aboutsummaryrefslogtreecommitdiffstats
path: root/lib/hipe
AgeCommit message (Collapse)Author
2016-02-25Merge branch 'kostis/hipe-icode-cleanup' into maintHenrik Nord
* kostis/hipe-icode-cleanup: Fix dialyzer warning and some code refactoring OTP-13379
2016-02-09dialyzer: Correct byte_size() and comparisonsHans Bolinder
The argument of byte_size() is a bitstring(). The code in erl_bif_types that finds cases where comparisons always return true or false is corrected when it comes to maps and bit strings.
2016-02-07Fix dialyzer warning and some code refactoringKostis Sagonas
A previous commit introduced a change that exposed dead code in this module and caused dialyzer warnings. This dead code was taken out. While at it, performed some code refactoring in the handling of the bs_get_integer primop.
2016-02-02Merge branch 'margnus1/bs_unit_fix' into maintZandra
* margnus1/bs_unit_fix: hipe: Fix signed compares of unsigned sizes beam: Fix overflow bug in i_bs_add_jId hipe: Add tests for bad bit syntax float sizes Add a case testing the handling of guards involving binaries Add some more binary syntax construction tests hipe: Guard against enormous numbers in ranges hipe: Fix constructing huge binaries hipe: Fix binary constructions failing with badarith Add missing corner-case to bs_construct_SUITE hipe: Allow unsigned args in hipe_rtl_arith hipe: test unit size match in bs_put_binary_all hipe: test unit size match in bs_append Fix hipe_rtl_binary_construct:floorlog2/1 OTP-13272
2016-01-27Merge branch 'kostis/hipe-tests-basic' into maintZandra
OTP-13269 * kostis/hipe-tests-basic: Fix compilation of matching with UTF binaries Cleanup and add one more test case Two tests that depend on inlining being turned on More tests for BIFs Include some more old HiPE tests to the test suite Add tests for the is_boolean/1 guard Two more tests added Test that apply/3 is tail recursive Three more tests added Minor cleanup Comment out tests that are not ready for to_llvm Cleanups & uncomment some code More tests for handling of UTF in bitstrings Minor code cleanup Add more generated test suites in Makefile Use function from hipe module instead of a local one Add function to prevent running tests in the LLVM backend More basic tests First part of the basic test suite for the HiPE compiler
2016-01-18Merge branch 'egil/improve-map-cerl-prettypr/OTP-13238' into maintBjörn-Egil Dahlberg
* egil/improve-map-cerl-prettypr/OTP-13238: compiler, hipe: Fix pretty printing of Core Maps hipe: Fix map pretty printing of pairs dialyzer: Update Maps tests
2016-01-15dialyzer: Correct handling of parameters of opaque typesHans Bolinder
Correction of commit d57f5e.
2016-01-14compiler, hipe: Fix pretty printing of Core MapsBjörn-Egil Dahlberg
Literal maps could cause dialyzer to crash when pretty printing the results. Reported-by: Chris McGrath <[email protected]>
2016-01-14hipe: Fix map pretty printing of pairsBjörn-Egil Dahlberg
In commit f667931e2905797ffab63e224e56eaf07f77178a the core format changed for map pairs. Let dialyzer and hipe pretty printing of maps also adhere to those changes. An Erlang map update, M#{foo := 1, bar => 2} will now be printed as: ~{ 'foo' := 1, 'bar' => 2 | M }~
2015-12-16Fix compilation of matching with UTF binariesKostis Sagonas
The code generated by the HiPE compiler for pattern matching with UTF binaries was such that sometimes THE_NON_VALUE was stored in the roots followed by the garbage collector. This was not an issue for the vanilla native code compiler, but was problematic for the ErLLVM back-end. Fix the issue by not storing THE_NON_VALUE in the live roots. An alternative fix would be to change the code of the garbage collector. With this fix, there are no more (known) failing test cases for the ErLLVM back-end (at least on x86_64 with LLVM 3.5, which is the configuation regularly tested). Thanks to @margnus1 for the fix.
