aboutsummaryrefslogtreecommitdiffstats
path: root/lib/compiler/src
AgeCommit message (Collapse)Author
2019-08-15Fix an internal consistency check failure caused by beam_exceptBjörn Gustavsson
https://bugs.erlang.org/browse/ERL-1026
2019-08-14Fix compiler crash when compiling some receive statementsBjörn Gustavsson
The compiler would crash when compiling the following code: do(Acc) -> receive {Pid, abc} -> ok; {Pid, []} -> ok; {Pid, _Res} -> exit(_Res) end, do([Pid | Acc]). The last clause that always raises an exception would confuse the compiler so that it would think that the `receive` statement was at the end of the function and it would generate incorrect code for the `do/1` call following the `receive`. https://bugs.erlang.org/browse/ERL-1022
2019-08-05Merge branch 'bjorn/compiler/fix-stack-init/ERL-1017/OTP-15968' into maintBjörn Gustavsson
* bjorn/compiler/fix-stack-init/ERL-1017/OTP-15968: Ensure that the stack slots are initialized when matching maps
2019-08-05Ensure that the stack slots are initialized when matching mapsBjörn Gustavsson
When matching a map, the compiler could fail to generate code that would initialize all stack slots (Y registers) properly. Here is a general outline of code that *could* cause this problem: foo(Key, Map) -> Res = case Map of #{Key := Val} -> %% Do something with Val here. . . . #{} -> [] end, %% The stack slot for Val might not have been initialized %% here if the key was not present in the map. . . . %% Use Res. . . . The code generator would wrongly assume that the map matching would always initialize the stack slot, and if nothing else happened to force that stack slot to be initialized, it would remain uninitialized, which would likely crash the runtime system at the next garbage collection. `beam_validator` is supposed to find these kind of problems, but a bug in `beam_validator` prevented it from detecting this problem. https://bugs.erlang.org/browse/ERL-1017
2019-08-05Merge branch 'john/compiler/fix-delayed-type-inference/OTP-15954/ERL-995' ↵John Högberg
into maint * john/compiler/fix-delayed-type-inference/OTP-15954/ERL-995: beam_validator: Values referenced by other values must be merged
2019-08-02Merge pull request #2336 from ↵Björn Gustavsson
bjorng/bjorn/compiler/fix-slow-beam_ssa_dead/ERL-1014/OTP-15966 Avoid extremely long compilation times for huge functions
2019-08-02Merge branch 'bjorn/compiler/length-misuse/ERL-1013' of ↵Björn Gustavsson
https://github.com/bjorng/otp into maint OTP-15970 * 'bjorn/compiler/length-misuse/ERL-1013' of https://github.com/bjorng/otp: Eliminate a crash in the type optimizer pass
2019-08-02Merge branch 'bjorn/compiler/fix-no_type_opt/ERL-997' of ↵Björn Gustavsson
https://github.com/bjorng/otp into maint OTP-15969 * 'bjorn/compiler/fix-no_type_opt/ERL-997' of https://github.com/bjorng/otp: Fix compiler crash when compiling with +no_type_opt
2019-08-01Avoid extremely long compilation times for huge functionsBjörn Gustavsson
Compiling this example takes less than a second for OTP 21: -define(B, {?A,?A,?A,?A,?A}). -define(C, {?B,?B,?B,?B,?B}). -define(D, {?C,?C,?C,?C,?C}). -define(E, {?D,?D,?D}). f() -> ?E = foo:bar(). The compilation time for OTP 22 is about 10 seconds. Most of the time is spent in `beam_ssa_dead`. This commit introduces several optimizations to bring the compilation time down to about a second. The most important of those optimizations is limiting the effort spent searching forward for a joining point for the success and failure labels for a two-way branch. This change is helped by the change of representation of variable sets from `ordsets` to `cerl_sets`. https://bugs.erlang.org/browse/ERL-1014
2019-08-01Merge branch 'bjorn/compiler/fix-unsafe-sharing/OTP-15963' into maintBjörn Gustavsson
* bjorn/compiler/fix-unsafe-sharing/OTP-15963: Fix unsafe code sharing
