aboutsummaryrefslogtreecommitdiffstats
path: root/erts/emulator/hipe
AgeCommit message (Collapse)Author
2015-08-31Merge branch 'sverk/hipe-fix-literal-crc' into maintSverker Eriksson
* sverk/hipe-fix-literal-crc: erts,hipe,dialyzer: Fix hipe checkum of target runtime system erts: Change THE_NON_VALUE to not be hard coded in hipe compiler OTP-12962 OTP-12963 OTP-12964
2015-08-31erts,hipe,dialyzer: Fix hipe checkum of target runtime systemSverker Eriksson
Main problem: A faulty HIPE_LITERAL_CRC was not detected by the loader. Strangeness #1: Dialyzer should ask the hipe compiler about the target checksum, not an internal bif. Strangeness #2: The HIPE_SYSTEM_CRC checksum was based on the HIPE_LITERALS_CRC checksum. Solution: New HIPE_ERTS_CHECKSUM which is an bxor of the two (now independent) HIPE_LITERALS_CRC and HIPE_SYSTEM_CRC. HIPE_LITERALS_CRC represents values that are assumed to stay constant for different VM configurations of the same arch, and are therefor hard coded into the hipe compiler. HIPE_SYSTEM_CRC represents values that may differ between VM variants. By default the hipe compiler asks the running VM for this checksum, in order to create beam files for the same running VM. The hipe compiler can be configured (with "make XCOMP=yes ...") to create beam files for another VM variant, in which case HIPE_SYSTEM_CRC is also hard coded. ToDo: Treat all erts properties the same. Either ask the running VM or hard coded into hipe (if XCOMP=yes). This will simplify and reduce the risk of dangerous mismatches. One concern might be the added overhead from more frequent calls to hipe_bifs:get_rts_param.
2015-08-28erts: Fix hipe bug for maps:merge/2Sverker Eriksson
Add forgotten HIPE_WRAPPER_BIF_DISABLE_GC which could lead to stack-heap overrun if unlucky with the yielding during maps:merge when called by native hipe code.
2015-08-28erts: Beautify hipe wrapper macroSverker Eriksson
2015-08-24erts: Change THE_NON_VALUE to not be hard coded in hipe compilerSverker Eriksson
Instead ask running VM for the value of THE_NON_VALUE, which is different between opt and debug VM. Same hipe compiler can now compile for both opt and debug VM.
2015-06-18Change license text to APLv2Bruce Yinhe
2015-05-27Change hipe_bifs:system_crc/1 to hipe_bifs:system_crc/0Magnus Henoch
The macro HIPE_SYSTEM_CRC used to contain a hidden cookie from the VM that generated hipe_literals.hrl. This means that BEAM files containing that macro would be tied to a particular version of the VM. Change hipe_bifs:system_crc such that it doesn't require a hidden cookie to return the desired value.
2015-05-08Merge branch 'rickard/timer-optimization/OTP-12650'Rickard Green
* rickard/timer-optimization/OTP-12650: Optimized timer implementation Reusable red-black tree implementation Conflicts: erts/emulator/beam/erl_bif_timer.c
2015-05-08Optimized timer implementationRickard Green
2015-05-06Merge branch 'sverk/hipe_match_wbin/OTP-12667'Sverker Eriksson
* sverk/hipe_match_wbin/OTP-12667: erts: Add debug assertions for match state sanity hipe: Add test for matching of writable binary erts,hipe: Optimize away calls to emasculate_binary erts,hipe: Fix bug in binary matching of writable binary Conflicts: erts/emulator/hipe/hipe_bif0.c
2015-04-22erts/hipe: unbreak arity 4 BIFsMikael Pettersson
This fixes arity 4 BIF support in HiPE, following its introduction on master (OTP 18) via the nox/ets-update_counter-4 merge. - define standard_bif_interface_4, nbif_4_gc_after_bif, and nbif_4_simple_exception on ARM: unbreaks the build on ARM - remove bogus stack re-alignment from standard_bif_interface_4 on AMD64: for 4-arg BIFs the stack is already aligned, and the code would misalign the C stack which violates the ABI and may cause alignment faults in vectorized code - the nbif_4_simple_exception OPD name on PPC64 was incorrectly using the nbif_3_simple_exception OPD name: this would have caused a multiple definition error in the assembler or linker In addition there are a few cleanups: - fix standard_bif_interface_N comment on x86 - fix standard_bif_interface_4 comment on SPARC - separate nbif_N_simple_exception blocks by empty lines on PPC, like on ARM, to clearly show which things belong together - fix standard_bif_interface_N comment on ARM - fix standard_bif_interface_4 on AMD64 to match the indentation and spacing conventions of the rest of that file
2015-04-21erts,hipe: Optimize away calls to emasculate_binarySverker Eriksson
Only call emasculate_binary if ProcBin.flags is set, which means it's a writable binary.
