KMSAN: kernel-infoleak in tty_compat_ioctl

[syzbot]

Hello,

syzbot found the following crash on:

HEAD commit: 8bbbc5cf kmsan: don't compile memmove
git tree: https://github.com/google/kmsan.git master
console output: https://syzkaller.appspot.com/x/log.txt?x=1280ea7ee00000
kernel config: https://syzkaller.appspot.com/x/.config?x=cd0e9a6b0e555cc3
dashboard link: https://syzkaller.appspot.com/bug?extid=8da9175e28eadcb203ce
compiler: clang version 10.0.0 (https://github.com/llvm/llvm-project/ c2443155a0fb245c8f17f2c1c72b6ea391e86e81)
userspace arch: i386
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=1722a3d9e00000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1071e265e00000

IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+8da917...@syzkaller.appspotmail.com

=====================================================
BUG: KMSAN: kernel-infoleak in kmsan_copy_to_user+0x81/0x90 mm/kmsan/kmsan_hooks.c:253
CPU: 1 PID: 11476 Comm: syz-executor814 Not tainted 5.6.0-rc2-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x1c9/0x220 lib/dump_stack.c:118
kmsan_report+0xf7/0x1e0 mm/kmsan/kmsan_report.c:118
kmsan_internal_check_memory+0x238/0x3d0 mm/kmsan/kmsan.c:423
kmsan_copy_to_user+0x81/0x90 mm/kmsan/kmsan_hooks.c:253
_copy_to_user+0x15a/0x1f0 lib/usercopy.c:33
copy_to_user include/linux/uaccess.h:174 [inline]
compat_tty_tiocgserial drivers/tty/tty_io.c:2748 [inline]
tty_compat_ioctl+0x1482/0x1850 drivers/tty/tty_io.c:2846
__do_compat_sys_ioctl fs/ioctl.c:857 [inline]
__se_compat_sys_ioctl+0x57c/0xed0 fs/ioctl.c:808
__ia32_compat_sys_ioctl+0xd9/0x110 fs/ioctl.c:808
do_syscall_32_irqs_on arch/x86/entry/common.c:339 [inline]
do_fast_syscall_32+0x3c7/0x6e0 arch/x86/entry/common.c:410
entry_SYSENTER_compat+0x68/0x77 arch/x86/entry/entry_64_compat.S:139
RIP: 0023:0xf7ff7d99
Code: 90 e8 0b 00 00 00 f3 90 0f ae e8 eb f9 8d 74 26 00 89 3c 24 c3 90 90 90 90 90 90 90 90 90 90 90 90 51 52 55 89 e5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 eb 0d 90 90 90 90 90 90 90 90 90 90 90 90
RSP: 002b:00000000ff97b20c EFLAGS: 00000213 ORIG_RAX: 0000000000000036
RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 000000000000541e
RDX: 0000000020000300 RSI: 00000000080ea078 RDI: 00000000ff97b260
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000

Local variable ----v32.i105@tty_compat_ioctl created at:
compat_tty_tiocgserial drivers/tty/tty_io.c:2735 [inline]
tty_compat_ioctl+0xf12/0x1850 drivers/tty/tty_io.c:2846
compat_tty_tiocgserial drivers/tty/tty_io.c:2735 [inline]
tty_compat_ioctl+0xf12/0x1850 drivers/tty/tty_io.c:2846

Bytes 50-51 of 60 are uninitialized
Memory access of size 60 starts at ffffb50b0158fce0
Data copied to user address 0000000020000300
=====================================================


---
This bug is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
syzbot can test patches for this bug, for details see:
https://goo.gl/tpsmEJ#testing-patches

[Eric Biggers]

From: Eric Biggers <ebig...@google.com>

Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start
copying a whole struct to userspace rather than individual fields, but
failed to initialize all padding and fields -- namely the hole after the
'iomem_reg_shift' field, and the 'reserved' field.

Fix this by initializing the struct to zero.

