aboutsummaryrefslogtreecommitdiffstats
path: root/erts/emulator
AgeCommit message (Collapse)Author
2010-03-10Add the BeamInstr data type for loaded BEAM codePatrik Nyblom
For cleanliness, use BeamInstr instead of the UWord data type to any machine-sized words that are used for BEAM instructions. Only use UWord for untyped words in general.
2010-03-10Fix the BEAM dissambler for the half-word emulatorPatrik Nyblom
2010-03-10Store pointers to heap data in 32-bit wordsPatrik Nyblom
Store Erlang terms in 32-bit entities on the heap, expanding the pointers to 64-bit when needed. This works because all terms are stored on addresses in the 32-bit address range (the 32 most significant bits of pointers to term data are always 0). Introduce a new datatype called UWord (along with its companion SWord), which is an integer having the exact same size as the machine word (a void *), but might be larger than Eterm/Uint. Store code as machine words, as the instructions are pointers to executable code which might reside outside the 32-bit address range. Continuation pointers are stored on the 32-bit stack and hence must point to addresses in the low range, which means that loaded beam code much be placed in the low 32-bit address range (but, as said earlier, the instructions themselves are full words). No Erlang term data can be stored on C stacks (enforced by an earlier commit). This version gives a prompt, but test cases still fail (and dump core). The loader (and emulator loop) has instruction packing disabled. The main issues has been in rewriting loader and actual virtual machine. Subsystems (like distribution) does not work yet.
2010-03-10Add a custom mmap wrapper to force heaps into the lower address rangePatrik Nyblom
The free list is still rudimentary for the mmap wrapper and a better implementation will be needed for production quality.
2010-03-10Fit all heap data into the 32-bit address rangePatrik Nyblom
This is the first step in the implementation of the half-word emulator, a 64-bit emulator where all pointers to heap data will be stored in 32-bit words. Code specific for this emulator variant is conditionally compiled when the HALFWORD_HEAP define has a non-zero value. First force all pointers to heap data to fall into a single 32-bit range, but still store them in 64-bit words. Temporary term data stored on C stack is moved into scheduler specific storage (allocated as heaps) and macros are added to make this happen only in emulators where this is needed. For a vanilla VM the temporary terms are still stored on the C stack.
2010-03-10OTP-8475 status lock needed when looking up ioq from async threadsRickard Green
Driver threads, such as async threads, using <seealso marker="erl_driver#ErlDrvPDL">port data locks</seealso> peeked at the port status field without proper locking when looking up the driver queue.
2010-03-10OTP-8487 unregister/1 on a port can cause the non-smp runtime to crashRickard Green
A call to the BIF unregister(RegName) when a port had the name RegName registered in the runtime system without SMP support caused a runtime system crash. (Thanks to Per Hedeland for the bugfix and test case.)
2010-02-18OTP-8451 Harmless buffer overflow by one byte in asn1 and ram_file_drv.Sverker Eriksson
2010-02-12Merge branch 'ms/pcre-security' into ccase/r13b04_devErlang/OTP
* ms/pcre-security: Fix CVE-2008-2371 (outer level option with alternatives caused crash). OTP-8438 The re module: A regular expression with an option change at the start of a pattern that had top-level alternatives could cause overwriting and/or a crash. (Thanks to Michael Santos.)
2010-02-11Merge branch 'ks/erts' into ccase/r13b04_devErlang/OTP
* ks/erts: hipe_bif0.c: Remove $Id$ erts/preloaded: Add types and specs
2010-02-11hipe_bif0.c: Remove $Id$Kostis Sagonas
2010-02-11OTP-8335 Even more NIF featuresSverker Eriksson
2010-02-11Fix CVE-2008-2371 (outer level option with alternatives caused crash).Michael Santos
The patch is from: http://vcs.pcre.org/viewvc?revision=360&view=revision Test case: re:compile(<<"(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]">>, [unicode]). An option change at the start of a pattern that had top-level alternatives could cause overwriting and/or a crash. This potential security problem was recorded as CVE-2008-2371.
2010-02-10OTP-8420 Fix processes in exiting status that are about to be scheduled,Björn-Egil Dahlberg
to not be allowed to garbage collect.
2010-02-10Merge branch 'egil/lcnt' into ccase/r13b04_devErlang/OTP
* egil/lcnt: Add test suite for lcnt in tools Add lcnt:rt_opt/1 bindings to erts_debug Add runtime option to enable/disable lcnt stats Add auto width on string output Add lcnt documentation Add lock profiling tool OTP-8424 Add lock profiling tool. The Lock profiling tool, lcnt, can make use of the internal lock statistics when the runtime system is built with this feature enabled. This provides a mechanism to examine potential lock bottlenecks within the runtime itself. - Add erts_debug:lock_counters({copy_save, bool()}). This option enables or disables statistics saving for destroyed processes and ets-tables. Enabling this might consume a lot of memory. - Add id-numbering for lock classes which is otherwise undefined.
