aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)Author
2017-06-12Remove minor diameter_config bloatAnders Svensson
Folded when I should have mapped.
2017-06-12Fix maximum AVP arity checkAnders Svensson
2017-06-12Use avp_arity/1 when creating a zero groupAnders Svensson
Converting with list_to_binary/1 appears to be faster than the equivalent binary comprehension: << (z(F,A)) || {F,A} <- avp_arity(Name) >>
2017-06-12Use avp_arity/1 when detecting missing AVPsAnders Svensson
Recursing over the entire list of arities and values is faster than retrieving them one at a time.
2017-06-12Add dictionary function avp_arity/1Anders Svensson
2017-06-12Don't prepend bit for sub binary optimizationAnders Svensson
2017-06-12Remove upgrade-related codeAnders Svensson
This and subsequent commits are destined for OTP 20.0.
2017-06-12Simplify value setting at decodeAnders Svensson
2017-06-12Simplify AVP decode as mapfoldlAnders Svensson
2017-06-12Be more flexible in what can be packed as 'AVP'Anders Svensson
In particular, allow {Name, Value} and {Dict, Name, Value} without requiring a diameter_avp wrapper.
2017-06-12Remove unnecessary value mappingAnders Svensson
Since value is ignored.
2017-06-12Use dictionary '#new-'/1 when creating CEAAnders Svensson
Which is the equivalent of what was done with '#new-'/1 and '#set-'/2.
2017-06-12Optimize sub binary creationAnders Svensson
2017-06-12Optimize sub binary creationAnders Svensson
base/diameter_codec.erl:716: Warning: OPTIMIZED: creation of sub binary delayed
2017-06-12Optimize sub binary creationAnders Svensson
base/diameter_codec.erl:545: Warning: OPTIMIZED: creation of sub binary delayed
2017-06-12Optimize sub binary creationAnders Svensson
base/diameter_codec.erl:600: Warning: OPTIMIZED: creation of sub binary delayed
2017-06-12Don't match for no useful reasonAnders Svensson
2017-06-12Avoid modifying records during encodeAnders Svensson
2017-06-12Don't create intermediate terms unnecessarily during encodeAnders Svensson
2017-06-12Don't create intermediate binaries unnecessarily during encodeAnders Svensson
Dict:avp(encode, Value, Name) no longer needs to return a binary, only an iolist(). Message encode runs list_to_binary/1 to convert accumulated lists into a message binary.
2017-06-12Fix encode of Grouped AVP as {Dict, Name, Data}Anders Svensson
This is a special case to allow encode of something other than an iolist. Eg. #diameter_avp{data = {diameter_gen_base_rfc6733, 'Proxy-Info', [{'Proxy-Host', "HOST"}, {'Proxy-State', "STATE"}]}} Only worked as expected for AVPs of type other than Grouped.
2017-06-12Remove minor encode bloatAnders Svensson
As when detecting missing AVPs, extract a list of field/value pairs in one step, which looks to be slightly more efficient. Flattening the list was unnecessary since the result is passed to list_to_binary.
2017-06-12Make encoding of Diameter header more efficientAnders Svensson
On the same theme as the parent commit, building binaries in fewer steps.
2017-06-12Make encoding of diameter_avp records more efficientAnders Svensson
Prepend the header in a single step. Before: {[{{diameter_codec,pack_avp,1}, 7000, 126.074, 51.058}], { {diameter_codec,pack_avp,2}, 7000, 126.074, 51.058}, % [{{diameter_codec,pack_avp,5}, 7000, 51.144, 25.758}, {{diameter_codec,pad,2}, 7000, 23.844, 23.570}, {suspend, 1, 0.028, 0.000}]}. After: {[{{diameter_codec,pack_avp,1}, 7000, 78.563, 26.986}], { {diameter_codec,pack_avp,2}, 7000, 78.563, 26.986}, % [{{diameter_codec,pack_avp,6}, 7000, 51.459, 26.381}, {suspend, 4, 0.118, 0.000}]}.
2017-06-12Replace slow binary comprehensionsAnders Svensson
Which appears to be about an order of magnitude slower than just creating a binary of the desired size.
