Re: WARNING: at drivers/tty/tty_ldisc.c:766 tty_ldisc_reinit+0x52/0x90() [3.0]

[Jiri Slaby <jirislaby@gmail.com>]

On 10/09/2011 08:59 AM, Dave Young wrote:
> On Fri, Sep 2, 2011 at 4:59 AM, Jiri Slaby <jslaby@suse.cz> wrote:
>> On 08/28/2011 03:15 PM, Dave Jones wrote:
>>>
>>> Just saw this on a box that was recieving an rsync while I ssh'd in.
>>> The ssh session hung.
>>
>> Yes, could you apply the attached debug patch and post a dmesg output?
> 
> I see this problem as well, the debug patch reports the refcount is 2
> 
> You can produce the WARN as below:
> in xterm run:
> 
> exec 0<>/dev/console 1<>/dev/console 2<>/dev/console

Thanks for the info. I suppose your console is something where also some
getty runs. Then the warning is perfectly OK. And with the hunk below
constifying timing, I guess your system dies after the 5s timeout of
tty_ldisc_wait_idle. (You don't need to test the patch, just confirm
there is a getty running on the console.)

So actually I'm not sure what we should do there to not die. Perhaps
wait infinitely? Because the user wants to steal a terminal?


--- a/drivers/tty/n_tty.c
+++ b/drivers/tty/n_tty.c
@@ -49,6 +49,7 @@
 #include <linux/file.h>
 #include <linux/uaccess.h>
 #include <linux/module.h>
+#include <linux/delay.h>

 #include <asm/system.h>

@@ -1575,6 +1576,7 @@ static int n_tty_open(struct tty_struct *tty)

        /* These are ugly. Currently a malloc failure here can panic */
        if (!tty->read_buf) {
+               msleep(100);
                tty->read_buf = kzalloc(N_TTY_BUF_SIZE, GFP_KERNEL);
                if (!tty->read_buf)
                        return -ENOMEM;
@@ -1815,6 +1817,7 @@ do_it_again:
                        /* FIXME: does n_tty_set_room need locking ? */
                        n_tty_set_room(tty);
                        timeout = schedule_timeout(timeout);
+                       msleep(20);
                        BUG_ON(!tty->read_buf);
                        continue;
                }


thanks,
-- 
js