aboutsummaryrefslogtreecommitdiffstats
path: root/lib/hipe
AgeCommit message (Collapse)Author
2018-09-03Introduce a put_tuple2 instructionBjörn Gustavsson
Sometimes when building a tuple, there is no way to avoid an extra `move` instruction. Consider this code: make_tuple(A) -> {ok,A}. The corresponding BEAM code looks like this: {test_heap,3,1}. {put_tuple,2,{x,1}}. {put,{atom,ok}}. {put,{x,0}}. {move,{x,1},{x,0}}. return. To avoid overwriting the source register `{x,0}`, a `move` instruction is necessary. The problem doesn't exist when building a list: %% build_list(A) -> [A]. {test_heap,2,1}. {put_list,{x,0},nil,{x,0}}. return. Introduce a new `put_tuple2` instruction that builds a tuple in a single instruction, so that the `move` instruction can be eliminated: %% make_tuple(A) -> {ok,A}. {test_heap,3,1}. {put_tuple2,{x,0},{list,[{atom,ok},{x,0}]}}. return. Note that the BEAM loader already combines `put_tuple` and `put` instructions into an internal instruction similar to `put_tuple2`. Therefore the introduction of the new instruction will not speed up execution of tuple building itself, but it will be less work for the loader to load the new instruction.
2018-08-17hipe_beam_to_icode: Correct translation of get_map_elementsBjörn Gustavsson
If one of the destination registers for get_map_elements is the same as the map source, extract that element last.
2018-07-27Change "can not" into "cannot"Raimo Niskanen
I did not find any legitimate use of "can not", however skipped changing e.g RFCs archived in the source tree.
2018-07-27Remove newline between can and notRaimo Niskanen
After this whitespace modification there should be no "can not"s separated by a newline in the entire OTP repository, so to find them all a simple git grep will do just fine.
2018-07-13docs: make clean all XMLDIRLukas Larsson
2018-06-19Prepare releaseErlang/OTP
2018-06-18Update copyright yearHenrik Nord
2018-06-14Merge pull request #1837 from kostis/hipe-lcm-21-fixBjörn Gustavsson
Fix a crash in HiPE's lazy code motion pass
2018-06-13Merge branch 'rickard/hipe-doc'Rickard Green
* rickard/hipe-doc: Update information about HiPE
2018-06-12No reason to refer to wxSizer in the testKostis Sagonas
2018-06-11Fix a crash ih HoPE's lazy code motion passKostis Sagonas
Some change in the BEAM compiler resulted in the creation of basic blocks that differed from those previously created by the compiler. As a result, the lazy code motion pass of RTL crashed when compiling some of the new code. Crashes were privately reported by @richcarl.
2018-05-30Update information about HiPERickard Green
2018-05-29Revert "Prepare release"Henrik
This reverts commit fd8e49b5bddceaae803670121b603b5eee8c5c08.
2018-05-25Prepare releaseErlang/OTP
2018-05-22hipe: Fix elimination of erlang:get_stacktrace/0Hans Bolinder
2018-05-16Merge branch 'richcarl/eliminate_lib_module/PR-1786/OTP-15072'Hans Bolinder
* richcarl/eliminate_lib_module/PR-1786/OTP-15072: Fix minor issues Eliminate call to ct:get_progname() in ts_erl_config Use \n escape instead of integer 10 Move error formatting to erl_error.erl and delete lib.erl Move extended parse functions in lib.erl to erl_eval.erl Move lib:eval_str/1 into mod_esi.erl Remove lib:progname/0 Eliminate call to lib:progname/1 in slave.erl Add ct:get_progname/0 Remove lib:error_message/2 Remove lib:flush_receive/0 Remove lib:send/2 and lib:sendw/2 Move lib:nonl/1 into yecc.erl
2018-05-07Merge pull request #1802 from michalmuskala/map-is-key-bifBjörn Gustavsson
Introduce is_map_key/2 guard BIF OTP-15037
2018-05-04Remove unused source files in lib/hipe/optBjörn Gustavsson
Three of the removed source files don't even compile. Uncompilable source files cause problems for tools such as scripts/diffable that want to compile all found source files.
