Re: [PATCH v4 0/9] crypto: caam - backlogging support

[Iuliana Prodan]

On 1/23/2020 9:43 PM, Corentin Labbe wrote:
> On Mon, Jan 20, 2020 at 02:23:59PM +0200, Iuliana Prodan wrote:
>> Integrate crypto_engine framework into CAAM, to make use of
>> the engine queue.
>> Added support for SKCIPHER, HASH, RSA and AEAD algorithms.
>> This is intended to be used for CAAM backlogging support.
>> The requests, with backlog flag (e.g. from dm-crypt) will be
>> listed into crypto-engine queue and processed by CAAM when free.
>>
>> While here, I've also made some refactorization.
>> Patches #1 - #4 include some refactorizations on caamalg, caamhash
>> and caampkc.
>> Patch #5 changes the return code of caam_jr_enqueue function
>> to -EINPROGRESS, in case of success, -ENOSPC in case the CAAM is
>> busy, -EIO if it cannot map the caller's descriptor.
>> Patches #6 - #9 integrate crypto_engine into CAAM, for
>> SKCIPHER/AEAD/RSA/HASH algorithms.
>>
>> ---
>> Changes since V3:
>> 	- update return on ahash_enqueue_req function from patch #9.
>>
>> Changes since V2:
>> 	- remove patch ("crypto: caam - refactor caam_jr_enqueue"),
>> 	  that added some structures not needed anymore;
>> 	- use _done_ callback function directly for skcipher and aead;
>> 	- handle resource leak in case of transfer request to crypto-engine;
>> 	- update commit messages.
>>
>> Changes since V1:
>> 	- remove helper function - akcipher_request_cast;
>> 	- remove any references to crypto_async_request,
>> 	  use specific request type;
>> 	- remove bypass crypto-engine queue, in case is empty;
>> 	- update some commit messages;
>> 	- remove unrelated changes, like whitespaces;
>> 	- squash some changes from patch #9 to patch #6;
>> 	- added Reviewed-by.
>> ---
>>
>> Iuliana Prodan (9):
>>    crypto: caam - refactor skcipher/aead/gcm/chachapoly {en,de}crypt
>>      functions
>>    crypto: caam - refactor ahash_done callbacks
>>    crypto: caam - refactor ahash_edesc_alloc
>>    crypto: caam - refactor RSA private key _done callbacks
>>    crypto: caam - change return code in caam_jr_enqueue function
>>    crypto: caam - support crypto_engine framework for SKCIPHER algorithms
>>    crypto: caam - add crypto_engine support for AEAD algorithms
>>    crypto: caam - add crypto_engine support for RSA algorithms
>>    crypto: caam - add crypto_engine support for HASH algorithms
>>
>>   drivers/crypto/caam/Kconfig    |   1 +
>>   drivers/crypto/caam/caamalg.c  | 416 ++++++++++++++++++-----------------------
>>   drivers/crypto/caam/caamhash.c | 341 ++++++++++++++++-----------------
>>   drivers/crypto/caam/caampkc.c  | 187 +++++++++++-------
>>   drivers/crypto/caam/caampkc.h  |  10 +
>>   drivers/crypto/caam/caamrng.c  |   4 +-
>>   drivers/crypto/caam/intern.h   |   2 +
>>   drivers/crypto/caam/jr.c       |  37 +++-
>>   drivers/crypto/caam/key_gen.c  |   2 +-
>>   9 files changed, 523 insertions(+), 477 deletions(-)
>>
> 
> Hello
> 
> I have tested this serie on a imx8mn-ddr4-evk and get the follwing crash:
> [   34.367787] Unable to handle kernel paging request at virtual address cccccccccccccccc
> [   34.375715] Mem abort info:
> [   34.378506]   ESR = 0x96000004
> [   34.381572]   EC = 0x25: DABT (current EL), IL = 32 bits
> [   34.386888]   SET = 0, FnV = 0
> [   34.389948]   EA = 0, S1PTW = 0
> [   34.393091] Data abort info:
> [   34.395978]   ISV = 0, ISS = 0x00000004
> [   34.399816]   CM = 0, WnR = 0
> [   34.402781] [cccccccccccccccc] address between user and kernel address ranges
> [   34.409922] Internal error: Oops: 96000004 [#1] PREEMPT SMP
> [   34.415492] Modules linked in: ctr des_generic caam_jr(+) caamhash_desc caamalg_desc rng_core authenc libdes fsl_imx8_ddr_perf snvs_pwrkey rtc_snvs caam error imx_cpufreq_dt
> [   34.430977] CPU: 2 PID: 267 Comm: cryptomgr_test Not tainted 5.5.0-rc7-next-20200122-00009-geed96b480af6 #39
> [   34.440799] Hardware name: NXP i.MX8MNano DDR4 EVK board (DT)
> [   34.446542] pstate: 60000005 (nZCv daif -PAN -UAO)
> [   34.451337] pc : __kmalloc+0xc4/0x290
> [   34.454997] lr : __kmalloc+0x48/0x290
> [   34.458655] sp : ffff800011b0b740
> [   34.461965] x29: ffff800011b0b740 x28: 0000000000000002
> [   34.467275] x27: 0000000000000080 x26: ffff800010c4b598
> [   34.472584] x25: 0000000000000cc0 x24: ffff00007b79b000
> [   34.477894] x23: 000000000000ac02 x22: ffff800008c5c254
> [   34.483203] x21: 0000000000000dc1 x20: cccccccccccccccc
> [   34.488512] x19: ffff00007d404200 x18: 0000000000000002
> [   34.493821] x17: 0000000008080040 x16: 0000000008182040
> [   34.499131] x15: 0000000002004800 x14: 0000000000000000
> [   34.504440] x13: 0000000000000000 x12: 0000000300000017
> [   34.509749] x11: fffffe0001ca2702 x10: 0000000000000000
> [   34.515058] x9 : 0000000000000000 x8 : ffff8000115e98c8
> [   34.520368] x7 : ffff800011776000 x6 : 0000000000000000
> [   34.525677] x5 : 000000000000ac02 x4 : 0000000000000011
> [   34.530986] x3 : 0000000000000002 x2 : 0000000000000000
> [   34.536295] x1 : ffff00007aaaf000 x0 : 0000000000000001
> [   34.541606] Call trace:
> [   34.544051]  __kmalloc+0xc4/0x290
> [   34.547377]  skcipher_edesc_alloc+0x154/0xab0 [caam_jr]
> [   34.552607]  skcipher_encrypt+0x58/0x118 [caam_jr]
> [   34.557397]  crypto_skcipher_encrypt+0x40/0x70
> [   34.561839]  test_skcipher_vec_cfg+0x254/0x768
> [   34.566279]  test_skcipher_vec+0xfc/0x148
> [   34.570285]  alg_test_skcipher+0xc0/0x1e8
> [   34.574294]  alg_test.part.32+0xb0/0x378
> [   34.578214]  alg_test+0x3c/0x68
> [   34.581354]  cryptomgr_test+0x44/0x50
> [   34.585015]  kthread+0x118/0x120
> [   34.588241]  ret_from_fork+0x10/0x18
> [   34.591818] Code: d5384101 b9402262 b9401020 11000400 (f8626a98)
> [   34.597910] ---[ end trace 3e4862893c99a748 ]---
> 

Thanks for testing this.
I'll fix it in v5.

Iulia