2015-04-20erts,hipe: Fix bug in binary matching of writable binarySverker Eriksson
Seen symptom: Hipe compiled code with <<C/utf8, ...>> = Bin does sometimes not match even though Bin contains a valid utf8 character. There might be other possible binary matching symptoms, as the problem is not utf8 specific. Problem: A writable binary was not "emasculated" when the matching started (as it should) by the hipe compiled code. Fix: Add a new primop emasculate_binary(Bin) that is called when a matchstate is created. ToDo: There are probably room for optimization. For example only call emasculate_binary if ProcBin.flags is set.
2015-03-16Merge branch 'nox/ets-update_counter-4'Zandra Hird
* nox/ets-update_counter-4: Create new BIF ets:update_counter/4 Allow 4-ary BIFs OTP-12563
2015-03-12Merge branch 'maint'Henrik Nord
Conflicts: erts/emulator/hipe/hipe_bif0.c
2015-03-12Create new BIF ets:update_counter/4Anthony Ramine
Conflicts: erts/emulator/beam/bif.tab lib/stdlib/src/ets.erl
2015-02-26hipe: change mfa_info_table lock to rwmutexMikael Pettersson
The hipe_mfa_info_table lock is currently a mutex, causing unnecessary contention and serialization of lookups for apply-like constructs. - change the lock from a mutex to an rwmutex - split hipe_get_na_nofail_locked into a "try" path which only needs a read lock, and a "slow" path which requires a write lock - reimplement hipe_get_na_nofail (used by apply ops etc) to first call the "try" path with a read lock, and if that fails the "slow" path with a write lock - reimplement hipe_get_na_nofail_locked (used by maintenance code) to call the "try" path and then optionally the "slow" path, but without taking locks since its caller already has the write lock - adjust remaining lock ops to take/release full write lock - use _rwlocked as suffix on functions requiring a write lock - change hipe_mfa_get_trampoline to call get_locked not put_locked, allowing it to take a read lock instead of a write lock - change hipe_find_mfa_from_ra() to only take a read lock
2015-02-24Allow 4-ary BIFsAnthony Ramine
2015-02-19remove perfctr supportMikael Pettersson
Perfctr is a Linux kernel extension that allows programmatic access to the performance monitoring counters found in most current CPUs. However, development of perfctr ceased after 2010, and it cannot be used with Linux kernels newer than 2.6.32. Therefore the perfctr support code in the Erlang VM is effectively dead code, so this patch removes it.
2015-01-26hipe: improve error handling at code allocation failureMikael Pettersson
2015-01-26hipe: remove two obsolete BIFsMikael Pettersson
The hipe_bifs:make_native_stub/2 and hipe_bifs:get_emu_address/1 BIFs were originally used by hipe_unified_loader.erl, but the code been obsolete and disabled for ages. Remove the BIFs and all references to them. In hipe_unified_loader.erl, remove the no-op emu_make_stubs/1 function.
2015-01-26hipe: remove HIPE_ALLOC_CODE macroMikael Pettersson
The HIPE_ALLOC_CODE macro in the HiPE runtime was introduced ages ago to allow x86 and amd64 to switch from erts_alloc() to an mmap() implementation with proper flag setting. Nowadays the macro is identical on all platforms, and serves no purpose. Delete the macro, move the hipe_alloc_code() prototype to hipe_arch.h, and simplify hipe_bifs_enter_code_2().