2018-05-04Move error formatting to erl_error.erl and delete lib.erlRichard Carlsson
2018-04-29Introduce is_map_key/2 guard BIFMichał Muskała
This complements the `map_get/2` guard BIF introduced in #1784. Rationale. `map_get/2` allows accessing map fields in guards, but it might be problematic in more complex guard expressions, for example: foo(X) when map_get(a, X) =:= 1 or is_list(X) -> ... The `is_list/1` part of the guard could never succeed since the `map_get/2` guard would fail the whole guard expression. In this situation, this could be solved by using `;` instead of `or` to separate the guards, but it is not possible in every case. To solve this situation, this PR proposes a `is_map_key/2` guard that allows to check if a map has key inside a guard before trying to access that key. When combined with `is_map/1` this allows to construct a purely boolean guard expression testing a value of a key in a map. Implementation. Given the use case motivating the introduction of this function, the PR contains compiler optimisations that produce optimial code for the following guard expression: foo(X) when is_map(X) and is_map_key(a, X) and map_get(a, X) =:= 1 -> ok; foo(_) -> error. Given all three tests share the failure label, the `is_map_key/2` and `is_map/2` tests are optimised away. As with `map_get/2` the `is_map_key/2` BIF is allowed in match specs.
2018-04-26Move erl_types test into a common_test suiteSean Cribbs
The existing test/0 function in erl_types was not being run in the OTP test suite, and it had not been updated to match the implementation in the module (maps vs. dict). This commit removes the macros that excluded some functions, exports the functions now included in the module, and extracts the test into a new common_test suite, erl_types_SUITE.
2018-04-25Merge branch 'map-get-bif' of git://github.com/michalmuskala/otpBjörn Gustavsson
* 'map-get-bif' of git://github.com/michalmuskala/otp: Introduce map_get guard-safe function OTP-15037
2018-04-24Introduce map_get guard-safe functionMichał Muskała
Rationale Today all compound data types except for maps can be deconstructed in guards. For tuples we have `element/2` and for lists `hd/1` and `tl/1`. Maps are completely opaque to guards. This means matching on maps can't be abstracted into macros, which is often done with repetitive guards. It also means that maps have to be always selected whole from ETS tables, even when only one field would be enough, which creates a potential efficiency issue. This PR introduces an `erlang:map_get/2` guard-safe function that allows extracting a map field in guard. An alternative to this function would be to introduce the syntax for extracting a value from a map that was planned in the original EEP: `Map#{Key}`. Even outside of guards, since this function is a guard-BIF it is more efficient than using `maps:get/2` (since it does not need to set up the stack), and more convenient from pattern matching on the map (compare: `#{key := Value} = Map, Value` to `map_get(key, Map)`). Performance considerations A common concern against adding this function is the notion that "guards have to be fast" and ideally execute in constant time. While there are some counterexamples (`length/1`), what is more important is the fact that adding those functions does not change in any way the time complexity of pattern matching - it's already possible to match on map fields today directly in patterns - adding this ability to guards will niether slow down or speed up the execution, it will only make certain programs more convenient to write. This first version is very naive and does not perform any optimizations.
2018-04-16hipe: Eliminate use of erlang:get_stacktrace/0Björn Gustavsson
2018-03-21Merge pull request #1740 from rickard-green/rickard/signals/OTP-14589Rickard Green
Implementation of true asynchronous signaling between processes
2018-03-21Implementation of true asynchronous signaling between processesRickard Green
Communication between Erlang processes has conceptually always been performed through asynchronous signaling. The runtime system implementation has however previously preformed most operation synchronously. In a system with only one true thread of execution, this is not problematic (often the opposite). In a system with multiple threads of execution (as current runtime system implementation with SMP support) it becomes problematic. This since it often involves locking of structures when updating them which in turn cause resource contention. Utilizing true asynchronous communication often avoids these resource contention issues. The case that triggered this change was contention on the link lock due to frequent updates of the monitor trees during communication with a frequently used server. The signal order delivery guarantees of the language makes it hard to change the implementation of only some signals to use true asynchronous signaling. Therefore the implementations of (almost) all signals have been changed. Currently the following signals have been implemented as true asynchronous signals: - Message signals - Exit signals - Monitor signals - Demonitor signals - Monitor triggered signals (DOWN, CHANGE, etc) - Link signals - Unlink signals - Group leader signals All of the above already defined as asynchronous signals in the language. The implementation of messages signals was quite asynchronous to begin with, but had quite strict delivery constraints due to the ordering guarantees of signals between a pair of processes. The previously used message queue partitioned into two halves has been replaced by a more general signal queue partitioned into three parts that service all kinds of signals. More details regarding the signal queue can be found in comments in the erl_proc_sig_queue.h file. The monitor and link implementations have also been completely replaced in order to fit the new asynchronous signaling implementation as good as possible. More details regarding the new monitor and link implementations can be found in the erl_monitor_link.h file.
