linux-kernel - Re: BUG: unable to handle kernel NULL pointer dereference in mem_serial

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: <CACT4Y+ZiOF0cwy4sDVw7T5wCpX-M9KarRj5S=rFY8zrHbrLFxA@mail.gmail.com>
Date:   Wed, 18 Dec 2019 07:53:48 +0100
From:   Dmitry Vyukov <dvyukov@...gle.com>
To:     Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
Cc:     Greg KH <gregkh@...uxfoundation.org>,
        syzbot <syzbot+f4f1e871965064ae689e@...kaller.appspotmail.com>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        asierra@...-inc.com, ext-kimmo.rautkoski@...sala.com,
        Jiri Slaby <jslaby@...e.com>,
        kai heng feng <kai.heng.feng@...onical.com>,
        LKML <linux-kernel@...r.kernel.org>,
        linux-serial <linux-serial@...r.kernel.org>,
        mika.westerberg@...ux.intel.com, o.barta89@...il.com,
        paulburton@...nel.org, sr@...x.de,
        syzkaller-bugs <syzkaller-bugs@...glegroups.com>,
        yegorslists@...glemail.com
Subject: Re: BUG: unable to handle kernel NULL pointer dereference in mem_serial_out

On Wed, Dec 18, 2019 at 1:56 AM Tetsuo Handa
<penguin-kernel@...ove.sakura.ne.jp> wrote:
>
> Hmm, this is a surprising bug. syzbot provided a C reproducer, but the definition
> of "struct serial_struct" used in that reproducer is wrong. As a result, syzbot was
> reporting crash caused by passing wrong arguments. ;-)

We are on it:
https://github.com/google/syzkaller/blob/master/sys/linux/dev_ptmx.txt.warn#L20-L25

> close_delay field used in the C reproducer is sizeof(unsigned int) bytes rather than
> sizeof(unsigned short) bytes, thus fields after close_delay field are incorrectly
> interpreted.
>
> ----------------------------------------
> #include <stdio.h>
> #include <sys/types.h>
> #include <sys/stat.h>
> #include <fcntl.h>
> #include <sys/ioctl.h>
> #include <linux/serial.h>
>
> struct bad_serial_struct {
>         int     type;
>         int     line;
>         unsigned int    port;
>         int     irq;
>         int     flags;
>         int     xmit_fifo_size;
>         int     custom_divisor;
>         int     baud_base;
>         unsigned int    close_delay; /* Correct type is "unsigned short". */
>         char    io_type;
>         char    reserved_char[1];
>         int     hub6;
>         unsigned short  closing_wait;
>         unsigned short  closing_wait2;
>         unsigned char   *iomem_base;
>         unsigned short  iomem_reg_shift;
>         unsigned int    port_high;
>         unsigned long   iomap_base;
> };
>
> int main(int argc, char *argv[])
> {
>         struct bad_serial_struct ss = { };
>         int fd = open("/dev/ttyS3", O_RDONLY);
>         ss.type = 0xa;
>         ss.line = 0x400000;
>         ss.port = 0x100;
>         ss.irq = 0;
>         ss.flags = 0x400000;
>         ss.xmit_fifo_size = 0;
>         ss.custom_divisor = 0;
>         ss.baud_base = 0x80000;
>         ss.close_delay = 0x200ff;
>         ss.io_type = 0;
>         ss.reserved_char[0] = 0x41;
>         ss.hub6 = 3;
>         ss.closing_wait = 0;
>         ss.closing_wait2 = 0x7c5;
>         ss.iomem_base = NULL;
>         ss.iomem_reg_shift = 0;
>         ss.port_high = 0;
>         ss.iomap_base = 0;
>         ioctl(fd, TIOCSSERIAL, &ss);
>         return 0;
> }
> ----------------------------------------