aboutsummaryrefslogtreecommitdiffstats
path: root/lib
AgeCommit message (Collapse)Author
2015-12-21Make peer handling more efficientAnders Svensson
Each service process maintains a dictionary of peers, mapping an application alias to a {pid(), #diameter_caps{}} list of connected peers. These lists are potentially large, peers were appended to the end of the list for no particular reason, and these long lists were constructed/deconstructed when filtering them for pick_peer callbacks. Many simultaneous outgoing request could then slow the VM to a crawl, with many scheduled processes mired in list manipulation. The pseudo-dicts are now replaced by plain ets tables. The reason for them was (once upon a time) to have an interface interchangeable with a plain dict for debugging purposes, but strict swapablity hasn't been the case for some time now, and in practice a swap has never taken place. Additional tables mapping Origin-Host/Realm have also been introduced, to minimize the size of the peers lists when peers are filtered on host/realm. For example, a filter like {any, [{all, [realm, host]}, realm]} is probably a very common case: preferring a Destination-Realm/Host match before falling back on Destination-Realm alone. This is now more efficiently (but not equivalently) expressed as {first, [{all, [realm, host]}, realm]} to stop the search when the best match is made, and extracts peers from host/realm tables instead of searching through the list of all peers supporting the application in question. The code to try and start with a lookup isn't exhaustive, and the 'any' filter is still as inefficient as previously.
2015-12-21Remove unnecessary erlang:monitor/2 qualificationAnders Svensson
See commit 862af31d.
2015-12-20Update release notesErlang/OTP
2015-12-20Merge branch 'anders/diameter/17.5.6.7/OTP-13211' into maint-17Erlang/OTP
* anders/diameter/17.5.6.7/OTP-13211: vsn -> 1.9.2.2 Update/fix appup for 17.5.6.7 Be resilient to diameter_service state upgrades
2015-12-20Merge branch 'anders/diameter/request_leak/OTP-13137' into maint-17Erlang/OTP
* anders/diameter/request_leak/OTP-13137: Fix request table leak at retransmission Fix request table leak at exit signal
2015-12-20Merge branch 'anders/diameter/17/watchdog/OTP-12969' into maint-17Erlang/OTP
* anders/diameter/17/watchdog/OTP-12969: Fix watchdog function_clause
2015-12-20Merge branch 'anders/diameter/M-bit/OTP-12947' into maint-17Erlang/OTP
* anders/diameter/M-bit/OTP-12947: Add service_opt() strict_mbit
2015-12-20vsn -> 1.9.2.2Anders Svensson
2015-12-20Update/fix appup for 17.5.6.7Anders Svensson
OTP-12947 strict_mbit OTP-12969 watchdog function_clause OTP-13137 request leak diameter_config (that allows the new option) should be loaded after the others. Anchor was missing from one regexp. Patches did not accumulate through older versions.
2015-12-20Be resilient to diameter_service state upgradesAnders Svensson
By not failing in code that looks up state: pick_peer and service_info.
2015-12-09Fix request table leak at retransmissionAnders Svensson
In the case of retranmission, a prepare_retransmit callback could modify End-to-End and/or Hop-by-Hop identifiers so that the resulting diameter_request entry was not removed, since the removal was of entries with the identifiers of the original request. The chances someone doing this in practice are probably minimal.
2015-12-09Fix request table leak at exit signalAnders Svensson
The storing of request records in the ets table diameter_request was wrapped in a try/after so that the latter would unconditionally remove written entries. The problem is that it didn't deal with the process exiting as a result of an exit signal, since this doesn't raise in an exception. Since the process in question applies callbacks to user code, we can potentially be linked to other process and exit as a result. Trapping exits changes the current behaviour of the process, so spawn a monitoring process that cleans up upon reception of 'DOWN'.
2015-12-03Prepare releaseErlang/OTP
2015-12-03Merge branch 'ia/ssl/maint-17/backport-of-18-fix' into maint-17Erlang/OTP
* ia/ssl/maint-17/backport-of-18-fix: ssl: Prepare for release ssl: Do not crash on proprietary hash_sign algorithms
2015-12-03ssl: Prepare for releaseIngela Anderton Andin
2015-12-01Update appupRaimo Niskanen
2015-11-19Testcase for inet:setopts/2 multiple raw optionsRaimo Niskanen
2015-11-19Fix inet:setopts/2 to take multiple raw optionsRaimo Niskanen
2015-09-07Prepare releaseErlang/OTP
2015-09-07Fix watchdog function_clauseAnders Svensson
Commit 4f365c07 introduced the error on set_watchdog/2, as a consequence of timeout/1 returning stop, which only happens with accepting transports with {restrict_connections, false}.