2019-07-30Fix compiler crash when compiling with +no_type_optBjörn Gustavsson
If the `no_type_opt` option was given, the compiler would crash when attempting to compile containing with a `try`...`after` construct, such as this code: foo() -> try make_ref() after ok end. To avoid having this bug re-appear, test the `no_type_opt` option in the test suites. https://bugs.erlang.org/browse/ERL-997
2019-07-30Eliminate a crash in the type optimizer passBjörn Gustavsson
https://bugs.erlang.org/browse/ERL-1013
2019-07-29Fix unsafe code sharingBjörn Gustavsson
2019-07-11beam_validator: Values referenced by other values must be mergedJohn Högberg
This is a more proper fix for ERIERL-348. We used to think that we wouldn't need to update the type of a variable that's no longer referenced by a register ("dead value"), but the attached test case pokes a hole in that assumption. To summarize, the result of '=:='/2 is kept alive longer than one of its arguments, which gets pruned in a state merge leaving us with nothing to work on when we finally compare the result. This is fine for most operations since there's no point in (say) updating the size of a tuple we can no longer reach, but '=:='/2 updates the types of both arguments and we risk missing out on important information when either of them is gone. This commit fixes the problem by merging all values that are *reachable* from a register, rather than just those that *exist* in a register, ensuring that all values stay around at least as long as they're needed.
2019-07-11Merge branch 'john/compiler/fix-bad-try_catch-recv-fix/OTP-15953/ERL-999' ↵John Högberg
into maint * john/compiler/fix-bad-try_catch-recv-fix/OTP-15953/ERL-999: compiler: Fix compiler crash introduced by OTP-15952
2019-07-10compiler: Fix compiler crash introduced by OTP-15952John Högberg
An assertion in code generation would fail when the common exit block was ?BADARG_BLOCK, as some operations expect to always "fail" directly to that block (= throw an exception) and we had inserted a dummy block in between. Other operations could also get funny fail labels, jumping to blocks that immediately jumped to {f,0}, but these were all cleaned up by beam_jump, sweeping the bug under the rug.
2019-07-09Merge branch 'john/compiler/fix-fail-path-exceptions-bsm/OTP-15946' into ↵Erlang/OTP
maint-22 * john/compiler/fix-fail-path-exceptions-bsm/OTP-15946: beam_ssa_bsm: Leave ?BADARG_BLOCK alone when cloning fail path
2019-07-09Merge branch 'john/compiler/fix-unsafe-tuple_size-opt/OTP-15945' into maint-22Erlang/OTP
* john/compiler/fix-unsafe-tuple_size-opt/OTP-15945: beam_ssa_opt: Do not apply tuple_size optimization outside guards
2019-07-09Merge branch 'john/compiler/fix-try_catch-receives/OTP-15952' into maintJohn Högberg
* john/compiler/fix-try_catch-receives/OTP-15952: compiler: Fix broken 'receive' in try/catch blocks
2019-07-09compiler: Fix broken 'receive' in try/catch blocksJohn Högberg
This fix is rather ugly and tacked-on, but I'm not comfortable refactoring the pass in an emergency patch.
2019-07-04Merge branch 'john/compiler/fix-fail-path-exceptions-bsm/OTP-15946' into maintJohn Högberg
* john/compiler/fix-fail-path-exceptions-bsm/OTP-15946: beam_ssa_bsm: Leave ?BADARG_BLOCK alone when cloning fail path
2019-07-04Merge branch 'john/compiler/fix-unsafe-tuple_size-opt/OTP-15945' into maintJohn Högberg
* john/compiler/fix-unsafe-tuple_size-opt/OTP-15945: beam_ssa_opt: Do not apply tuple_size optimization outside guards
2019-07-03beam_ssa_bsm: Leave ?BADARG_BLOCK alone when cloning fail pathJohn Högberg
2019-07-03beam_ssa_opt: Do not apply tuple_size optimization outside guardsJohn Högberg
Rewriting `tuple_size` to `is_tuple` + `tuple_size` will cause it not to throw an exception, either crashing the compiler or the emulator when the code runs.