Reported-by: syzbot+8da917...@syzkaller.appspotmail.com
Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
Cc: <sta...@vger.kernel.org> # v4.20+
Signed-off-by: Eric Biggers <ebig...@google.com>
---
drivers/tty/tty_io.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 1fcf7ad83dfa0..d24c250312edf 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -2731,6 +2731,7 @@ static int compat_tty_tiocgserial(struct tty_struct *tty,
struct serial_struct v;
int err;
memset(&v, 0, sizeof(struct serial_struct));
+ memset(&v32, 0, sizeof(struct serial_struct32));

if (!tty->ops->set_serial)
return -ENOTTY;
--
2.25.1

[Jiri Slaby]

sizeof(v32) please.

thanks,
--
js
suse labs

[Eric Biggers]

Sure, I was trying to be consistent with the other line, but I'll just change
them both to sizeof.

- Eric

[Eric Biggers]

Fix two bugs in the compat implementation of TIOCGSERIAL, both of which
were introduced by the compat ioctl refactoring in Linux v4.20.

Eric Biggers (2):
tty: fix compat TIOCGSERIAL leaking uninitialized memory
tty: fix compat TIOCGSERIAL checking wrong function ptr

drivers/tty/tty_io.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

--
2.25.0.265.gbab2e86ba0-goog

[Eric Biggers]

From: Eric Biggers <ebig...@google.com>

Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start

copying a whole 'serial_struct32' to userspace rather than individual

fields, but failed to initialize all padding and fields -- namely the
hole after the 'iomem_reg_shift' field, and the 'reserved' field.

Fix this by initializing the struct to zero.

[v2: use sizeof, and convert the adjacent line for consistency.]

Reported-by: syzbot+8da917...@syzkaller.appspotmail.com
Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
Cc: <sta...@vger.kernel.org> # v4.20+
Signed-off-by: Eric Biggers <ebig...@google.com>
---

drivers/tty/tty_io.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 1fcf7ad83dfa0a..db4a13bc855ed6 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -2730,7 +2730,9 @@ static int compat_tty_tiocgserial(struct tty_struct *tty,
struct serial_struct32 v32;

struct serial_struct v;
int err;

- memset(&v, 0, sizeof(struct serial_struct));
+
+ memset(&v, 0, sizeof(v));
+ memset(&v32, 0, sizeof(v32));

if (!tty->ops->set_serial)
return -ENOTTY;
--

2.25.0.265.gbab2e86ba0-goog

[Eric Biggers]

From: Eric Biggers <ebig...@google.com>

Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start

checking for the presence of the ->set_serial function pointer rather
than ->get_serial. This appears to be a copy-and-paste error, since
->get_serial is the function pointer that is called as well as the
pointer that is checked by the non-compat version of TIOCGSERIAL.

Fix this by checking the correct function pointer.

Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
Cc: <sta...@vger.kernel.org> # v4.20+
Signed-off-by: Eric Biggers <ebig...@google.com>
---

drivers/tty/tty_io.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index db4a13bc855ed6..5a6f36b391d95d 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -2734,7 +2734,7 @@ static int compat_tty_tiocgserial(struct tty_struct *tty,
memset(&v, 0, sizeof(v));
memset(&v32, 0, sizeof(v32));

- if (!tty->ops->set_serial)
+ if (!tty->ops->get_serial)
return -ENOTTY;
err = tty->ops->get_serial(tty, &v);
if (!err) {
--
2.25.0.265.gbab2e86ba0-goog

[Jiri Slaby]

On 24. 02. 20, 19:20, Eric Biggers wrote:
> From: Eric Biggers <ebig...@google.com>
>
> Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
> tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start
> checking for the presence of the ->set_serial function pointer rather
> than ->get_serial. This appears to be a copy-and-paste error, since
> ->get_serial is the function pointer that is called as well as the
> pointer that is checked by the non-compat version of TIOCGSERIAL.
>
> Fix this by checking the correct function pointer.
>
> Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
> Cc: <sta...@vger.kernel.org> # v4.20+
> Signed-off-by: Eric Biggers <ebig...@google.com>

