Re: [PATCH net] net/sched: fix NULL dereference in the error path of tcf_bpf_init()

[Lucas Bates]

On Thu, Apr 5, 2018 at 7:19 PM, Davide Caratti  wrote:
> when tcf_bpf_init_from_ops() fails (e.g. because of program having invalid
> number of instructions), tcf_bpf_cfg_cleanup() calls bpf_prog_put(NULL) or
> bpf_prog_destroy(NULL). Unless CONFIG_BPF_SYSCALL is unset, this causes
> the following error:
>
>  BUG: unable to handle kernel NULL pointer dereference at 0020
>  PGD 80007345a067 P4D 80007345a067 PUD 340e1067 PMD 0
>  Oops:  [#1] SMP PTI
>  Modules linked in: act_bpf(E) ip6table_filter ip6_tables iptable_filter 
> binfmt_misc ext4 mbcache jbd2 crct10dif_pclmul crc32_pclmul 
> ghash_clmulni_intel snd_hda_codec_generic pcbc snd_hda_intel snd_hda_codec 
> snd_hda_core snd_hwdep snd_seq snd_seq_device snd_pcm aesni_intel crypto_simd 
> glue_helper cryptd joydev snd_timer snd virtio_balloon pcspkr soundcore 
> i2c_piix4 nfsd auth_rpcgss nfs_acl lockd grace sunrpc ip_tables xfs libcrc32c 
> ata_generic pata_acpi qxl drm_kms_helper syscopyarea sysfillrect sysimgblt 
> fb_sys_fops ttm virtio_blk drm virtio_net virtio_console i2c_core 
> crc32c_intel serio_raw virtio_pci ata_piix libata virtio_ring floppy virtio 
> dm_mirror dm_region_hash dm_log dm_mod [last unloaded: act_bpf]
>  CPU: 3 PID: 5654 Comm: tc Tainted: GE4.16.0.bpf_test+ #408
>  Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011
>  RIP: 0010:__bpf_prog_put+0xc/0xc0
>  RSP: 0018:9594003ef728 EFLAGS: 00010202
>  RAX:  RBX: 9594003ef758 RCX: 0024
>  RDX:  RSI: 0001 RDI: 
>  RBP:  R08: 0001 R09: 0044
>  R10: 0220 R11: 8a7ab9f17131 R12: 
>  R13: 8a7ab7c3c8e0 R14: 0001 R15: 8a7ab88f1054
>  FS:  7fcb2f17c740() GS:8a7abfd8() knlGS:
>  CS:  0010 DS:  ES:  CR0: 80050033
>  CR2: 0020 CR3: 7c888006 CR4: 001606e0
>  Call Trace:
>   tcf_bpf_cfg_cleanup+0x2f/0x40 [act_bpf]
>   tcf_bpf_cleanup+0x4c/0x70 [act_bpf]
>   __tcf_idr_release+0x79/0x140
>   tcf_bpf_init+0x125/0x330 [act_bpf]
>   tcf_action_init_1+0x2cc/0x430
>   ? get_page_from_freelist+0x3f0/0x11b0
>   tcf_action_init+0xd3/0x1b0
>   tc_ctl_action+0x18b/0x240
>   rtnetlink_rcv_msg+0x29c/0x310
>   ? _cond_resched+0x15/0x30
>   ? __kmalloc_node_track_caller+0x1b9/0x270
>   ? rtnl_calcit.isra.29+0x100/0x100
>   netlink_rcv_skb+0xd2/0x110
>   netlink_unicast+0x17c/0x230
>   netlink_sendmsg+0x2cd/0x3c0
>   sock_sendmsg+0x30/0x40
>   ___sys_sendmsg+0x27a/0x290
>   ? mem_cgroup_commit_charge+0x80/0x130
>   ? page_add_new_anon_rmap+0x73/0xc0
>   ? do_anonymous_page+0x2a2/0x560
>   ? __handle_mm_fault+0xc75/0xe20
>   __sys_sendmsg+0x58/0xa0
>   do_syscall_64+0x6e/0x1a0
>   entry_SYSCALL_64_after_hwframe+0x3d/0xa2
>  RIP: 0033:0x7fcb2e58eba0
>  RSP: 002b:7ffc93c496c8 EFLAGS: 0246 ORIG_RAX: 002e
>  RAX: ffda RBX: 7ffc93c497f0 RCX: 7fcb2e58eba0
>  RDX:  RSI: 7ffc93c49740 RDI: 0003
>  RBP: 5ac6a646 R08: 0002 R09: 
>  R10: 7ffc93c49120 R11: 0246 R12: 
>  R13: 7ffc93c49804 R14: 0001 R15: 0066afa0
>  Code: 5f 00 48 8b 43 20 48 c7 c7 70 2f 7c b8 c7 40 10 00 00 00 00 5b e9 a5 
> 8b 61 00 0f 1f 44 00 00 0f 1f 44 00 00 41 54 55 48 89 fd 53 <48> 8b 47 20 f0 
> ff 08 74 05 5b 5d 41 5c c3 41 89 f4 0f 1f 44 00
>  RIP: __bpf_prog_put+0xc/0xc0 RSP: 9594003ef728
>  CR2: 0020
>
> Fix it in tcf_bpf_cfg_cleanup(), ensuring that bpf_prog_{put,destroy}(f)
> is called only when f is not NULL.
>
> Fixes: bbc09e7842a5 ("net/sched: fix idr leak on the error path of 
> tcf_bpf_init()")
> Reported-by: Lucas Bates 
> Signed-off-by: Davide Caratti 