2019-06-26Fix slow compilation of huge functionsBjörn Gustavsson
Some huge functions would compile very slowly because of a bottleneck in `beam_ssa:def_used/2`. One example is the `cuter_binlib` module in https://github.com/cuter-testing/cuter. On my computer, this commit reduces the compilatation time for `cuter_binlib` to 45 seconds down from more than 4 minutes. Noticed-by: Kostis Sagonas
2019-06-24Eliminate dialyzer warningsBjörn Gustavsson
Eliminate the Dialyzer warnings shown when the limits in lib/cerl/erl_types.erl were raised as follows: -define(TUPLE_TAG_LIMIT, 10). -define(TUPLE_ARITY_LIMIT, 10). -define(SET_LIMIT, 64).
2019-06-11Merge branch 'john/compiler/fix-bad-bitstring-type-opt/OTP-15872' into maintJohn Högberg
* john/compiler/fix-bad-bitstring-type-opt/OTP-15872: beam_ssa_type: Fix incorrect bitstring unit determination
2019-06-10beam_ssa_type: Fix incorrect bitstring unit determinationJohn Högberg
The compiler would treat the "Unit" of bs_init instructions as the unit of the result instead of the required unit of the input, causing is_binary checks to be wrongly optimized away.
2019-05-29Merge branch 'bjorn/compiler/fix-beam_ssa_dead-patch/OTP-15845' into maint-22Erlang/OTP
* bjorn/compiler/fix-beam_ssa_dead-patch/OTP-15845: Fix unsafe optimizations where guard tests could be removed
2019-05-29Merge branch ↵Erlang/OTP
'bjorng_ghub/bjorn/compiler/fix-beam_ssa_dead-crash/ERL-956/OTP-15848' into maint-22 * bjorng_ghub/bjorn/compiler/fix-beam_ssa_dead-crash/ERL-956/OTP-15848: Eliminate crash in the beam_ssa_dead compiler pass
2019-05-29Merge branch 'bjorn/compiler/fix-unloadable-code-patch/ERL-955/OTP-15846' ↵Erlang/OTP
into maint-22 * bjorn/compiler/fix-unloadable-code-patch/ERL-955/OTP-15846: Fix loading of Core Erlang code for extracting a map element
2019-05-29Merge branch 'bjorn/compiler/fix-beam_except/ERL-954/OTP-15839' into maint-22Erlang/OTP
* bjorn/compiler/fix-beam_except/ERL-954/OTP-15839: Fix compiler crash in beam_except
2019-05-29Merge branch 'bjorn/compiler/fix-unsafe-type-inference/OTP-15838' into maint-22Erlang/OTP
* bjorn/compiler/fix-unsafe-type-inference/OTP-15838: Fix unsafe negative type inference # Conflicts: # lib/compiler/src/beam_ssa_type.erl
2019-05-29Merge branch 'john/compiler/list_append_type/OTP-15841' into maint-22Erlang/OTP
* john/compiler/list_append_type/OTP-15841: compiler: Fix broken type for erlang:'++'/2
2019-05-29Merge branch 'bjorn/compiler/fix-receive-patch/ERL-950/OTP-15832' into maint-22Erlang/OTP
* bjorn/compiler/fix-receive-patch/ERL-950/OTP-15832: Eliminate compiler crash when compiling complex receive statements
2019-05-29Merge pull request #2263 from ↵Björn Gustavsson
bjorng/bjorn/compiler/fix-beam_ssa_dead-crash/ERL-956/OTP-15848 Eliminate crash in the beam_ssa_dead compiler pass
2019-05-28Merge branch 'john/compiler/list_append_type/OTP-15841' into maintJohn Högberg
2019-05-28Fix unsafe optimizations where guard tests could be removedBjörn Gustavsson
A repeated test could be optimized away. Example: bar(A) -> if is_bitstring(A) -> if is_binary(A) -> binary; true -> bitstring end; true -> other end. In the example, the `is_binary/1` test would be optimized away, basically turning the example into: bar(A) -> if is_bitstring(A) -> bitstring; true -> other end. Thanks user Marcus Kruse in the Elixir forum for noticing this bug.