2015-01-15Merge branch 'maint'Bruce Yinhe
2015-01-11hipe: fix ARM/Thumb interworkingMikael Pettersson
HiPE on ARM is currently severely broken if the rest of the VM is compiled to run in Thumb mode -- calling native code quickly ends up executing code in the wrong mode and crashing the VM. This is a problem on e.g. Ubuntu which configures its system GCC to generate Thumb by default. It can also be triggered by overriding CC or CFLAGS when compiling the VM. There were three issues that caused the breakage: 1. Assembly-coded functions in hipe_arm_glue.S weren't explicitly tagged as functions, preventing the linker from generating the correct mode-switching call instructions for calls from C to these functions. Fixed by tagging those symbols as functions. 2. A few BIF wrappers were so simple that they performed tailcalls to the C BIFs. This fails to switch mode when C is in Thumb. Fixed by performing ordinary recursive calls when C is in Thumb. 3. The assembly-coded source files weren't explicitly tagged as ARM. Tested with the HiPE testsuite on ARMv7, with the VM built as ARM and as Thumb. Also manually inspected the object code for the beam executable and checked that call sites from C to HiPE's ARM runtime code and vice versa used the correct mode-switching instructions.
2014-11-03Merge branch 'sverk/yielding-distr-send/OTP-12232'Sverker Eriksson
* sverk/yielding-distr-send/OTP-12232: erts: Add constant TERM_TO_BINARY_MEMCPY_FACTOR erts: Optimize some repeated calls to {E,W}STACK_PUSH erts: Yield in term_to_binary when encoding big maps erts: Remove unnecessary goto for fun encoding erts: Yield in term_to_binary while copying large binaries erts: Implement yielding for distributed send of large messages
2014-10-28Merge branch 'maint'Lukas Larsson
* maint: Use isfinite() instead of finite() when available
2014-10-28Merge branch 'sv/isfinite/OTP-12268' into maintLukas Larsson
* sv/isfinite/OTP-12268: Use isfinite() instead of finite() when available
2014-10-28Use isfinite() instead of finite() when availableAnthony Ramine
OS X Mavericks builds result in a number of warnings about finite() being deprecated, like these: beam/erl_arith.c:451:7: warning: 'finite' is deprecated: first deprecated in OS X 10.9 [-Wdeprecated-declarations] ERTS_FP_ERROR(p, f1.fd, goto badarith); ^ sys/unix/erl_unix_sys.h:319:33: note: expanded from macro 'ERTS_FP_ERROR' ^ sys/unix/erl_unix_sys.h:244:51: note: expanded from macro '__ERTS_FP_ERROR' ^ /usr/include/math.h:718:12: note: 'finite' has been explicitly marked deprecated here extern int finite(double) __OSX_AVAILABLE_BUT_DEPRECATED(__MAC_10_0, __MAC_10_9, __IPHONE_NA, __IPHONE_NA); Add checks to use isfinite() instead of finite() where available. Verified on OS X Mavericks 10.9.5 and Ubuntu 12.04.
2014-10-24Merge branch 'sverk/hipe-inline-reserve-trap-frame'Sverker Eriksson
* sverk/hipe-inline-reserve-trap-frame: erts: Extend usage of ASM macro to avoid including asm macros in C code erts: Make hipe_{un}reserve_beam_trap_frame INLINE
2014-10-24erts: Extend usage of ASM macro to avoid including asm macros in C codeSverker Eriksson
and reduce the probability of macro name collisions Catalyst: ppc macro "r0" conflicting with local variable in external.c
2014-10-24erts: Make hipe_{un}reserve_beam_trap_frame INLINESverker Eriksson
2014-10-24Merge branch 'maint'Sverker Eriksson
2014-10-21erts: Fix bug when hipe tailcalls trapping BIF that disables GCSverker Eriksson
Symptom: base64_SUITE:roundtrip crashes with hipe compiled stdlib. Problem: HIPE_WRAPPER_BIF_DISABLE_GC pushed a "trap frame", but the frame was only popped if the call was recursive. Solution: Only reserve "trap frame" if BIF call is recursive.
