linux-kernel - Re: [PATCH] drivers/tty: Don't hangup shared ttys

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: <CADMs+9ZWnz4zFqa69QoHfbOLc7Yxn2Cvo-NTNdAhw8-uPm9M6A@mail.gmail.com>
Date:	Tue, 11 Jun 2013 16:19:47 -0700
From:	Stéphane Marchesin <marcheu@...omium.org>
To:	Greg KH <gregkh@...uxfoundation.org>
Cc:	linux-kernel@...r.kernel.org, jslaby@...e.cz, olof@...om.net,
	linux-serial@...r.kernel.org
Subject: Re: [PATCH] drivers/tty: Don't hangup shared ttys

On Tue, Jun 11, 2013 at 4:15 PM, Greg KH <gregkh@...uxfoundation.org> wrote:
> On Tue, Jun 11, 2013 at 04:03:07PM -0700, Stéphane Marchesin wrote:
>> When quickly restarting X servers, we can run into a situation where
>> one X server quits while another one starts on the same tty. For a
>> while, two X servers share the tty, and when the old X server
>> eventually quits, the tty layer hangs up the tty, which among other
>> things stubs out the tty's ioctl functions. Later on, the new X
>> server (which shares the tty functions) tries to call some ioctls
>> on the tty and fails because they have been replaced with the hungup
>> versions. This in turn causes the new X server to abort.
>>
>> This patch checks the tty->count to make sure we're the last
>> consumer before hanging up a tty.
>>
>> Signed-off-by: Stéphane Marchesin <marcheu@...omium.org>
>> ---
>>  drivers/tty/tty_io.c | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
>> index 6464029..62a0f02 100644
>> --- a/drivers/tty/tty_io.c
>> +++ b/drivers/tty/tty_io.c
>> @@ -619,6 +619,9 @@ static void __tty_hangup(struct tty_struct *tty, int exit_session)
>>       if (!tty)
>>               return;
>>
>> +     /* Don't hangup if there are other users */
>> +     if (tty->count > 1)
>> +             return;
>
> What happens when you have a "real" tty that was hungup because it was
> disconnected physically from the system yet userspace had a tty open?
> You want those ttys to be hungup properly, right?  Doesn't this change
> break that?

My understanding was that they'd have a different tty_struct. Is that
not the case? If so how would you recommend fixing the problem I
described?

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