2009-07-04 08:36:54

by Eugene Kapun

[permalink] [raw]
Subject: PROBLEM: tun/tap crashes if open() /dev/net/tun and then poll() it.

This small program causes kernel oops:

#include <fcntl.h>
#include <poll.h>

int main() {
int fd;
struct pollfd pfd;
fd = open("/dev/net/tun", O_RDWR);
pfd.fd = fd;
pfd.events = POLLIN | POLLOUT;
poll(&pfd, 1, 0);
}

Software versions:
Gnu C 4.4.0
Gnu make 3.81
binutils 2.19.51.20090622
util-linux 2.15.1-rc1
mount support
module-init-tools 3.8
e2fsprogs 1.41.5
reiserfsprogs 3.6.21
pcmciautils 014
Linux C Library 2.9
Dynamic linker (ldd) 2.9
Procps 3.2.8
Net-tools 1.60
Kbd 1.15
Sh-utils 7.4
wireless-tools 29
Modules Loaded tun ppp_async option usb_storage binfmt_misc
ppdev kqemu sbp2 lp parport snd_hda_codec_realtek joydev snd_hda_intel
snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm arc4 snd_seq_dummy
snd_seq_oss ecb snd_seq_midi snd_rawmidi ath5k snd_seq_midi_event
snd_seq snd_timer snd_seq_device pcmcia mac80211 nsc_ircc ath psmouse
snd sdhci_pci sdhci acer_wmi yenta_socket rsrc_nonstatic uvcvideo
tifm_7xx1 tifm_core serio_raw soundcore iTCO_wdt iTCO_vendor_support
irda led_class pcmcia_core videodev v4l1_compat v4l2_compat_ioctl32
cfg80211 pcspkr snd_page_alloc crc_ccitt ohci1394 ieee1394 tg3 usbhid
fbcon tileblit font bitblit softcursor i915 drm i2c_algo_bit video
output intel_agp

