linux-kernel - [PATCH 3.16 053/202] uart: Fix crash in uart_write and uart_put

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: <lsq.1556377989.891094013@decadent.org.uk>
Date:   Sat, 27 Apr 2019 16:13:09 +0100
From:   Ben Hutchings <ben@...adent.org.uk>
To:     linux-kernel@...r.kernel.org, stable@...r.kernel.org
CC:     akpm@...ux-foundation.org, Denis Kirjanov <kda@...ux-powerpc.org>,
        "Greg Kroah-Hartman" <gregkh@...uxfoundation.org>,
        "Samir Virmani" <samir@...edur.com>,
        "Tycho Andersen" <tycho@...ho.ws>
Subject: [PATCH 3.16 053/202] uart: Fix crash in uart_write and uart_put_char

3.16.66-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Samir Virmani <samir@...edur.com>

commit aff9cf5955185d1f183227e46c5f8673fa483813 upstream.

We were experiencing a crash similar to the one reported as part of
commit:a5ba1d95e46e ("uart: fix race between uart_put_char() and
uart_shutdown()") in our testbed as well. We continue to observe the same
crash after integrating the commit a5ba1d95e46e ("uart: fix race between
uart_put_char() and uart_shutdown()")

On reviewing the change, the port lock should be taken prior to checking for
if (!circ->buf) in fn. __uart_put_char and other fns. that update the buffer
uart_state->xmit.

Traceback:

[11/27/2018 06:24:32.4870] Unable to handle kernel NULL pointer dereference
                           at virtual address 0000003b

[11/27/2018 06:24:32.4950] PC is at memcpy+0x48/0x180
[11/27/2018 06:24:32.4950] LR is at uart_write+0x74/0x120
[11/27/2018 06:24:32.4950] pc : [<ffffffc0002e6808>]
                           lr : [<ffffffc0003747cc>] pstate: 000001c5
[11/27/2018 06:24:32.4950] sp : ffffffc076433d30
[11/27/2018 06:24:32.4950] x29: ffffffc076433d30 x28: 0000000000000140
[11/27/2018 06:24:32.4950] x27: ffffffc0009b9d5e x26: ffffffc07ce36580
[11/27/2018 06:24:32.4950] x25: 0000000000000000 x24: 0000000000000140
[11/27/2018 06:24:32.4950] x23: ffffffc000891200 x22: ffffffc01fc34000
[11/27/2018 06:24:32.4950] x21: 0000000000000fff x20: 0000000000000076
[11/27/2018 06:24:32.4950] x19: 0000000000000076 x18: 0000000000000000
[11/27/2018 06:24:32.4950] x17: 000000000047cf08 x16: ffffffc000099e68
[11/27/2018 06:24:32.4950] x15: 0000000000000018 x14: 776d726966205948
[11/27/2018 06:24:32.4950] x13: 50203a6c6974755f x12: 74647075205d3333
[11/27/2018 06:24:32.4950] x11: 3a35323a36203831 x10: 30322f37322f3131
[11/27/2018 06:24:32.4950] x9 : 5b205d303638342e x8 : 746164206f742070
[11/27/2018 06:24:32.4950] x7 : 7520736920657261 x6 : 000000000000003b
[11/27/2018 06:24:32.4950] x5 : 000000000000817a x4 : 0000000000000008
[11/27/2018 06:24:32.4950] x3 : 2f37322f31312a5b x2 : 000000000000006e
[11/27/2018 06:24:32.4950] x1 : ffffffc0009b9cf0 x0 : 000000000000003b

[11/27/2018 06:24:32.4950] CPU2: stopping
[11/27/2018 06:24:32.4950] CPU: 2 PID: 0 Comm: swapper/2 Tainted: P      D    O    4.1.51 #3
[11/27/2018 06:24:32.4950] Hardware name: Broadcom-v8A (DT)
[11/27/2018 06:24:32.4950] Call trace:
[11/27/2018 06:24:32.4950] [<ffffffc0000883b8>] dump_backtrace+0x0/0x150
[11/27/2018 06:24:32.4950] [<ffffffc00008851c>] show_stack+0x14/0x20
[11/27/2018 06:24:32.4950] [<ffffffc0005ee810>] dump_stack+0x90/0xb0
[11/27/2018 06:24:32.4950] [<ffffffc00008e844>] handle_IPI+0x18c/0x1a0
[11/27/2018 06:24:32.4950] [<ffffffc000080c68>] gic_handle_irq+0x88/0x90

Fixes: a5ba1d95e46e ("uart: fix race between uart_put_char() and uart_shutdown()")
Signed-off-by: Samir Virmani <samir@...edur.com>
Acked-by: Tycho Andersen <tycho@...ho.ws>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
[bwh: Backported to 3.16:
 - Keep open-coding uart_port_lock()
 - Adjust context]
Signed-off-by: Ben Hutchings <ben@...adent.org.uk>
---
 drivers/tty/serial/serial_core.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

--- a/drivers/tty/serial/serial_core.c
+++ b/drivers/tty/serial/serial_core.c
@@ -480,10 +480,12 @@ static inline int __uart_put_char(struct
 	unsigned long flags;
 	int ret = 0;
 
-	if (!circ->buf)
+	spin_lock_irqsave(&port->lock, flags);
+	if (!circ->buf) {
+		spin_unlock_irqrestore(&port->lock, flags);
 		return 0;
+	}
 
-	spin_lock_irqsave(&port->lock, flags);
 	if (uart_circ_chars_free(circ) != 0) {
 		circ->buf[circ->head] = c;
 		circ->head = (circ->head + 1) & (UART_XMIT_SIZE - 1);
@@ -524,12 +526,14 @@ static int uart_write(struct tty_struct
 	}
 
 	port = state->uart_port;
+	spin_lock_irqsave(&port->lock, flags);
 	circ = &state->xmit;
 
-	if (!circ->buf)
+	if (!circ->buf) {
+		spin_unlock_irqrestore(&port->lock, flags);
 		return 0;
+	}
 
-	spin_lock_irqsave(&port->lock, flags);
 	while (1) {
 		c = CIRC_SPACE_TO_END(circ->head, circ->tail, UART_XMIT_SIZE);
 		if (count < c)