[PATCH 1/1] TTY Fix NULL pointer error in tty_wakeup

[PATCH 1/1] TTY Fix NULL pointer error in tty_wakeup

[Frankie Lin]

When calling tty_wakeup and the parameter is NULL,the oops will happen.
The virtual address &tty->flags can not be handled in the kernel space,
while PC goto test_bit(TTY_DO_WRITE_WAKEUP, &tty->flags).

Tested on ARM Machine msm8625.

Signed-off-by: Frankie Lin <frankiefrankie.lin@gmail.com>
---
 drivers/tty/tty_io.c |    5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 05400ac..9c4fb2e 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -519,6 +519,11 @@ void tty_wakeup(struct tty_struct *tty)
 {
 	struct tty_ldisc *ld;
 
+	if (!tty) {
+		pr_err("tty is null in function: %s.\n", __func__);
+		return;
+	}
+
 	if (test_bit(TTY_DO_WRITE_WAKEUP, &tty->flags)) {
 		ld = tty_ldisc_ref(tty);
 		if (ld) {
-- 
1.7.9.5

Re: [PATCH 1/1] TTY Fix NULL pointer error in tty_wakeup

[Greg Kroah-Hartman]

On Sat, Mar 30, 2013 at 03:58:43PM +0800, Frankie Lin wrote:
> When calling tty_wakeup and the parameter is NULL,the oops will happen.

Who is calling that function with NULL?  Shouldn't that be fixed
instead?

thanks,

greg k-h

Re: [PATCH 1/1] TTY Fix NULL pointer error in tty_wakeup

[Greg Kroah-Hartman]

A: No.
Q: Should I include quotations after my reply?

http://daringfireball.net/2007/07/on_top


On Sun, Mar 31, 2013 at 04:09:05PM +0800, frankie lin wrote:
> Hi, greg, it's excited to get your reply.
> In our case, it is gs_start_io(struct gs_port *port) in the file u_serial.c who
> call the tty_wakeup. We are trying to find the root cause.  Something is wrong
> before calling tty_wakeup, nevertheless it is reasonable to check the parameter
> first before test_bit.

No, please find the root problem here in your driver.

> To some extent, this patch can improve the robustness of function tty_wakeup,
> isn't it?

Not at the expense of papering over other bugs, sorry.

greg k-h