Acked-by: Jiri Slaby <jsl...@suse.cz>

> ---
> drivers/tty/tty_io.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
> index db4a13bc855ed6..5a6f36b391d95d 100644
> --- a/drivers/tty/tty_io.c
> +++ b/drivers/tty/tty_io.c
> @@ -2734,7 +2734,7 @@ static int compat_tty_tiocgserial(struct tty_struct *tty,
> memset(&v, 0, sizeof(v));
> memset(&v32, 0, sizeof(v32));
>
> - if (!tty->ops->set_serial)
> + if (!tty->ops->get_serial)
> return -ENOTTY;
> err = tty->ops->get_serial(tty, &v);
> if (!err) {
>


--

js
suse labs

[Jiri Slaby]

On 24. 02. 20, 19:20, Eric Biggers wrote:

> From: Eric Biggers <ebig...@google.com>
>
> Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
> tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start
> copying a whole 'serial_struct32' to userspace rather than individual
> fields, but failed to initialize all padding and fields -- namely the
> hole after the 'iomem_reg_shift' field, and the 'reserved' field.
>
> Fix this by initializing the struct to zero.
>
> [v2: use sizeof, and convert the adjacent line for consistency.]
>
> Reported-by: syzbot+8da917...@syzkaller.appspotmail.com
> Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
> Cc: <sta...@vger.kernel.org> # v4.20+
> Signed-off-by: Eric Biggers <ebig...@google.com>

Acked-by: Jiri Slaby <jsl...@suse.cz>

> ---
> drivers/tty/tty_io.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
> index 1fcf7ad83dfa0a..db4a13bc855ed6 100644
> --- a/drivers/tty/tty_io.c
> +++ b/drivers/tty/tty_io.c
> @@ -2730,7 +2730,9 @@ static int compat_tty_tiocgserial(struct tty_struct *tty,
> struct serial_struct32 v32;
> struct serial_struct v;
> int err;
> - memset(&v, 0, sizeof(struct serial_struct));
> +
> + memset(&v, 0, sizeof(v));
> + memset(&v32, 0, sizeof(v32));
>
> if (!tty->ops->set_serial)
> return -ENOTTY;
>


--

js
suse labs

[Eric Biggers]

On Tue, Feb 25, 2020 at 08:30:35AM +0100, Jiri Slaby wrote:
> On 24. 02. 20, 19:20, Eric Biggers wrote:
> > From: Eric Biggers <ebig...@google.com>
> >
> > Commit 77654350306a ("take compat TIOC[SG]SERIAL treatment into
> > tty_compat_ioctl()") changed the compat version of TIOCGSERIAL to start
> > copying a whole 'serial_struct32' to userspace rather than individual
> > fields, but failed to initialize all padding and fields -- namely the
> > hole after the 'iomem_reg_shift' field, and the 'reserved' field.
> >
> > Fix this by initializing the struct to zero.
> >
> > [v2: use sizeof, and convert the adjacent line for consistency.]
> >
> > Reported-by: syzbot+8da917...@syzkaller.appspotmail.com
> > Fixes: 77654350306a ("take compat TIOC[SG]SERIAL treatment into tty_compat_ioctl()")
> > Cc: <sta...@vger.kernel.org> # v4.20+
> > Signed-off-by: Eric Biggers <ebig...@google.com>
>
> Acked-by: Jiri Slaby <jsl...@suse.cz>
>

Thanks. Greg, are you planning to take these patches?

- Eric

[Greg Kroah-Hartman]

Yes, please give me a chance to catch up on things :)

[Greg Kroah-Hartman]

On Mon, Mar 02, 2020 at 01:24:25PM -0800, Eric Biggers wrote:

Yes, sorry, they were not cc: linux-serial and fell through my initial
filters, to go into my generic "to-review" bucket. Will take them
now...

greg k-h

[Eric Biggers]

If people are supposed to send tty patches to linux-serial, then you need to add
it to MAINTAINERS:

$ ./scripts/get_maintainer.pl drivers/tty/
Greg Kroah-Hartman <gre...@linuxfoundation.org> (supporter:TTY LAYER)
Jiri Slaby <jsl...@suse.com> (supporter:TTY LAYER)
linux-...@vger.kernel.org (open list)