresseses outside different objects
is undefined. With LTO gcc 10 tries to exploits this and creates an
unconditional BUG_ON in the scheduler initialization, resulting
in a boot hang.

Use RELOC_HIDE to make this work. This hides the symbols from gcc,
so the optimizer won't make these assumption. I also split
the BUG_ONs in multiple.

Signed-off-by: Andi Kleen <andi@...stfloor.org>
---
 kernel/sched/core.c | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 98191218d891..272a654f5293 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -8086,12 +8086,20 @@ void __init sched_init(void)
 	unsigned long ptr = 0;
 	int i;
 
-	/* Make sure the linker didn't screw up */
-	BUG_ON(&idle_sched_class + 1 != &fair_sched_class ||
-	       &fair_sched_class + 1 != &rt_sched_class ||
-	       &rt_sched_class + 1   != &dl_sched_class);
+	/*
+	 * Make sure the linker didn't screw up.
+	 * We have to use RELOC_HIDE to prevent gcc from optimizing
+	 * them to true because they're technically undefined in ISO-C.
+	 */
+	BUG_ON(RELOC_HIDE(&idle_sched_class, 0) + 1 !=
+			RELOC_HIDE(&fair_sched_class, 0));
+	BUG_ON(RELOC_HIDE(&fair_sched_class, 0) + 1 !=
+			RELOC_HIDE(&rt_sched_class, 0));
+	BUG_ON(RELOC_HIDE(&rt_sched_class, 0) + 1 !=
+			RELOC_HIDE(&dl_sched_class, 0));
 #ifdef CONFIG_SMP
-	BUG_ON(&dl_sched_class + 1 != &stop_sched_class);
+	BUG_ON(RELOC_HIDE(&dl_sched_class, 0) + 1 !=
+			RELOC_HIDE(&stop_sched_class, 0));
 #endif
 
 	wait_bit_init();
-- 
2.25.4