2015-12-16Cleanup and add one more test caseKostis Sagonas
2015-12-16Two tests that depend on inlining being turned onKostis Sagonas
2015-12-16More tests for BIFsKostis Sagonas
2015-12-16Include some more old HiPE tests to the test suiteKostis Sagonas
2015-12-16Add tests for the is_boolean/1 guardKostis Sagonas
2015-12-16Two more tests addedKostis Sagonas
- A test for proper handling of negative numbers in binary search key tables - A test for HiPE's ICode range analysis
2015-12-16Test that apply/3 is tail recursiveKostis Sagonas
2015-12-16Three more tests addedKostis Sagonas
2015-12-16Minor cleanupKostis Sagonas
Take out extraneous spaces at the end of lines.
2015-12-16Comment out tests that are not ready for to_llvmKostis Sagonas
2015-12-16Cleanups & uncomment some codeKostis Sagonas
2015-12-16More tests for handling of UTF in bitstringsKostis Sagonas
2015-12-16Minor code cleanupKostis Sagonas
2015-12-16Add more generated test suites in MakefileKostis Sagonas
2015-12-16Use function from hipe module instead of a local oneKostis Sagonas
2015-12-16Add function to prevent running tests in the LLVM backendKostis Sagonas
2015-12-16More basic testsKostis Sagonas
and some tests that the HiPE compiler is not causing trouble
2015-12-16First part of the basic test suite for the HiPE compilerKostis Sagonas
2015-12-15Update release notesErlang/OTP
2015-12-15Update version numbersErlang/OTP
2015-12-02hipe: Fix signed compares of unsigned sizesMagnus Lång
Also, some of the branches were testing sizes in bits against a constant ?MAX_BINSIZE, which was in bytes. The signed comparisons masked this mistake. These branches have been removed since all sizes in bits that fit in a machine word are valid binary sizes. Finally, a test that reproduces the issue was added to bs_construct, along with a test for one of the cases (bs_init<0>(...)) when the test against ?MAX_BINSIZE must be changed to unsigned rather than removed.
2015-11-30hipe: Add tests for bad bit syntax float sizesMagnus Lång
2015-11-27Add a case testing the handling of guards involving binariesKostis Sagonas
The HiPE compiler was leaking exceptions out of guards involving binaries with `strange' arithmetic in them. Fixed by the set of changes in this pull request.
2015-11-27Add some more binary syntax construction testsKostis Sagonas
Some more cases (taken from the BEAM bs_construct suite) that used to show different behaviour than BEAM when compiled to native code were included in the appropriate test file of the HiPE test suite.
2015-11-27hipe: Guard against enormous numbers in rangesMagnus Lång
This allows us to compile bs_match_int_SUITE with HiPE without a system_limit crash.
2015-11-27hipe: Fix constructing huge binariesMagnus Lång
Bugs were fixed in hipe_rtl_binary_match:{first_part/3,make_size/3,set_high/1} in commit 5aea81c49, but it turns out these had been copy-pasted verbatim into hipe_rtl_binary_construct, where they were causing further bugs. They have now moved to hipe_rtl_binary, from where they are included by the other two modules. Furthermore, first_part/3 (reamed get_word_integer/3, since it loads integers that fits into an unsigned word), and make_size/3 now accepts a fourth argument to distinguish too large arguments (which should cause a system_limit exception) from negative or non-integral arguments. The use of first_part/3 (get_word_integer/3) from 5aea81c49 in hipe_rtl_binary_construct now allows several binary construction instructions to accept bignum sizes, as they were supposed to. Additionally, calls to hipe_rtl_binary_construct:check_and_untag_fixnum/3 were replaced with get_word_integer/4 since all of them were also supposed to accept sufficiently small bignums, but didn't, and check_and_untag_fixnum/3 was essentially identical to first_part/3 anyway. HiPE is now capable of passing bs_construct_SUITE completely unmodified.
2015-11-27hipe: Fix binary constructions failing with badarithMagnus Lång
Expressions like <<0:(fun(X)->X end(anka))>> would fail with 'badarith' in HiPE, but 'badarg' in BEAM. hipe_beam_to_icode was attempting to optimise the amount of code generated by letting the arithmetic error propagate instead of catching it. However, since it was emitting a block to throw 'badarg' in that case anyway, we can just reuse it to achieve just as compact code that behaves exactly like BEAM.
