LRW/XTS + Via Padlock Bug in 2.6.24-rc7?

Hello,

I have a Via Epia Sn1000 Board with a padlock-aes crypto accelerator, if I uses this with cryptsetup and LRW or XTS, I get a kernel panic and cryptsetup never ends.

XTS:
cryptsetup -c aes-xts-plain -s 256 luksFormat /dev/raid/test

general protection fault: 0000 [#1]
Modules linked in: lrw padlock_aes xts xt_TCPMSS xt_tcpmss iptable_mangle pppoe pppox ipt_MASQUERADE xt_tcpudp iptable_nat nf_nat nf_conntrack_ipv4 iptable_filter ip_tables x_tables af_packet ppp_generic slhc aes_i586 dm_crypt dm_mod

Pid: 5991, comm: kcryptd Not tainted (2.6.24-rc7 #4)
EIP: 0060:[<f88cf028>] EFLAGS: 00010296 CPU: 0
EIP is at aes_crypt_copy+0x28/0x40 [padlock_aes]
EAX: f6385ff0 EBX: f61ce450 ECX: 00000001 EDX: f61ce430
ESI: f6235dac EDI: f6385ff0 EBP: f6235e00 ESP: f6235dac
DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068
Process kcryptd (pid: 5991, ti=f6234000 task=f63fcab0 task.ti=f6234000)

Stack: 811cd642 f3a4d1ea d12d0cfa 3098ed1e 00000000 ffffff24 c02695d8 00000060
      f61ce450 f6385ff0 f6385ff0 f88cf096 f61ce430 f6385ff0 f6385ff0 
00000010
      f88c90fc f581f3e0 f6235e34 f6235ea8 00000010 f89b59df c29b34ff 
391d88f0
Call Trace:
[<c02695d8>] gf128mul_x_ble+0x68/0x90
[<f88cf096>] aes_decrypt+0x56/0x60 [padlock_aes]
[<f88c90fc>] crypt+0xdc/0x110 [xts]
[<f88cf040>] aes_decrypt+0x0/0x60 [padlock_aes]
[<f88c9222>] decrypt+0x42/0x50 [xts]
[<f88cf810>] aes_encrypt+0x0/0x60 [padlock_aes]
[<f88cf040>] aes_decrypt+0x0/0x60 [padlock_aes]
[<f886a7de>] crypt_convert_scatterlist+0x6e/0xe0 [dm_crypt]
[<f886a9ea>] crypt_convert+0x19a/0x1c0 [dm_crypt]
[<f886aa10>] kcryptd_do_crypt+0x0/0x260 [dm_crypt]
[<f886aa59>] kcryptd_do_crypt+0x49/0x260 [dm_crypt]
[<c01195c0>] update_curr+0x70/0x110
[<f886aa10>] kcryptd_do_crypt+0x0/0x260 [dm_crypt]
[<c012b116>] run_workqueue+0x66/0xe0
[<c03f5279>] schedule+0x149/0x270
[<c012b8f0>] worker_thread+0x0/0x100
[<c012b98d>] worker_thread+0x9d/0x100
[<c012e790>] autoremove_wake_function+0x0/0x50
[<c012b8f0>] worker_thread+0x0/0x100
[<c012e452>] kthread+0x42/0x70
[<c012e410>] kthread+0x0/0x70
[<c0104acf>] kernel_thread_helper+0x7/0x18
=======================
Code: <f3> 0f a7 c8 8b 5c 24 20 8b 74 24 24 8b 7c 24 28 83 c4 2c c3 8d 74
EIP: [<f88cf028>] aes_crypt_copy+0x28/0x40 [padlock_aes] SS:ESP 0068:f6235dac
---[ end trace 9ae4eb03d4e4c477 ]---

note: kcryptd[5991] exited with preempt_count 1

I'm not on the lklm, so please CC me.

I hope you have enough information.