That does the trick, thanks Davide.

Tested-by: Lucas Bates 

[PATCH net] net/sched: fix NULL dereference in the error path of tcf_bpf_init()

[Davide Caratti]

when tcf_bpf_init_from_ops() fails (e.g. because of program having invalid
number of instructions), tcf_bpf_cfg_cleanup() calls bpf_prog_put(NULL) or
bpf_prog_destroy(NULL). Unless CONFIG_BPF_SYSCALL is unset, this causes
the following error:

 BUG: unable to handle kernel NULL pointer dereference at 0020
 PGD 80007345a067 P4D 80007345a067 PUD 340e1067 PMD 0
 Oops:  [#1] SMP PTI
 Modules linked in: act_bpf(E) ip6table_filter ip6_tables iptable_filter 
binfmt_misc ext4 mbcache jbd2 crct10dif_pclmul crc32_pclmul ghash_clmulni_intel 
snd_hda_codec_generic pcbc snd_hda_intel snd_hda_codec snd_hda_core snd_hwdep 
snd_seq snd_seq_device snd_pcm aesni_intel crypto_simd glue_helper cryptd 
joydev snd_timer snd virtio_balloon pcspkr soundcore i2c_piix4 nfsd auth_rpcgss 
nfs_acl lockd grace sunrpc ip_tables xfs libcrc32c ata_generic pata_acpi qxl 
drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm virtio_blk drm 
virtio_net virtio_console i2c_core crc32c_intel serio_raw virtio_pci ata_piix 
libata virtio_ring floppy virtio dm_mirror dm_region_hash dm_log dm_mod [last 
unloaded: act_bpf]
 CPU: 3 PID: 5654 Comm: tc Tainted: GE4.16.0.bpf_test+ #408
 Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011
 RIP: 0010:__bpf_prog_put+0xc/0xc0
 RSP: 0018:9594003ef728 EFLAGS: 00010202
 RAX:  RBX: 9594003ef758 RCX: 0024
 RDX:  RSI: 0001 RDI: 
 RBP:  R08: 0001 R09: 0044
 R10: 0220 R11: 8a7ab9f17131 R12: 
 R13: 8a7ab7c3c8e0 R14: 0001 R15: 8a7ab88f1054
 FS:  7fcb2f17c740() GS:8a7abfd8() knlGS:
 CS:  0010 DS:  ES:  CR0: 80050033
 CR2: 0020 CR3: 7c888006 CR4: 001606e0
 Call Trace:
  tcf_bpf_cfg_cleanup+0x2f/0x40 [act_bpf]
  tcf_bpf_cleanup+0x4c/0x70 [act_bpf]
  __tcf_idr_release+0x79/0x140
  tcf_bpf_init+0x125/0x330 [act_bpf]
  tcf_action_init_1+0x2cc/0x430
  ? get_page_from_freelist+0x3f0/0x11b0
  tcf_action_init+0xd3/0x1b0
  tc_ctl_action+0x18b/0x240