dmesg output:
[14610.772456] BUG: unable to handle kernel NULL pointer dereference at
0000000000000080
[14610.772464] IP: [<ffffffffa0455b63>] tun_chr_poll+0x33/0xf0 [tun]
[14610.772473] PGD 2885f067 PUD 2bd8f067 PMD 0
[14610.772477] Oops: 0000 [#4] SMP
[14610.772480] last sysfs file:
/sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:29/PNP0C09:00/PNP0C0A:00/power_supply/BAT0/voltage_now
[14610.772484] CPU 0
[14610.772486] Modules linked in: tun ppp_async option usb_storage
binfmt_misc ppdev kqemu sbp2 lp parport snd_hda_codec_realtek joydev
snd_hda_intel snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm arc4
snd_seq_dummy snd_seq_oss ecb snd_seq_midi snd_rawmidi ath5k
snd_seq_midi_event snd_seq snd_timer snd_seq_device pcmcia mac80211
nsc_ircc ath psmouse snd sdhci_pci sdhci acer_wmi yenta_socket
rsrc_nonstatic uvcvideo tifm_7xx1 tifm_core serio_raw soundcore iTCO_wdt
iTCO_vendor_support irda led_class pcmcia_core videodev v4l1_compat
v4l2_compat_ioctl32 cfg80211 pcspkr snd_page_alloc crc_ccitt ohci1394
ieee1394 tg3 usbhid fbcon tileblit font bitblit softcursor i915 drm
i2c_algo_bit video output intel_agp
[14610.772532] Pid: 7896, comm: tun-crash Tainted: G D
2.6.31-rc1-git10-generic-vanilla #1 Extensa 5220
[14610.772534] RIP: 0010:[<ffffffffa0455b63>] [<ffffffffa0455b63>]
tun_chr_poll+0x33/0xf0 [tun]
[14610.772539] RSP: 0018:ffff88002bdbdb08 EFLAGS: 00010246
[14610.772541] RAX: 0000000000000000 RBX: 0000000000000000 RCX:
ffff88002bdbdb88
[14610.772544] RDX: 0000000000000000 RSI: 0000000000000000 RDI:
ffff88007add0fe0
[14610.772546] RBP: ffff88002bdbdb28 R08: 00007feea2ddb2e0 R09:
00007feea2def160
[14610.772548] R10: 00007fffc0c0a150 R11: 0000000000000246 R12:
ffff88002bdf8000
[14610.772550] R13: 0000000000000000 R14: ffff88002bdf8000 R15:
0000000000000000
[14610.772553] FS: 00007feea2fdf6f0(0000) GS:ffff88000199d000(0000)
knlGS:0000000000000000
[14610.772555] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[14610.772557] CR2: 0000000000000080 CR3: 000000002bdf7000 CR4:
00000000000006b0
[14610.772560] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
0000000000000000
[14610.772562] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
0000000000000400
[14610.772565] Process tun-crash (pid: 7896, threadinfo
ffff88002bdbc000, task ffff880028838000)
[14610.772566] Stack:
[14610.772568] ffff88002bdbdf04 ffff88002bdbde0c ffff88002bdbde04
ffff88002bdf8000
[14610.772572] <0> ffff88002bdbdf38 ffffffff8111eef5 ffff88002bdbdb58
01ffffff810ccaf2
[14610.772575] <0> 00007fffc0c0a3d0 0000000000000000 ffff88002bdbdb88
00000001810e221d
[14610.772580] Call Trace:
[14610.772587] [<ffffffff8111eef5>] do_sys_poll+0x235/0x510
[14610.772591] [<ffffffff8111eaf0>] ? __pollwait+0x0/0xf0
[14610.772596] [<ffffffff81504b49>] ? _spin_lock+0x9/0x10
[14610.772598] [<ffffffff81504b49>] ? _spin_lock+0x9/0x10
[14610.772601] [<ffffffff811219de>] ? __d_lookup+0x12e/0x150
[14610.772605] [<ffffffffa0455862>] ? tun_chr_open+0x32/0xe0 [tun]
[14610.772609] [<ffffffffa0455862>] ? tun_chr_open+0x32/0xe0 [tun]
[14610.772612] [<ffffffff81504dcd>] ? unlock_kernel+0x2d/0x40
[14610.772617] [<ffffffff812e351b>] ? misc_open+0x1bb/0x200
[14610.772622] [<ffffffff811118da>] ? chrdev_open+0xfa/0x1b0
[14610.772626] [<ffffffff810cb849>] ? find_get_page+0x19/0xa0
[14610.772629] [<ffffffff810cd831>] ? filemap_fault+0xb1/0x450
[14610.772632] [<ffffffff810ccaf2>] ? unlock_page+0x22/0x30
[14610.772636] [<ffffffff810e4ce9>] ? __do_fault+0x419/0x4e0
[14610.772639] [<ffffffff810e8d12>] ? handle_mm_fault+0x1b2/0x950
[14610.772643] [<ffffffff811271bb>] ? mntput_no_expire+0x2b/0x100
[14610.772647] [<ffffffff81030389>] ? default_spin_lock_flags+0x9/0x10
[14610.772651] [<ffffffff81260f9a>] ? __up_read+0x9a/0xc0
[14610.772656] [<ffffffff81073199>] ? up_read+0x9/0x10
[14610.772659] [<ffffffff8111f3b7>] sys_poll+0x77/0x110
[14610.772663] [<ffffffff81011ec2>] system_call_fastpath+0x16/0x1b
[14610.772665] Code: 48 89 1c 24 4c 89 64 24 08 49 89 fc 4c 89 6c 24 10
4c 89 74 24 18 48 89 f3 48 8b bf a0 00 00 00 e8 73 f6 ff ff 48 85 db 49
89 c5 <4c> 8b b0 80 00 00 00 0f 85 98 00 00 00 49 8d 45 18 49 39 45 18
[14610.772694] RIP [<ffffffffa0455b63>] tun_chr_poll+0x33/0xf0 [tun]
[14610.772698] RSP <ffff88002bdbdb08>
[14610.772699] CR2: 0000000000000080
[14610.772702] ---[ end trace 9813be85d2509042 ]---


2009-07-05 22:11:24

by Mariusz Kozlowski

[permalink] [raw]
Subject: Re: PROBLEM: tun/tap crashes if open() /dev/net/tun and then poll() it.