2019-05-28Eliminate crash in the beam_ssa_dead compiler passBjörn Gustavsson
The compiler could crash in the beam_ssa_dead pass while compiling complex nested `case` expressions. See the added test case for an example and explanation. https://bugs.erlang.org/browse/ERL-956
2019-05-27Merge branch 'bjorn/compiler/fix-unloadable-code-patch/ERL-955/OTP-15846' ↵Björn Gustavsson
into maint * bjorn/compiler/fix-unloadable-code-patch/ERL-955/OTP-15846: Fix loading of Core Erlang code for extracting a map element
2019-05-27Merge branch 'bjorn/compiler/fix-beam_ssa_dead-patch/OTP-15845' into maintBjörn Gustavsson
* bjorn/compiler/fix-beam_ssa_dead-patch/OTP-15845: Fix unsafe optimizations where guard tests could be removed
2019-05-27Merge branch 'bjorn/compiler/fix-beam_except/ERL-954/OTP-15839' into maintBjörn Gustavsson
* bjorn/compiler/fix-beam_except/ERL-954/OTP-15839: Fix compiler crash in beam_except
2019-05-27Fix compiler crash in beam_exceptBjörn Gustavsson
The compiler would crash in `beam_except` while compiling this function: bar(Req) -> ok = case Req of "POST" -> {error, <<"BAD METHOD ", Req/binary>>, Req}; _ -> ok end. https://bugs.erlang.org/browse/ERL-954
2019-05-27Merge branch 'bjorn/compiler/fix-unsafe-type-inference/OTP-15838' into maintBjörn Gustavsson
* bjorn/compiler/fix-unsafe-type-inference/OTP-15838: Fix unsafe negative type inference
2019-05-27compiler: Fix broken type for erlang:'++'/2John Högberg
2019-05-27Fix loading of Core Erlang code for extracting a map elementBjörn Gustavsson
The following Core Erlang code could not be loaded: 'f'/1 = fun (_1) -> case <_1> of <~{'foo':='foo'}~> when 'true' -> _1 end Loading would fail with the following message: beam/beam_load.c(2314): Error loading function example:f/1: op i_get_map_element_hash p x a u x: no specific operation found https://bugs.erlang.org/browse/ERL-955
2019-05-25Fix unsafe optimizations where guard tests could be removedBjörn Gustavsson
A repeated test could be optimized away. Example: bar(A) -> if is_bitstring(A) -> if is_binary(A) -> binary; true -> bitstring end; true -> other end. In the example, the `is_binary/1` test would be optimized away, basically turning the example into: bar(A) -> if is_bitstring(A) -> bitstring; true -> other end. Thanks user Marcus Kruse in the Elixir forum for noticing this bug.
2019-05-24Fix unsafe negative type inferenceBjörn Gustavsson
The type optimizer pass (`beam_ssa_type`) could make unsafe negative inferences. That is, incorrectly infer that a variable could *not* have a particular type. This bug was found when adding another optimization. It is not clear how write a failing test case without that added optimization.
2019-05-23Merge branch 'bjorn/compiler/fix-receive-patch/ERL-950/OTP-15832' into maintBjörn Gustavsson
* bjorn/compiler/fix-receive-patch/ERL-950/OTP-15832: Eliminate compiler crash when compiling complex receive statements
2019-05-22Merge branch 'bjorn/compiler/fix-freeze/ERL-948/OTP-15828' into maintBjörn Gustavsson
* bjorn/compiler/fix-freeze/ERL-948/OTP-15828: Fix non-terminating compilation