2015-08-25Add service_opt() strict_mbitAnders Svensson
There are differing opinions on whether or not reception of an arbitrary AVP setting the M-bit is an error. 1.3.4 of RFC 6733 says this about how an existing Diameter application may be modified: o The M-bit allows the sender to indicate to the receiver whether or not understanding the semantics of an AVP and its content is mandatory. If the M-bit is set by the sender and the receiver does not understand the AVP or the values carried within that AVP, then a failure is generated (see Section 7). It is the decision of the protocol designer when to develop a new Diameter application rather than extending Diameter in other ways. However, a new Diameter application MUST be created when one or more of the following criteria are met: M-bit Setting An AVP with the M-bit in the MUST column of the AVP flag table is added to an existing Command/Application. An AVP with the M-bit in the MAY column of the AVP flag table is added to an existing Command/Application. The point here is presumably interoperability: that the command grammar should specify explicitly what mandatory AVPs much be understood, and that anything more is an error. On the other hand, 3.2 says thus about command grammars: avp-name = avp-spec / "AVP" ; The string "AVP" stands for *any* arbitrary AVP ; Name, not otherwise listed in that Command Code ; definition. The inclusion of this string ; is recommended for all CCFs to allow for ; extensibility. This renders 1.3.4 pointless unless "*any* AVP" is qualified by "not setting the M-bit", since the sender can effectively violate 1.3.4 without this necessitating an error at the receiver. If clients add arbitrary AVPs setting the M-bit then request handling becomes more implementation-dependent. The current interpretation in diameter is strict: if a command grammar doesn't explicitly allow an AVP setting the M-bit then reception of such an AVP is regarded as an error. The strict_mbit option now allows this behaviour to be changed, false turning all responsibility for the M-bit over to the user.
2015-08-14debugger: Fix broken debugger:quick/3 startDan Gudmundsson
Too much code was removed in commit 560f73141af
2015-08-13Update release notesErlang/OTP
2015-08-13Merge branch 'anders/diameter/17.5.6.3/OTP-12927' into maint-17Erlang/OTP
* anders/diameter/17.5.6.3/OTP-12927: vsn -> 1.9.2.1 Update appup for 17.5.6.3
2015-08-13Merge branch 'anders/diameter/17/time/OTP-12926' into maint-17Erlang/OTP
* anders/diameter/17/time/OTP-12926: Simplify time manipulation Remove use of monotonic time in pre-18 code Remove unnecessary redefinition of erlang:max/2
2015-08-13Merge branch 'anders/diameter/grouped_errors/OTP-12930' into maint-17Erlang/OTP
* anders/diameter/grouped_errors/OTP-12930: Fix decode of Grouped AVPs containing errors Simplify logic Simplify logic
2015-08-13Merge branch 'anders/diameter/transport/OTP-12929' into maint-17Erlang/OTP
* anders/diameter/transport/OTP-12929: Fix start order of alternate transports Log discarded answers
2015-08-13Merge branch 'anders/diameter/lcnt/OTP-12912' into maint-17Erlang/OTP
* anders/diameter/lcnt/OTP-12912: Make ets diameter_stats a set Remove unnecessary sorting in stats suite Set ets {write_concurrency, true} on diameter_stats Don't start watchdog timers unnecessarily Remove unnecessary erlang:monitor/2 qualification Add missing watchdog suite clause
2015-08-13Merge branch 'anders/diameter/caseless/OTP-12902' into maint-17Erlang/OTP
* anders/diameter/caseless/OTP-12902: Match allowable peer addresses case insensitively Replace calls to module inet_parse to equivalents in inet
2015-08-13Merge branch 'anders/diameter/grouped_decode/OTP-12879' into maint-17Erlang/OTP
* anders/diameter/grouped_decode/OTP-12879: Fix relay encode of decoded diameter_avp lists
2015-08-13Merge branch 'anders/diameter/decode/OTP-12891' into maint-17Erlang/OTP
* anders/diameter/decode/OTP-12891: Don't compute AVP list length unnecessarily at AVP decode
2015-08-13Merge branch 'anders/diameter/decode/OTP-12871' into maint-17Erlang/OTP
* anders/diameter/decode/OTP-12871: Don't traverse errors list unnecessarily when detecting missing AVPs Don't flag AVP as missing as a consequence of decode error Correct inaccurate doc Truncate potentially large terms passed to diameter_lib:log/4
2015-08-07Make ets diameter_stats a setAnders Svensson
There's no need for it to be ordered, and the ordering has been seen to have an unexpectedly negative impact on performance in some cases. Order when retrieving statistics instead, so as not to change the presentation in diameter:service_info/2.
2015-08-07Remove unnecessary sorting in stats suiteAnders Svensson
The ordering of (ets) diameter_stats (also unnecessary) ensures the sorting.
2015-08-05Simplify time manipulationAnders Svensson
By doing away with more wrapping that the parent commit started to remove.