On Sat, 04 Jul 2009 12:36:16 +0400
Eugene Kapun <[email protected]> wrote:

> This small program causes kernel oops:
>
> #include <fcntl.h>
> #include <poll.h>
>
> int main() {
> int fd;
> struct pollfd pfd;
> fd = open("/dev/net/tun", O_RDWR);
> pfd.fd = fd;
> pfd.events = POLLIN | POLLOUT;
> poll(&pfd, 1, 0);
> }

Hi,

Can you try this patch?

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index 11a0ba4..c599327 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -486,12 +486,14 @@ static unsigned int tun_chr_poll(struct file *file, poll_table * wait)
{
struct tun_file *tfile = file->private_data;
struct tun_struct *tun = __tun_get(tfile);
- struct sock *sk = tun->sk;
+ struct sock *sk;
unsigned int mask = 0;

if (!tun)
return POLLERR;

+ sk = tun->sk;
+
DBG(KERN_INFO "%s: tun_chr_poll\n", tun->dev->name);

poll_wait(file, &tun->socket.wait, wait);



> Software versions:
> Gnu C 4.4.0
> Gnu make 3.81
> binutils 2.19.51.20090622
> util-linux 2.15.1-rc1
> mount support
> module-init-tools 3.8
> e2fsprogs 1.41.5
> reiserfsprogs 3.6.21
> pcmciautils 014
> Linux C Library 2.9
> Dynamic linker (ldd) 2.9
> Procps 3.2.8
> Net-tools 1.60
> Kbd 1.15
> Sh-utils 7.4
> wireless-tools 29
> Modules Loaded tun ppp_async option usb_storage binfmt_misc
> ppdev kqemu sbp2 lp parport snd_hda_codec_realtek joydev snd_hda_intel
> snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm arc4 snd_seq_dummy
> snd_seq_oss ecb snd_seq_midi snd_rawmidi ath5k snd_seq_midi_event
> snd_seq snd_timer snd_seq_device pcmcia mac80211 nsc_ircc ath psmouse
> snd sdhci_pci sdhci acer_wmi yenta_socket rsrc_nonstatic uvcvideo
> tifm_7xx1 tifm_core serio_raw soundcore iTCO_wdt iTCO_vendor_support
> irda led_class pcmcia_core videodev v4l1_compat v4l2_compat_ioctl32
> cfg80211 pcspkr snd_page_alloc crc_ccitt ohci1394 ieee1394 tg3 usbhid
> fbcon tileblit font bitblit softcursor i915 drm i2c_algo_bit video
> output intel_agp
>
> dmesg output:
> [14610.772456] BUG: unable to handle kernel NULL pointer dereference at
> 0000000000000080
> [14610.772464] IP: [<ffffffffa0455b63>] tun_chr_poll+0x33/0xf0 [tun]
> [14610.772473] PGD 2885f067 PUD 2bd8f067 PMD 0
> [14610.772477] Oops: 0000 [#4] SMP
> [14610.772480] last sysfs file:
> /sys/devices/LNXSYSTM:00/device:00/PNP0A08:00/device:29/PNP0C09:00/PNP0C0A:00/power_supply/BAT0/voltage_now
> [14610.772484] CPU 0
> [14610.772486] Modules linked in: tun ppp_async option usb_storage
> binfmt_misc ppdev kqemu sbp2 lp parport snd_hda_codec_realtek joydev
> snd_hda_intel snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm arc4
> snd_seq_dummy snd_seq_oss ecb snd_seq_midi snd_rawmidi ath5k
> snd_seq_midi_event snd_seq snd_timer snd_seq_device pcmcia mac80211
> nsc_ircc ath psmouse snd sdhci_pci sdhci acer_wmi yenta_socket
> rsrc_nonstatic uvcvideo tifm_7xx1 tifm_core serio_raw soundcore iTCO_wdt
> iTCO_vendor_support irda led_class pcmcia_core videodev v4l1_compat
> v4l2_compat_ioctl32 cfg80211 pcspkr snd_page_alloc crc_ccitt ohci1394
> ieee1394 tg3 usbhid fbcon tileblit font bitblit softcursor i915 drm
> i2c_algo_bit video output intel_agp
> [14610.772532] Pid: 7896, comm: tun-crash Tainted: G D
> 2.6.31-rc1-git10-generic-vanilla #1 Extensa 5220