2017-06-12Decode message header in a single matchAnders Svensson
2017-06-12Simplify missing AVP detectionAnders Svensson
By using the existing '#get-'/1 in generated dictionary modules to retrieve fields and values at the same time. Before: {[{{diameter_gen_base_rfc6733,missing,3}, 1000, 211.722, 8.741}, {{diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',4},12000, 0.000, 95.764}], { {diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',4},13000, 211.722, 104.505}, % [{{diameter_gen_base_rfc6733,'#get-',2}, 12000, 49.917, 28.221}, {{diameter_gen_base_rfc6733,has_arity,2}, 12000, 31.811, 23.442}, {{diameter_gen_base_rfc6733,avp_arity,2}, 12000, 21.076, 20.975}, {garbage_collect, 457, 3.918, 3.918}, {suspend, 31, 0.495, 0.000}, {{diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',4},12000, 0.000, 95.764}]}. After: {[{{diameter_gen_base_rfc6733,missing,3}, 1000, 134.098, 2.402}, {{diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',3},13000, 0.000, 77.327}], { {diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',3},14000, 134.098, 79.729}, % [{{diameter_gen_base_rfc6733,has_arity,2}, 12000, 31.084, 22.913}, {{diameter_gen_base_rfc6733,avp_arity,2}, 12000, 20.526, 20.440}, {garbage_collect, 253, 2.504, 2.504}, {suspend, 17, 0.255, 0.000}, {{diameter_gen_base_rfc6733,'-missing/3-lc$^0/1-0-',3},13000, 0.000, 77.327}]}.
2017-06-12Use maps when detecting missing AVPsAnders Svensson
Instead of the slower sets. Bump application dependencies to 17.5, even though earlier versions may do fine.
2017-06-12Make AVP splitting more efficientAnders Svensson
Profiling with fprof showed this prior to this commit: {[{{diameter_codec,decode,3}, 1000, 231.122, 4.092}, {{diameter_codec,collect_avps,1}, 1000, 0.000, 3.929}], { {diameter_codec,collect_avps,1}, 2000, 231.122, 8.021}, % [{{diameter_codec,collect_avps,3}, 1000, 222.932, 11.644}, {garbage_collect, 19, 0.169, 0.169}, {{diameter_codec,collect_avps,1}, 1000, 0.000, 3.929}]}. {[{{diameter_codec,collect_avps,1}, 1000, 222.932, 11.644}, {{diameter_codec,collect_avps,3}, 7000, 0.000, 68.186}], { {diameter_codec,collect_avps,3}, 8000, 222.932, 79.830}, % [{{diameter_codec,split_avp,1}, 7000, 120.886, 72.382}, {{erlang,setelement,3}, 7000, 21.830, 21.830}, {garbage_collect, 48, 0.386, 0.386}, {{diameter_codec,collect_avps,3}, 7000, 0.000, 68.186}]}. Note the time consumed in split_avp/1 and erlang:setelement/3. This commit does more matching in one go, without intermediate results, giving this: {[{{diameter_codec,decode,3}, 1000, 42.512, 3.701}, {{diameter_codec,collect_avps,1}, 1000, 0.000, 3.594}], { {diameter_codec,collect_avps,1}, 2000, 42.512, 7.295}, % [{{diameter_codec,collect_avps,3}, 1000, 35.217, 4.577}, {{diameter_codec,collect_avps,1}, 1000, 0.000, 3.594}]}. {[{{diameter_codec,collect_avps,1}, 1000, 35.217, 4.577}, {{diameter_codec,collect_avps,3}, 7000, 0.000, 27.754}], { {diameter_codec,collect_avps,3}, 8000, 35.217, 32.331}, % [{garbage_collect, 262, 2.647, 2.647}, {suspend, 9, 0.239, 0.000}, {{diameter_codec,collect_avps,3}, 7000, 0.000, 27.754}]}.
2017-06-12Remove bloat from zero encodeAnders Svensson
Don't call a function when we know the result, and consistently return a binary.
2017-06-12Add {init,end}_per_suite/1 to codec suiteAnders Svensson
Do nothing, but convenient for adding trace.
2017-06-12Fix comment typosAnders Svensson
2017-06-12Remove client/server string decode from traffic suiteAnders Svensson
Decode on both ends or not, since the choice doesn't affect the peer.
2017-06-12Add diameter_sctp option packetAnders Svensson
To determine the wrapping of messages passed to recv callbacks and into diameter. The default passing of the input stream in transport_data is probably of no practical use, but has been set since time immemorial.
2017-06-12Add diameter_sctp send/recv callbacksAnders Svensson
Corresponding to diameter_tcp callbacks a few commits back. Exercise the callbacks in the traffic suite.
2017-06-12Let diameter_tcp send/recv callbacks deal in diameter_packetAnders Svensson
To let a recv callback for an incoming request set transport_data and have it returned in a send callback.
2017-06-12Randomly select traffic testcasesAnders Svensson
Since the number of configuration variants tested makes for (too) many. Randomly select a subset of testcases in each configuration group.
2017-06-12Exercise diameter_tcp message callbacks in traffic suiteAnders Svensson
2017-06-12Exercise diameter_{tcp,sctp} sender in traffic suiteAnders Svensson
2017-06-12Remove upgrade from diameter_trafficAnders Svensson
2017-06-12Add diameter_tcp send/recv callbacksAnders Svensson
From the receiver process, that can return binaries to send/receive and stop the transport process from reading on the socket. This is still undocumented, and may change.
