princeton-cabernet / beaucoup Goto Github PK

There is another issue here. When I run the python simulator, it reports errors as follows.

$ python3 py/simulator.py --seed=1 test/IR.json test/caida0.npy test/reports.npz
Traceback (most recent call last): File "py/simulator.py", line 34, in <module> lib_eval.simulate_run(args.npy_filename, args.out_filename, IR, args.seed, debug=args.verbose) File "/home/ltang/Workspace/research/store/BeauCoup/py/core/lib_eval.py", line 189, in simulate_run val=h_compute[j](Trace[i]) File "/home/ltang/Workspace/research/store/BeauCoup/py/core/lib_eval.py", line 108, in <lambda> return lambda val:pull_pack_and_compute(list_indexes,val, mask, seedbytes, Pack_formatstring) File "/home/ltang/Workspace/research/store/BeauCoup/py/core/lib_eval.py", line 105, in pull_pack_and_compute base+=struct.pack(Pack_formatstring[i],value_list[i]) struct.error: short format requires (-0x7fff - 1) <= number <= 0x7fff

The IR.json is compiled from the query_examples.yaml file in BeauCoup folder.

How to reproduce:

• Clone repository
• Install dependencies
• Run python3 py/compiler.py -v --gamma=1.0 query_examples.yaml IR.json from repository root
• Run python3 py/p4gen.py -v IR.json p4src/beaucoup.p4template beaucoup.p4

I pinpointed this issue to the following blocks in the Jinja2 template:

{% for i in range(16) %}
{% if K1 in keydefn[i] %}
  {{i}} : {{A1}}();
{% endif %}
{% if K2 in keydefn[i] %}
  {{i}} : {{A2}}();
{% endif %}
{% endfor %}

The template receives the following dictionary as keydefn argument, which does not have a "0"-element:

{1: ('hdr.ipv4.src_addr',), 2: ('hdr.ipv4.dst_addr', 'hdr.tcp.dst_port'), 3: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr'), 4: ('hdr.ipv4.dst_addr',), 5: ('hdr.ipv4.src_addr', 'hdr.tcp.dst_port'), 6: ('hdr.ipv4.src_addr', 'hdr.udp.dst_port'), 7: ('hdr.ipv4.dst_addr', 'hdr.udp.dst_port'), 8: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr', 'hdr.tcp.src_port'), 9: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr', 'hdr.udp.src_port'), 10: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr', 'hdr.tcp.dst_port'), 11: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr', 'hdr.udp.dst_port'), 12: ('hdr.ipv4.src_addr', 'hdr.ipv4.dst_addr', 'hdr.tcp.src_port', 'hdr.tcp.dst_port')}

Am I doing something wrong? Or is there an issue with the template?

The program work just fine, but I was wondering why you choose P4 switch?
Is there a necessary need for using P4 switch.

when running python3 py/trace_prep.py ../../dataset/pcap/test.pcap test/caida0.npy I have the following error:

Traceback (most recent call last): File "py/trace_prep.py", line 22, in <module> nparray = lib_eval.parse_pcap_into_npy(args.PCAP_filename, args.count, args.verbose) File "/home/ltang/Workspace/research/store/BeauCoup/py/core/lib_eval.py", line 79, in parse_pcap_into_npy packets_numerical.append(tuple(to_list(pdict))) File "/home/ltang/Workspace/research/store/BeauCoup/py/core/lib_eval.py", line 73, in to_list line[i]=int(line[i]) TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

Note that test.pcap is a small pcap file extract from CAIDA18 traces

Hi Xiaoqi,
In py/core/lib_eval.py, line293, should it be
vect=trial_finish[QID] ?

I cannot get correct error results with vect=trial_finish[qid] .