2015-08-05Remove use of monotonic time in pre-18 codeAnders Svensson
This has been seen to be a bottleneck at high load: each undef results in a loop out to the code server. Originally implemented as suggested in the erts user's guide, in commits e6d19a18 and d4386254.
2015-08-05vsn -> 1.9.2.1Anders Svensson
2015-08-05Update appup for 17.5.6.3Anders Svensson
OTP-12871: 5005 decode OTP-12791: decode performance OTP-12879: grouped decode OTP-12902: caseless address match OTP-12912: fewer timers OTP-12926: pre-18 time diameter_lib must be loaded after modules calling its time-related functions (that have been removed).
2015-08-05Fix start order of alternate transportsAnders Svensson
A transport configured with diameter:add_transport/2 can be passed multiple transport_module/transport_config tuples in order to specify alternate configuration, modules being attempted in order until one succeeds. This is primarily for the connecting case, to allow a transport to be configured to first attempt connection over SCTP, and then TCP in case SCTP fails, with configuration like that documented: {transport_module, diameter_sctp}, {transport_config, [...], 5000}, {transport_module, diameter_tcp}, {transport_config, [...]} If the options are the same in both cases, another possibility would be configuration like this, which attaches the same transport_config to both modules: {transport_module, diameter_sctp}, {transport_module, diameter_tcp}, {transport_config, [...], 5000}, However, in this case the start order was reversed relative to the documented order: first tcp, then sctp. This commit restores the intended order.
2015-08-05Log discarded answersAnders Svensson
To diameter_lib:log/4, which was last motivated in commit 39acfdb0.
2015-08-04Remove unnecessary redefinition of erlang:max/2Anders Svensson
The function already operates on arbitrary terms.
2015-08-04Fix relay encode of decoded diameter_avp listsAnders Svensson
Commit c74b593a fixed the problem that a decoded deep diameter_avp list couldn't be encoded, but did so in the wrong way: there's no need to reencode component AVPs since the Grouped AVP itself already contains the encoded binary. The blunder caused diameter_codec:pack_avp/1 to fail if the first element of the AVP list to be encoded was itself a list. Thanks to Andrzej TrawiƄski for reporting the problem.
2015-08-04Match allowable peer addresses case insensitivelyAnders Svensson
Both diameter_tcp and diameter_sctp can be configured with one or more IP addresses from which connections should be accepted (an 'accept' tuple), specified either as a tuple-valued address or as a regular expression. In the latter case, peer addresses are mapped to string using inet:ntoa/1 and the result matched against the regexp. Since (ipv6) addresses are case insensitive, this should also be the case with the match, but was not.
2015-08-04Replace calls to module inet_parse to equivalents in inetAnders Svensson
Commits b563c796 (R16B) and 0fad6449 (R16B02) added parse_address/1 and ntoa/1 to module inet, providing documented alternatives to address/1 and ntoa/1 in the undocumented (save comments in inet(3)) inet_parse.
2015-08-04Don't compute AVP list length unnecessarily at AVP decodeAnders Svensson
This has had a hugely negative impact on performance when decoding messages containing many AVP: each decode of an AVP having variable arity computed the length of the list of previously decoded AVPs when checking that the allowed arity was not exceeded, even if the allowed arity was infinite, making for O(n^2) cost. Here are some execution times, for diameter_codec:decode/2 on a representative message with n integer AVPs in the Common application (on the host at hand): Before After ------- --------- n = 1K 5 ms 2 ms n = 10K 500 ms 25 ms n = 100K 75 sec 225 ms n = 1M 2.6 sec Note the nearly linear increase following the change. Remove the dire documentation warning for incoming_maxlen as a consequence. It can still be useful to set, but not doing so won't have the same consequences as previously.
2015-08-04Don't traverse errors list unnecessarily when detecting missing AVPsAnders Svensson
Since the list can potentially be long.
2015-08-04Don't flag AVP as missing as a consequence of decode errorAnders Svensson
The decode of an incoming Diameter message uses the record representation to determine whether or not an AVP has been received with the expected arity, the number of AVPs in each field following decode being compared with the arity specified in the message grammar. The problem with this is that decode failure isn't reflected in the record representation, so that an AVP can be appended to the errors field of a diameter_packet record despite an entry for the same AVP already existing. This isn't a fault as much as a misleading error indication, but now only append AVPs that aren't already represented.
2015-08-04Correct inaccurate docAnders Svensson
The warning report was removed in commit 00584303.
2015-08-04Truncate potentially large terms passed to diameter_lib:log/4Anders Svensson
Last visited in commit 00584303.
2015-08-04Set ets {write_concurrency, true} on diameter_statsAnders Svensson
lcnt:inspect/1 recently showed this: lock id #tries collisions [%] time [us] ----- --- ------- --------------- ---------- db_tab diameter_stats 932920 92.9326 330332554