2017-06-11Make diameter_{tcp,sctp} sender configurableAnders Svensson
With sends still from the receiving process by default, since changing the default behaviour may well have negative effects. A separate sender probably implies a greater need for some form of load regulation for one, since a blocking send would no longer imply that incoming messages are no longer recevied. Dealing with this could result in the same deadlock that the sending process intends to avoid, but the user should be in control over how/when incoming traffic is regulated.
2017-06-11Remove upgrade from diameter_sctp; tweak diameter_tcp to matchAnders Svensson
Added in commit 2afd1fe5. Only rename variables in diameter_tcp, no functional change.
2017-06-11Fix incomprehensible dialyzer warningAnders Svensson
This: diameter_tcp.erl:241: Record construction #transport{parent::'false',ssl::boolean() | maybe_improper_list(),frag::<<>>,tref::'false',flush::'false',pending::0,reset::{1 | 4,0 | 2},throttled::boolean(),q::{0,queue:queue(_)},monitor::'undefined' | pid()} violates the declared type of field parent::pid() The problem isn't #transport.pid at all, it's #monitor.pid, and the only relation is that the pid that's assigned to the latter is also (later) assigned to the former. There is no record construction that assigns false to #transport.parent. Introduced in commit 33a535e4.
2017-06-11Simplify acks to transport processesAnders Svensson
What's interesting when implementing some form of load regulation is when an incoming request has been answered or discarded. Acknowledge exactly this, not the identity of handler processes as previously. A transport process can request acks of nonforthcoming answers by sending {diameter, ack} to the parent peer_fsm, a handler processes identifies itself with a {handler, pid()} message, and the peer_fsm monitors on this to be able to send a notification to the transport if the handler dies before sending an answer.
2017-06-11Strip throttling callbacks from diameter_tcpAnders Svensson
Commits starting at 472a080c added a throttle_cb option to diameter_tcp to let a callback apply backpressure when it decides that additional requests should not be read. It didn't provide a hook for knowing that an answer was sent however, which is needed when sends no longer take place in the receiver process, and is more complicated than it should be. Strip it all away, in preparation for a simpler incarnation.
2017-06-11Deal with (another) SCTP association id quirk on SolarisAnders Svensson
Shutdown events have been seen to get a different association id. For example, first incoming message with association id = 0: + {trace_ts,<6421.268.0>,call, {diameter_sctp,handle_info, [{sctp,#Port<6421.2588>, {10,67,16,179}, 44159, {[{sctp_sndrcvinfo,0,0,[],0,0,0,269950872,269950872,0}], <<1,0,0,156,128,0,1,1,0,0,0,0,6,193,40,137,6,193,40,137,0,0, 1,8,64,0,0,30,67,45,49,51,52,50,49,55,52,52,49,46,101,114, 108,97,110,103,46,111,114,103,0,0,0,0,1,40,64,0,0,18,101, 114,108,97,110,103,46,111,114,103,0,0,0,0,1,1,64,0,0,14,0, 1,127,0,0,1,0,0,0,0,1,10,64,0,0,12,0,0,48,57,0,0,1,13,0,0, 0,20,79,84,80,47,100,105,97,109,101,116,101,114,0,0,1,22, 64,0,0,12,0,0,0,1,0,0,1,2,64,0,0,12,0,0,0,0,0,0,1,3,64,0,0, 12,0,0,0,3>>}}, {transport,<6421.252.0>,accept,#Port<6421.2588>,true,undefined, {32,32}, 0,undefined}]}, {1493,21505,577938}} Later, a shutdown event with association id 1536: + {trace_ts,<6421.268.0>,call, {diameter_sctp,handle_info, [{sctp,#Port<6421.2588>, {10,67,16,179}, 44159, {[],{sctp_shutdown_event,1536}}}, {transport,<6421.252.0>,accept,#Port<6421.2588>,0,undefined, {32,32}, 2,<6421.304.0>}]}, {1493,21505,746929}} Both this and the grandparent commit are on this: $ uname -a SunOS beren 5.10 Generic_118833-33 sun4v sparc SUNW,Netra-T2000
2017-06-11Use binary:copy/2 when generating largish data in test suitesAnders Svensson
Faster than lists:duplicate/2.
2017-06-11Deal with SCTP association id quirk on SolarisAnders Svensson
In particular, that the association id received in messages on a one-to-one socket after peeloff may be different from the id received on the listen socket at comm_up. This seems odd, since it's then not possible to send until the id is discovered by reception of an SCTP message containing it, but it's unclear if this is a bug or a feature, or if it's specific to certain platforms. Treat it as a feature in this commit, and get the association id as mentioned, an incoming CER being expected before anything is sent. Commit da3e5d67 has more history.
2017-06-11Make skipped SCTP testcases more visible in traffic suiteAnders Svensson
By explicitly skipping instead of omitting testcases from groups.