2018-03-21Merge pull request #1721 from josevalim/jv-no-messageanBjörn Gustavsson
Remove unused cerl_messagean module
2018-03-14Merge branch 'maint'Henrik
Conflicts: OTP_VERSION
2018-03-09Update release notesErlang/OTP
2018-03-09Update version numbersErlang/OTP
2018-02-23Sort entries and correct typosMichał Muskała
2018-02-19Remove unused cerl_messagean moduleJosé Valim
2018-02-17Inline more type test BIFs in HiPEMichał Muskała
2018-02-06dialyzer: Assign a type to the primop build_stacktraceHans Bolinder
2018-02-01Merge pull request #1701 from bjorng/bjorn/get_hd_tlBjörn Gustavsson
Eliminate get_list/3 internally in the compiler
2018-01-29make HiPE work on x86_64 when PIE is enabledbhuztez
Currently HiPE amd64 assumes the runtime system code is loaded into the low 2G of the address space. However, this is not the case when PIE is enabled, it is loaded into a random location. So trampolines are required to call BIFs, and also we have first to load the address of sse2_fnegate_mask to a regisiter before xorpd in fchs.
2018-01-29Merge PR-1685 from sverker/sverker/hipe-verify-gcsafe OTP-14900Sverker Eriksson
Make hipe compile option verify_gcsafe the default
2018-01-26Eliminate get_list/3 internally in the compilerBjörn Gustavsson
Instructions that produce more than one result complicate optimizations. get_list/3 is one of two instructions that produce multiple results (get_map_elements/3 is the other). Introduce the get_hd/2 and get_tl/2 instructions that return the head and tail of a cons cell, respectively, and use it internally in all optimization passes. For efficiency, we still want to use get_list/3 if both head and tail are used, so we will translate matching pairs of get_hd and get_tl back to get_list instructions.
2018-01-22Don't build a stacktrace if it's only passed to erlang:raise/3Björn Gustavsson
Consider the following function: function({function,Name,Arity,CLabel,Is0}, Lc0) -> try %% Optimize the code for the function. catch Class:Error:Stack -> io:format("Function: ~w/~w\n", [Name,Arity]), erlang:raise(Class, Error, Stack) end. The stacktrace is retrieved, but it is only used in the call to erlang:raise/3. There is no need to build a stacktrace in this function. We can avoid the building if we introduce an instruction called raw_raise/3 that works exactly like the erlang:raise/3 BIF except that its third argument must be a raw stacktrace.
2018-01-22Merge branch 'maint'Björn Gustavsson
* maint: ErLLVM: Preserve precise BEAM tailcall semantics observer: Fix change accum Remove double calls observer: Don't crash for late messages observer: Optimize tv tab for many tables
2018-01-19ErLLVM: Preserve precise BEAM tailcall semanticsMagnus Lång
The BEAM compiler chooses not to perform tailcall optimisations for some calls in tail position, for example to some built-in functions. However, when the ErLLVM HiPE backend is used, LLVM may choose to perform tailcall optimisation on these calls, breaking the expected semantics. To preserve the precise semantics exhibited by BEAM, the 'notail' marker, present in LLVM since version 3.8, is added to call instructions that BEAM has not turned into tail calls, which inhibits LLVM from performing tail-call optimisation in turn.
2018-01-19Merge branch 'maint'Hans Bolinder
* maint: dialyzer: Fix bsl/2 bug
2018-01-19Merge branch 'hasse/dialyzer/fix_bsl' into maintHans Bolinder
* hasse/dialyzer/fix_bsl: dialyzer: Fix bsl/2 bug
2018-01-16hipe: Make option 'verify_gcsafe' the defaultSverker Eriksson
2018-01-16hipe: Mark primop 'bs_put_utf8' as gcsafeSverker Eriksson
which is has been since 3d21f793538927ae88f78504a11dd898e8ca1a7a
2018-01-15Merge branch 'maint'Sverker Eriksson
2018-01-15Merge branch 'sverker/cuddle-big-bin-tests' into maintSverker Eriksson
2018-01-12dialyzer: Fix bsl/2 bugHans Bolinder
Also modified erl_bif_types:infinity_bsl() when called with zero as first argument. As of writing this, erlang:'bsl'/2 is modified on the master branch to never fail if called with a huge second argument.
2018-01-09Merge branch 'maint'Hans Bolinder
* maint: dialyzer: Correct handling of erlang:abs/1
2018-01-08dialyzer: Correct handling of erlang:abs/1Hans Bolinder
See also https://bugs.erlang.org/browse/ERL-551.