Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Hongbo Yao <yaohongbo@huawei.com>
To:     <tglx@linutronix.de>, <edumazet@google.com>,
        <linux-kernel@vger.kernel.org>, <yaohongbo@huawei.com>
Subject: [RFC PATCH 2/2] hrtimer: Prevent overflow for relative refrences
Date:   Wed, 6 Mar 2019 21:13:26 +0800
Message-ID: <20190306131326.10275-3-yaohongbo@huawei.com>
In-Reply-To: <20190306131326.10275-1-yaohongbo@huawei.com>
References: <20190306131326.10275-1-yaohongbo@huawei.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 7BIT
Content-Type:   text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

I ran Syzkaller testsuite, and got the following call trace.

===============================================================
UBSAN: Undefined behaviour in ../kernel/time/hrtimer.c:514:11
signed integer overflow:
9223372036854775807 - -240652746 cannot be represented in type 'long
long int'
CPU: 1 PID: 9308 Comm: trinity-c5 Not tainted 4.19.25-dirty #4
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014
Call Trace:
 <IRQ>
 dump_stack+0xdd/0x12a
 ubsan_epilogue+0x10/0x65
 handle_overflow+0x1a8/0x212
 ? val_to_string.constprop.2+0x137/0x137
 ? print_lock_contention_bug+0x190/0x190
 ? __next_base+0x2b/0xc0
 ? __hrtimer_run_queues+0xca/0x830
 __ubsan_handle_sub_overflow+0x11/0x19
 __hrtimer_next_event_base+0x1a9/0x1b0
 __hrtimer_get_next_event+0x96/0x140
 hrtimer_interrupt+0x391/0x610
 smp_apic_timer_interrupt+0x137/0x3d0
 apic_timer_interrupt+0xf/0x20

===============================================================

Use ktime_sub_safe() which has the necessary sanity checks in place and
limits the result to the valid range.

Signed-off-by: Hongbo Yao <yaohongbo@huawei.com>
---
 kernel/time/hrtimer.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index cadc5bcbfc9e..2195b35af085 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -527,7 +527,8 @@ static ktime_t __hrtimer_next_event_base(struct hrtimer_cpu_base *cpu_base,
 
 			timer = container_of(next, struct hrtimer, node);
 		}
-		expires = ktime_sub(hrtimer_get_expires(timer), base->offset);
+		expires = ktime_sub_safe(hrtimer_get_expires(timer),
+					 base->offset);
 		if (expires < expires_next) {
 			expires_next = expires;
 
@@ -781,7 +782,8 @@ static void hrtimer_reprogram(struct hrtimer *timer, bool reprogram)
 {
 	struct hrtimer_cpu_base *cpu_base = this_cpu_ptr(&hrtimer_bases);
 	struct hrtimer_clock_base *base = timer->base;
-	ktime_t expires = ktime_sub(hrtimer_get_expires(timer), base->offset);
+	ktime_t expires = ktime_sub_safe(hrtimer_get_expires(timer),
+					 base->offset);
 
 	WARN_ON_ONCE(hrtimer_get_expires_tv64(timer) < 0);
 
-- 
2.20.1