2015-11-27hipe: Allow unsigned args in hipe_rtl_arithMagnus Lång
hipe_rtl_arith is only used by hipe_rtl_ssa_const_prop, which applies it to any RTL, including RTL where the intent is to do unsigned math. Since signed and unsigned operations produce the same 2's complement result, this change is harmless. On 32-bit architectures it caused HiPE crashes when compiling code like <<0:((1 bsl 32)-1)>>, because the size of the field is too large to fit in a signed integer.
2015-11-27hipe: test unit size match in bs_put_binary_allMagnus Lång
The unit size field was previously completely discarded when lowering this instruction from BEAM to Icode. This feature was previously missing and expressions such as <<0, <<1:1>>/binary>> would succeed construction when compiled with HiPE.
2015-11-27hipe: test unit size match in bs_appendMagnus Lång
This feature was previously missing and expressions such as <<<<1:1>>/binary>> would succeed construction when compiled with HiPE. A primop is_divisible is introduced to handle the case when the unit size is not a power of two.
2015-11-27Fix hipe_rtl_binary_construct:floorlog2/1Magnus Lång
Relying on double-precision floating-point arithmetic to compute the log2 of an integer up to 64 bits long leads to rounding errors.
2015-11-09Merge branch 'kostis/hipe-bs-match-huge-bin' into maintHenrik Nord
* kostis/hipe-bs-match-huge-bin: Fix matching with huge binaries Compile without errors for exported variables OTP-13092
2015-11-04[hipe] Correct documentationHans Bolinder
Fix mistakes found by 'xmllint'.
2015-10-27Merge branch 'lucafavatella/dialyzer-fun-literal-arity' into maintHenrik Nord
* lucafavatella/dialyzer-fun-literal-arity: Teach Dialyzer arity of funs with literal arity OTP-13068
2015-10-12Fix edge case of Size = 0 in bs_put_integerKostis Sagonas
copy_offset_int_big was assuming (Offset + Size - 1) (Tmp9 in the first BB) would not underflow. It was also unconditionally reading and writing the binary even when Size was zero, unlike copy_int_little, which is the only other case of bs_put_integer that does not have a short-circuit on Size = 0. This was causing segfaults when constructing binaries starting with a zero-length integer field, because a logical right shift was used to compute an offset in bytes (which became 0x1fffffffffffffff) to read in the binary. Tests, taken from the emulator bs_construct_SUITE, were also added. The complete credit for the report and the fix goes to Magnus Lång.
2015-10-02Fix matching with huge binariesKostis Sagonas
In certain cases of matching with very big binaries, the HiPE compiler generated code that would fail the match, even in cases that the matching was successful. The problem was more quite noticeable on 32-bit platforms where certain integer quantities would be represented as bignums. Brief summary of changes: * gen_rtl({bs_skip_bits, ...}, ...) could not handle too large constants. Previously the constants were truncated to word size. * hipe_rtl_binary_match:make_size/3 erroneously assumed that the output of first_part/3 would not overflow when multiplied by 8, which is no longer true. To maintain full performance, the overflow test is only performed when BitsVar was a bignum. Thus, the fast path is identical to before. * hipe_rtl_binary_match:set_high/2 was assuming that only bits below bit 27 were ever set in arguments to bs_skip_bits, which is not only false when the arguments are bignums, but also on 64-bit platforms. The commit includes a test taken from the bs_match_bin_SUITE. Most of the credit for finding these HiPE compiler errors and for creating appropriate fixes for them should go to Magnus Lång.
2015-10-02Compile without errors for exported variablesKostis Sagonas
2015-09-21Prepare releaseErlang/OTP
2015-09-16Update application versionsHans Bolinder
2015-09-09dialyzer: Correct the handling of parameters of opaque typesHans Bolinder
Prior to this commit, the fact that parameters of opaque types are expanded differently depending on the current values of limits used during expansion, caused problems later when the types of parameters are used for determining if opaque types are comparable.