> [14610.772534] RIP: 0010:[<ffffffffa0455b63>] [<ffffffffa0455b63>]
> tun_chr_poll+0x33/0xf0 [tun]
> [14610.772539] RSP: 0018:ffff88002bdbdb08 EFLAGS: 00010246
> [14610.772541] RAX: 0000000000000000 RBX: 0000000000000000 RCX:
> ffff88002bdbdb88
> [14610.772544] RDX: 0000000000000000 RSI: 0000000000000000 RDI:
> ffff88007add0fe0
> [14610.772546] RBP: ffff88002bdbdb28 R08: 00007feea2ddb2e0 R09:
> 00007feea2def160
> [14610.772548] R10: 00007fffc0c0a150 R11: 0000000000000246 R12:
> ffff88002bdf8000
> [14610.772550] R13: 0000000000000000 R14: ffff88002bdf8000 R15:
> 0000000000000000
> [14610.772553] FS: 00007feea2fdf6f0(0000) GS:ffff88000199d000(0000)
> knlGS:0000000000000000
> [14610.772555] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [14610.772557] CR2: 0000000000000080 CR3: 000000002bdf7000 CR4:
> 00000000000006b0
> [14610.772560] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
> 0000000000000000
> [14610.772562] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
> 0000000000000400
> [14610.772565] Process tun-crash (pid: 7896, threadinfo
> ffff88002bdbc000, task ffff880028838000)
> [14610.772566] Stack:
> [14610.772568] ffff88002bdbdf04 ffff88002bdbde0c ffff88002bdbde04
> ffff88002bdf8000
> [14610.772572] <0> ffff88002bdbdf38 ffffffff8111eef5 ffff88002bdbdb58
> 01ffffff810ccaf2
> [14610.772575] <0> 00007fffc0c0a3d0 0000000000000000 ffff88002bdbdb88
> 00000001810e221d
> [14610.772580] Call Trace:
> [14610.772587] [<ffffffff8111eef5>] do_sys_poll+0x235/0x510
> [14610.772591] [<ffffffff8111eaf0>] ? __pollwait+0x0/0xf0
> [14610.772596] [<ffffffff81504b49>] ? _spin_lock+0x9/0x10
> [14610.772598] [<ffffffff81504b49>] ? _spin_lock+0x9/0x10
> [14610.772601] [<ffffffff811219de>] ? __d_lookup+0x12e/0x150
> [14610.772605] [<ffffffffa0455862>] ? tun_chr_open+0x32/0xe0 [tun]
> [14610.772609] [<ffffffffa0455862>] ? tun_chr_open+0x32/0xe0 [tun]
> [14610.772612] [<ffffffff81504dcd>] ? unlock_kernel+0x2d/0x40
> [14610.772617] [<ffffffff812e351b>] ? misc_open+0x1bb/0x200
> [14610.772622] [<ffffffff811118da>] ? chrdev_open+0xfa/0x1b0
> [14610.772626] [<ffffffff810cb849>] ? find_get_page+0x19/0xa0
> [14610.772629] [<ffffffff810cd831>] ? filemap_fault+0xb1/0x450
> [14610.772632] [<ffffffff810ccaf2>] ? unlock_page+0x22/0x30
> [14610.772636] [<ffffffff810e4ce9>] ? __do_fault+0x419/0x4e0
> [14610.772639] [<ffffffff810e8d12>] ? handle_mm_fault+0x1b2/0x950
> [14610.772643] [<ffffffff811271bb>] ? mntput_no_expire+0x2b/0x100
> [14610.772647] [<ffffffff81030389>] ? default_spin_lock_flags+0x9/0x10
> [14610.772651] [<ffffffff81260f9a>] ? __up_read+0x9a/0xc0
> [14610.772656] [<ffffffff81073199>] ? up_read+0x9/0x10
> [14610.772659] [<ffffffff8111f3b7>] sys_poll+0x77/0x110
> [14610.772663] [<ffffffff81011ec2>] system_call_fastpath+0x16/0x1b
> [14610.772665] Code: 48 89 1c 24 4c 89 64 24 08 49 89 fc 4c 89 6c 24 10
> 4c 89 74 24 18 48 89 f3 48 8b bf a0 00 00 00 e8 73 f6 ff ff 48 85 db 49
> 89 c5 <4c> 8b b0 80 00 00 00 0f 85 98 00 00 00 49 8d 45 18 49 39 45 18
> [14610.772694] RIP [<ffffffffa0455b63>] tun_chr_poll+0x33/0xf0 [tun]
> [14610.772698] RSP <ffff88002bdbdb08>
> [14610.772699] CR2: 0000000000000080
> [14610.772702] ---[ end trace 9813be85d2509042 ]---
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

