[dpdk-dev] [Bug 307] ACL (librte_acl) field of type RANGE and size U32 is not working properly

[dpdk-dev] [Bug 307] ACL (librte_acl) field of type RANGE and size U32 is not working properly

[bugzilla]

https://bugs.dpdk.org/show_bug.cgi?id=307

            Bug ID: 307
           Summary: ACL (librte_acl) field of type RANGE and size U32 is
                    not working properly
           Product: DPDK
           Version: 18.11
          Hardware: All
                OS: All
            Status: CONFIRMED
          Severity: normal
          Priority: Normal
         Component: other
          Assignee: dev@dpdk.org
          Reporter: ido@cgstowernetworks.com
  Target Milestone: ---

Created attachment 47
  --> https://bugs.dpdk.org/attachment.cgi?id=47&action=edit
acl-test for single field match - can be compiled like any dpdk/example

Hi,

Such ACL field doesn't seem to work properly 
~~~
{
        .type = RTE_ACL_FIELD_TYPE_RANGE,
        .size = sizeof(uint32_t),
...
}
~~~

I found same complain here
https://mails.dpdk.org/archives/users/2017-June/002096.html (by Doohwan Lee)
with no resolution.

Also all DPDK ACL test/examples RTE_ACL_FIELD_TYPE_RANGE fields are with
uint16_t and never with uint32_t

Attached is a small acl-test app 
It sets a single ACL field (though there are few in the acl_fields array all
the rest are don't-care)
And matches a data argument (no need for traffic/packets)

usage:
~~~
./acl-test [EAL options] -- [app options]
app options:
        --size=16|32
        [--type=RANGE --min=<min> --max=<max>]
        [--type=BITMASK --value=<value> --bitmask=<bitmask>]
        --data=<data>
example:
        ./acl-test --no-huge -c 1 -- --size=32 --type=RANGE --min=100 --max=200
--data=150
~~~

Output example for RANGE U16
~~~
cgs@ubuntu:~/workspace/KAZ/output/build/arpeggio/host/dpdk/app$ ./acl-test
--no-huge -c 1 -- --size=16 --type=RANGE --min=1 --max=500 --data=250
EAL: Detected 4 lcore(s)
EAL: Detected 1 NUMA nodes
EAL: Multi-process socket /run/user/1000/dpdk/rte/mp_socket
EAL: Probing VFIO support...
EAL: Started without hugepages support, physical addresses not available
EAL: PCI device 0000:02:01.0 on NUMA socket -1
EAL:   Invalid NUMA socket, default to 0
EAL:   probe driver: 8086:100f net_e1000_em
EAL: PCI device 0000:02:06.0 on NUMA socket -1
EAL:   Invalid NUMA socket, default to 0
EAL:   probe driver: 8086:100f net_e1000_em
RULE 1: | 00/00 | 0001/01F4   0000/0000   | 00000000/FFFFFFFF       |
00000000/00000000       | 0x1-0x1-0x1 
acl context <ACL w RANGE U16>@0x7f9dede13b40
  socket_id=0
  alg=3
  max_rules=256
  rule_size=96
  num_rules=1
  num_categories=1
  num_tries=1
250 is in range 1-500    should MATCH
ACL MATCH        test PASSED
~~~

Bad output example for RANGE U32 (same test just size change)
~~~
./acl-test --no-huge -c 1 -- --size=32 --type=RANGE --min=1 --max=500
--data=250
EAL: Detected 4 lcore(s)
EAL: Detected 1 NUMA nodes
EAL: Multi-process socket /run/user/1000/dpdk/rte/mp_socket
EAL: Probing VFIO support...
EAL: Started without hugepages support, physical addresses not available
EAL: PCI device 0000:02:01.0 on NUMA socket -1
EAL:   Invalid NUMA socket, default to 0
EAL:   probe driver: 8086:100f net_e1000_em
EAL: PCI device 0000:02:06.0 on NUMA socket -1
EAL:   Invalid NUMA socket, default to 0
EAL:   probe driver: 8086:100f net_e1000_em
RULE 1: | 00/00 | 0000/FFFF   0000/0000   | 00000001/000001F4       |
00000000/00000000       | 0x1-0x1-0x1 
acl context <ACL w RANGE U32>@0x7ff3d08a3b40
  socket_id=0
  alg=3
  max_rules=256
  rule_size=96
  num_rules=1
  num_categories=1
  num_tries=1
250 is in range 1-500    should MATCH
ACL NO-MATCH     test FAILED!
~~~

Thanx,
- Ido

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