[PATCH V2] tty: serial: omap: Increase max consoles and add check to prevent crash

From: Nishanth Menon
Date: Wed Oct 22 2014 - 08:47:38 EST


Increase the maximum number of consoles possible to 10 since DRA7 now
has the maximum number of consoles possible. without doing this, for
example, enabling DRA7 UART10 results in internal data structures and
console cannot match up and we endup with a crash as follows:

[ 1.903503] omap_uart 48424000.serial: [UART-1]: failure [serial_omap_probe]: -22
[ 1.911437] omap_uart: probe of 48424000.serial failed with error -22
[ 1.920379] Unable to handle kernel NULL pointer dereference at virtual address 00000004
[ 1.928894] pgd = c0004000
[ 1.931732] [00000004] *pgd=00000000
[ 1.935485] Internal error: Oops: 5 [#1] SMP ARM
[ 1.940338] Modules linked in:
[ 1.943542] CPU: 1 PID: 12 Comm: kworker/1:0 Tainted: G W 3.18.0-rc1-00001-g8821bc4-dirty #6
[ 1.953521] task: ed8a2d00 ti: ed8a4000 task.ti: ed8a4000
[ 1.959197] PC is at process_one_work+0x38/0x4a4
[ 1.964050] LR is at 0x0
[ 1.966705] pc : [<c00548e0>] lr : [<00000000>] psr: 40000093
[ 1.966705] sp : ed8a5ea8 ip : ed8a5ec8 fp : eeb9abc0
[ 1.978759] r10: ed8a4000 r9 : 00000008 r8 : ed842458
[ 1.984252] r7 : 00000000 r6 : eeb9abc0 r5 : ed842440 r4 : edbf26a8
[ 1.991119] r3 : 00000000 r2 : 00000000 r1 : 00000000 r0 : 00000000
[ 1.997985] Flags: nZcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
[ 2.005767] Control: 10c5387d Table: 8000406a DAC: 00000015
[ 2.011810] Process kworker/1:0 (pid: 12, stack limit = 0xed8a4248)
[ 2.018371] Stack: (0xed8a5ea8 to 0xed8a6000)
[ 2.022949] 5ea0: 00000001 00000000 60000093 c008346c 00000001 ed8a5ec8
[ 2.031555] 5ec0: c0054de4 eeb9abc0 ed8a4000 ed842458 00000008 ed8a4000 eeb9abc0 ed842440
[ 2.040161] 5ee0: eeb9abc0 eeb9abf0 ed8a4000 ed842458 00000008 ed8a4000 eeb9abc0 c0054ec4
[ 2.048767] 5f00: ed8a4000 eeb9ac4c a0000053 00000000 ed845bc0 ed842440 c0054d80 00000000
[ 2.057342] 5f20: 00000000 00000000 00000000 c005999c 00000001 00000000 c05eaa64 ed842440
[ 2.065948] 5f40: 00000000 00000000 dead4ead ffffffff ffffffff c097c680 00000000 00000000
[ 2.074554] 5f60: c078acd4 ed8a5f64 ed8a5f64 00000000 00000000 dead4ead ffffffff ffffffff
[ 2.083160] 5f80: c097c680 00000000 00000000 c078acd4 ed8a5f90 ed8a5f90 600000d3 ed845bc0
[ 2.091766] 5fa0: c00598d4 00000000 00000000 c000e668 00000000 00000000 00000000 00000000
[ 2.100341] 5fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[ 2.108947] 5fe0: 00000000 00000000 00000000 00000000 00000013 00000000 90005148 11010482
[ 2.117553] [<c00548e0>] (process_one_work) from [<c0054ec4>] (worker_thread+0x144/0x498)
[ 2.126159] [<c0054ec4>] (worker_thread) from [<c005999c>] (kthread+0xc8/0xe4)
[ 2.133758] [<c005999c>] (kthread) from [<c000e668>] (ret_from_fork+0x14/0x2c)
[ 2.141357] Code: e1a04001 e1a05000 e893000f e31e0004 (e5978004)
[ 2.147766] ---[ end trace 5798e2803311b69f ]---
<snip>

The final solution is to transition off to use 8250 driver and no
dependency on console structures and move away from omap-serial
driver, hence no major cleanups are done for this driver, so add a
simple check to prevent future crashes in a similar manner in case of
platform with additional ports.

Signed-off-by: Nishanth Menon <nm@xxxxxx>
---
drivers/tty/serial/omap-serial.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/tty/serial/omap-serial.c b/drivers/tty/serial/omap-serial.c
index 18c30ca..b0fe462 100644
--- a/drivers/tty/serial/omap-serial.c
+++ b/drivers/tty/serial/omap-serial.c
@@ -46,7 +46,7 @@

#include <dt-bindings/gpio/gpio.h>

-#define OMAP_MAX_HSUART_PORTS 6
+#define OMAP_MAX_HSUART_PORTS 10

#define UART_BUILD_REVISION(x, y) (((x) << 8) | (y))

@@ -1693,6 +1693,13 @@ static int serial_omap_probe(struct platform_device *pdev)
goto err_port_line;
}

+ if (up->port.line >= OMAP_MAX_HSUART_PORTS) {
+ dev_err(&pdev->dev, "uart ID %d > MAX %d.\n", up->port.line,
+ OMAP_MAX_HSUART_PORTS);
+ ret = -ENXIO;
+ goto err_port_line;
+ }
+
ret = serial_omap_probe_rs485(up, pdev->dev.of_node);
if (ret < 0)
goto err_rs485;
--
1.7.9.5

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