2014-10-21erts: Fix hipe bug when gc-disabling bif traps with gc enabledSverker Eriksson
The trap_frame got pushed twice, first by the wrapper then by hipe_push_beam_trap_frame as it was looking at F_DISABLE_GC.
2014-09-22Merge remote-tracking branch 'upstream/maint'Ingela Anderton Andin
2014-09-17Merge branch 'sverk/bin-alloc-refactor'Sverker Eriksson
* sverk/bin-alloc-refactor: erts: Fix gdb command etp-carrier-blocks for 32-bit erts: Refactor binary allocation interface to also initialize Binary
2014-09-04erts: Refactor binary allocation interface to also initialize BinarySverker Eriksson
except the reference counter 'refc', as different callers have different strategies regarding the lifetime of the binary.
2014-08-29erts: Implement yielding for distributed send of large messagesSverker Eriksson
Use same mechanism as term_to_binary to yield while encoding large messages for distributed send.
2014-08-16Fix HiPE debug lock checking on OS X 64bitAnthony Ramine
Position-independent code is mandatory on OS X. We use r11 as an intermediate register to fill BIF_P->hipe.bif_callee. This fixes the following error when doing `make debug FLAVOR=smp`: clang -cc1as: fatal error: error in backend: 32-bit absolute addressing is not supported in 64-bit mode
2014-06-06Merge branch 'maint'Rickard Green
2014-06-05Replace erlang:binary_to_term() Erlang wrappersRickard Green
Replace the 'erlang:binary_to_term/1' and 'erlang:binary_to_term/2' Erlang wrappers taking care of failure after yield with management of this in the hidden yield BIF.
2014-05-22Make binary BIFs converting from lists yield on large inputRickard Green
- erlang:list_to_binary/1 - erlang:iolist_to_binary/1 - erlang:list_to_bitstring/1 - binary:list_to_bin/1
2014-05-22Make binary BIFs converting to lists yield on large inputRickard Green
- erlang:binary_to_list/1 - erlang:binary_to_list/3 - erlang:bitstring_to_list/1
2014-04-29erts: Remove some dead hipe specific codeSverker Eriksson
2014-04-29erts: Fix global tracing of beam function when called from hipe codeSverker Eriksson
Make hipe to beam calls use export entry. Makes it a lot easier to handle global tracing correctly (breakpoints in export entry). A beam function should now be traced correctly regardless how it is called. This will also fix a SEGV crash when a hipe stub is made pointing into a traced export entry and tracing is then stopped which clears the export entry causing the hipe stub to execute beam instruction NULL. This commit assumes that hipe code never calls local beam functions, which should be the case nowadays as we only hipe compile entire modules.
2014-04-29erts: Save some space in process struct for hipeSverker Eriksson
by combining hipe.ncallee and hipe.closure in a union as the comment indicate should be possible.
2014-03-21Merge branch 'yiannist/hipe-llvm-backend'Henrik Nord
* yiannist/hipe-llvm-backend: Support the LLVM backend in HiPE Implement the LLVM backend Extend RTL API to support the LLVM backend Add support for llvm unique symbols in hipe_gensym Add a BIF that only returns the atom ok Move some common code in hipe_pack_constants Add better specs in hipe_pack_constants and cleanup OTP-11801
2014-03-13Merge branch 'sverk/valgrind-leaks'Sverker Eriksson
* sverk/valgrind-leaks: erts: Suppress false leak in hipe_thread_signal_init erts: Fix leak in nif_SUITE:resource_takeover (again)
2014-03-07Add a BIF that only returns the atom okYiannis Tsiouris
A call in llvm_fix_pinned_regs/0 is inserted in the beginning of LLVM unwind blocks (i.e. code executed when an LLVM 'invoke' call triggers an exception) in order to get the BP and SP registers right. This is needed because LLVM exception handling doesn't provide the return value (that also contains the values for these registers, as defined in the HiPE Calling Convention for LLVM) on the 'landingpad's.
2014-02-28erts: Suppress false leak in hipe_thread_signal_initSverker Eriksson