[dpdk-dev] [Bug 669] Integrity check failure occurs when ciphering and integrity both enabled with snow3g algorithm

[bugzilla at dpdk.org]

https://bugs.dpdk.org/show_bug.cgi?id=669

            Bug ID: 669
           Summary: Integrity check failure occurs when ciphering and
                    integrity both enabled with snow3g algorithm
           Product: DPDK
           Version: 20.02
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: critical
          Priority: Normal
         Component: cryptodev
          Assignee: dev at dpdk.org
          Reporter: tiger.liu at mavenir.com
  Target Milestone: ---

Created attachment 153
  --> https://bugs.dpdk.org/attachment.cgi?id=153&action=edit
information about the environment

Environment: 200Mbps traffic, 1 core, crypto_snow3g PMD without QAT for
ciphering and integrity

Detail: 
1.Integrity check failure immediately occurs when sending 200Mbps traffic.
2.Not every packets are failing in the integrity check, the situation is like
this: packet NO.1 passed the check, NO.2-9 failed and 10-18 passed again, these
numbers are not stable every time.
3.The procedure of processing packets is firstly appending 4 bytes MACI by
integrity algorithm then ciphering the packets including the MACI.
4.Checking the error packets with dpdk-test-crypto-perf tool, the ciphered data
does not match with the original data.
5.Before pushing the data for ciphering by function
rte_cryptodev_enqueue_burst() to DPDK, the content is the same as the original
packets and the 4 bytes MACI is correct.
6.After getting the ciphered data by rte_cryptodev_dequeue_burst(),the ciphered
data is not correct, which probably means some error happens in DPDK code.

Additionally, Integrity check failure won’t occur when ciphering and integrity
both enabled with AES algorithm in the same environment. 

We have got a response from Jayakumar Muthurajan, he said it is a missing
feature in the crypto-test when the size is different for encryption and
authentication. Please help check it.

-- 
You are receiving this mail because:
You are the assignee for the bug.