2009-07-06 01:12:46

by Herbert Xu

[permalink] [raw]
Subject: Re: PROBLEM: tun/tap crashes if open() /dev/net/tun and then poll() it.

On Mon, Jul 06, 2009 at 12:11:14AM +0200, Mariusz Kozlowski wrote:
>
> Can you try this patch?
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index 11a0ba4..c599327 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -486,12 +486,14 @@ static unsigned int tun_chr_poll(struct file *file, poll_table * wait)
> {
> struct tun_file *tfile = file->private_data;
> struct tun_struct *tun = __tun_get(tfile);
> - struct sock *sk = tun->sk;
> + struct sock *sk;
> unsigned int mask = 0;
>
> if (!tun)
> return POLLERR;
>
> + sk = tun->sk;
> +
> DBG(KERN_INFO "%s: tun_chr_poll\n", tun->dev->name);
>
> poll_wait(file, &tun->socket.wait, wait);

Good catch. Can you please resend with a sign-off?

Thanks,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

2009-07-06 05:48:45

by Mariusz Kozlowski

[permalink] [raw]
Subject: Re: PROBLEM: tun/tap crashes if open() /dev/net/tun and then poll() it.

On Mon, 6 Jul 2009 09:12:30 +0800
Herbert Xu <[email protected]> wrote:

> On Mon, Jul 06, 2009 at 12:11:14AM +0200, Mariusz Kozlowski wrote:
> >
> > Can you try this patch?
...
> Good catch. Can you please resend with a sign-off?

Sure. Just wanted to wait for confirmation from Eugene.

Fix NULL pointer dereference in tun_chr_pool() introduced by
commit 33dccbb050bbe35b88ca8cf1228dcf3e4d4b3554 and triggered
by this code:

int fd;
struct pollfd pfd;
fd = open("/dev/net/tun", O_RDWR);
pfd.fd = fd;
pfd.events = POLLIN | POLLOUT;
poll(&pfd, 1, 0);

Reported-by: Eugene Kapun <[email protected]>
Signed-off-by: Mariusz Kozlowski <[email protected]>

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index a1b0697..bcbb25e 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -482,12 +482,14 @@ static unsigned int tun_chr_poll(struct file *file, poll_table * wait)
{
struct tun_file *tfile = file->private_data;
struct tun_struct *tun = __tun_get(tfile);
- struct sock *sk = tun->sk;
+ struct sock *sk;
unsigned int mask = 0;

if (!tun)
return POLLERR;

+ sk = tun->sk;
+
DBG(KERN_INFO "%s: tun_chr_poll\n", tun->dev->name);

poll_wait(file, &tfile->read_wait, wait);

2009-07-06 20:08:52

by David Miller

[permalink] [raw]
Subject: Re: PROBLEM: tun/tap crashes if open() /dev/net/tun and then poll() it.

From: Mariusz Kozlowski <[email protected]>
Date: Mon, 6 Jul 2009 07:48:35 +0200

> Fix NULL pointer dereference in tun_chr_pool() introduced by
> commit 33dccbb050bbe35b88ca8cf1228dcf3e4d4b3554 and triggered
> by this code:
>
> int fd;
> struct pollfd pfd;
> fd = open("/dev/net/tun", O_RDWR);
> pfd.fd = fd;
> pfd.events = POLLIN | POLLOUT;
> poll(&pfd, 1, 0);
>
> Reported-by: Eugene Kapun <[email protected]>
> Signed-off-by: Mariusz Kozlowski <[email protected]>

Applied, thanks!