Thank you and greetings
Torben Viets

Hello again,

I've tried a little bit more with the XTS and the LRW Cipher, XTS is unuseable it crashes with the kernel panic below. LRW doenst work in 2.6.23.12 with/without the patch from http://www.mail-archive.com/[email protected]/msg244470.html or with 2.6.24-rc7

My System ist Ubuntu 7.10 with cryptsetup 2:1.0.5-2ubuntu2

everytime I use cryptsetup -c aes-lrw-benbi -s 256 luksFormat /dev/raid/test

I get the following in dmesg:

eneral protection fault: 0000 [#1]

Modules linked in: xt_TCPMSS xt_tcpmss iptable_mangle ipt_MASQUERADE xt_tcpudp xt_mark xt_state iptable_nat nf_nat nf_conntrack_ipv4 iptable_filter ip_tables x_tables pppoe pppox af_packet ppp_generic slhc aes_i586
CPU:    0
EIP:    0060:[<c035b828>]    Not tainted VLI
EFLAGS: 00010292   (2.6.23.12 #7)

EIP is at aes_crypt_copy+0x28/0x40
eax: f7639ff0 ebx: f6c24050 ecx: 00000001 edx: f6c24030
esi: f7e89dc8 edi: f7639ff0 ebp: 00010000 esp: f7e89dc8
ds: 007b es: 007b fs: 0000 gs: 0000 ss: 0068
Process kcryptd/0 (pid: 937, ti=f7e88000 task=f7d85ab0 task.ti=f7e88000)

Stack: dd7f6714 1b8c0a97 19fd5394 4492154e f7e89ef8 c0262f42 0000007b f7e89ec8
      f6c24050 f7639ff0 f7639ff0 c035bda6 f6c24030 00000000 f7639ff0 
f7e89e28
      c02681e6 f7639ff0 f76c6020 f7e89e58 f7e89ec8 00000010 f7e89ea0 
00000010
Call Trace:
[<c0262f42>] blkcipher_walk_next+0x132/0x2f0
[<c035bda6>] aes_decrypt+0x56/0x60
[<c02681e6>] crypt+0xe6/0x220
[<c035bd50>] aes_decrypt+0x0/0x60
[<c0268352>] decrypt+0x32/0x40
[<c035bd50>] aes_decrypt+0x0/0x60
[<c0357d5e>] crypt_convert_scatterlist+0x6e/0xe0
[<c0271e43>] __make_request+0x1a3/0x480
[<c0357f08>] crypt_convert+0x138/0x150
[<c03580e0>] kcryptd_do_work+0x0/0x310
[<c0358333>] kcryptd_do_work+0x253/0x310
[<c0118647>] update_stats_wait_end+0xa7/0xe0
[<c03580e0>] kcryptd_do_work+0x0/0x310
[<c012a206>] run_workqueue+0x66/0xe0
[<c012a9e0>] worker_thread+0x0/0x100
[<c012aa7d>] worker_thread+0x9d/0x100
[<c012d5c0>] autoremove_wake_function+0x0/0x50
[<c012a9e0>] worker_thread+0x0/0x100
[<c012d282>] kthread+0x42/0x70
[<c012d240>] kthread+0x0/0x70
[<c0104acf>] kernel_thread_helper+0x7/0x18
=======================
Code: 90 90 90 83 ec 2c 89 5c 24 20 89 cb 89 74 24 24 89 c6 89 d0 89 7c 24 28 8b 54 24 30 89 e7 a5 a5 a5 a5 b9 01 00 00 00 89 e6 89 c7 <f3> 0f a7 c8 8b 5c 24 20 8b 74 24 24 8b 7c 24 28 83 c4 2c c3 8d
EIP: [<c035b828>] aes_crypt_copy+0x28/0x40 SS:ESP 0068:f7e89dc8

note: kcryptd/0[937] exited with preempt_count 1

I'm not on the list, please CC me.

Thanks and greetings
Torben Viets

Thanks for the report. It would appear that your CPU goes one step further than the other report and insists on having two blocks in the destination too.

Please let me know whether this patch fixes the problem.

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
--

index a337b69..ed3ae47 100644
@@ -433,7 +433,8 @@ static void aes_crypt_copy(const u8 *in, u8 *out, u32 *key, struct cword *cword)
 		__attribute__ ((__aligned__(PADLOCK_ALIGNMENT)));
 
 	memcpy(tmp, in, AES_BLOCK_SIZE);
-	padlock_xcrypt(tmp, out, key, cword);
+	padlock_xcrypt(tmp, tmp, key, cword);
+	memcpy(out, tmp, AES_BLOCK_SIZE);
 }
 
 static inline void aes_crypt(const u8 *in, u8 *out, u32 *key,