  rtnetlink_rcv_msg+0x29c/0x310
  ? _cond_resched+0x15/0x30
  ? __kmalloc_node_track_caller+0x1b9/0x270
  ? rtnl_calcit.isra.29+0x100/0x100
  netlink_rcv_skb+0xd2/0x110
  netlink_unicast+0x17c/0x230
  netlink_sendmsg+0x2cd/0x3c0
  sock_sendmsg+0x30/0x40
  ___sys_sendmsg+0x27a/0x290
  ? mem_cgroup_commit_charge+0x80/0x130
  ? page_add_new_anon_rmap+0x73/0xc0
  ? do_anonymous_page+0x2a2/0x560
  ? __handle_mm_fault+0xc75/0xe20
  __sys_sendmsg+0x58/0xa0
  do_syscall_64+0x6e/0x1a0
  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
 RIP: 0033:0x7fcb2e58eba0
 RSP: 002b:7ffc93c496c8 EFLAGS: 0246 ORIG_RAX: 002e
 RAX: ffda RBX: 7ffc93c497f0 RCX: 7fcb2e58eba0
 RDX:  RSI: 7ffc93c49740 RDI: 0003
 RBP: 5ac6a646 R08: 0002 R09: 
 R10: 7ffc93c49120 R11: 0246 R12: 
 R13: 7ffc93c49804 R14: 0001 R15: 0066afa0
 Code: 5f 00 48 8b 43 20 48 c7 c7 70 2f 7c b8 c7 40 10 00 00 00 00 5b e9 a5 8b 
61 00 0f 1f 44 00 00 0f 1f 44 00 00 41 54 55 48 89 fd 53 <48> 8b 47 20 f0 ff 08 
74 05 5b 5d 41 5c c3 41 89 f4 0f 1f 44 00
 RIP: __bpf_prog_put+0xc/0xc0 RSP: 9594003ef728
 CR2: 0020

Fix it in tcf_bpf_cfg_cleanup(), ensuring that bpf_prog_{put,destroy}(f)
is called only when f is not NULL.

Fixes: bbc09e7842a5 ("net/sched: fix idr leak on the error path of 
tcf_bpf_init()")
Reported-by: Lucas Bates 
Signed-off-by: Davide Caratti 
---
 net/sched/act_bpf.c | 12 
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/net/sched/act_bpf.c b/net/sched/act_bpf.c
index 9092531d45d8..18089c02e557 100644
--- a/net/sched/act_bpf.c
+++ b/net/sched/act_bpf.c
@@ -248,10 +248,14 @@ static int tcf_bpf_init_from_efd(struct nlattr **tb, 
struct tcf_bpf_cfg *cfg)
 
 static void tcf_bpf_cfg_cleanup(const struct tcf_bpf_cfg *cfg)
 {
-   if (cfg->is_ebpf)
-   bpf_prog_put(cfg->filter);
-   else
-   bpf_prog_destroy(cfg->filter);
+   struct bpf_prog *filter = cfg->filter;
+
+   if (filter) {
+   if (cfg->is_ebpf)
+   bpf_prog_put(filter);
+   else
+   bpf_prog_destroy(filter);
+   }
 
kfree(cfg->bpf_ops);
kfree(cfg->bpf_name);
-- 
2.14.3