2010-02-08Add runtime option to enable/disable lcnt statsBjörn-Egil Dahlberg
Add erts_debug:lock_counters({copy_save, bool()}). This option enables or disables statistics saving for destroyed processes and ets-tables. Enabling this might consume a lot of memory. Add id-numbering for lock classes which is otherwise undefined.
2010-02-08Merge branch 'gc/hipe_darwin_amd64' into ccase/r13b04_devErlang/OTP
* gc/hipe_darwin_amd64: Fix hipe memory allocation problems on darwin/amd64 Porting x86 darwin fixes to amd64 darwin hipe asm/m4 code Automatically enable hipe for darwin/amd64 builds Allow configure to enable_hipe for darwin/amd64 builds OTP-8416 HiPE now works in the 64-bit emulator on Mac OS X. (Thanks to Geoff Cant.)
2010-02-08OTP-8412 Fixed numerous compiler warnings generated by gcc 4.4.1 andRickard Green
tile-cc 2.0.1.78377 when compiling the runtime system.
2010-02-06Fix hipe memory allocation problems on darwin/amd64Geoff Cant
HiPE on amd64 needs to be able to allocate memory in the 0x20000000+ range so that compiled code will be addressable with 32bit pointers (when sign extended). By default, when building 64bit executables on OS X, PAGEZERO occupies the first 4GB of virtual memory. The PAGEZERO segment cannot be mmap(ed) or vm_allocate(d) into. To avoid this problem we use the -pagezero_size linker flag to reduce the size of PAGEZERO and free up the 0x20000000+ space.
2010-02-06Porting x86 darwin fixes to amd64 darwin hipe asm/m4 codeGeoff Cant
Compared to GNU as, Mac OS X's assembler uses different directives for text and global sections, and omits type and size directives entirely. We also need to mangle symbol names in order to allow linking with C object files.
2010-02-05Fixed erroneous assertion.Rickard Green
2010-02-04Fixed an erroneous assertion.Rickard Green
2010-02-03Merge branch 'jb/atom-table-size' into ccase/r13b04_devErlang/OTP
* jb/atom-table-size: Add the +t emulator option to change the maximum number of atoms OTP-8405 There is a new +t emulator option for changing the maximum number of atoms. (Thanks to Julien Barbot.)
2010-02-03OTP-8386 Immediately repeated multi-scheduling block/unblock cycles usingRickard Green
erlang:system_flag(multi_scheduling, block | unblock) could deadlock the runtime system.
2010-02-02Add the +t emulator option to change the maximum number of atomsJulien Barbot
It is now possible to increase or decrease the maximum number of atoms the VM can handle. The default value is 1048576 (1024*1024).
2010-01-31hipe_mfait_lock needs to be below proc_main.Rickard Green
2010-01-30Merge branch 'mp/hipe-smp-fixes' into ccase/r13b04_devErlang/OTP
* mp/hipe-smp-fixes: work around hipe_mfa_info_table lock omission fix hipe loader SMP non-atomicity error OTP-8397 The loading of native code was not properly atomic in the SMP emulator, which could cause crashes. Also a per-MFA information table for the native code has now been protected with a lock since it turns that it could be accessed concurrently in the SMP emulator. (Thanks to Mikael Pettersson.)
2010-01-27work around hipe_mfa_info_table lock omissionMikael Pettersson
HiPE maintains per-MFA information such as native code entry point in a table. This table was thought to be read-only at runtime, except when the loader populates it, so it employed no locking. That turned out to be incorrect: if there is an apply of a previously unseen MFA, a native code stub for that MFA is created and recorded in the table, causing it to grow. Work around this for now by slapping a mutex around accesses to that table. Signed-off-by: Mikael Pettersson <[email protected]>
2010-01-27Fix re:replace/4 to handle unicode charlist RE argumentPatrik Nyblom
The real problem is in the re:run/3 BIF. Noticed-by: Rory Byrne Tests-by: Rory Byrne
2010-01-25Merge branch 'ta/os_timestamp_no_mutex' into ccase/r13b04_devErlang/OTP
* ta/os_timestamp_no_mutex: Remove mutex lock around sys_gettimeofday() in os:timestamp/0 OTP-8390 An unecessary lock operation in os:timestamp/0 has been eliminated, making it slightly more efficient. (Thanks to Jonas Falkevik and Tuncer Ayaz.)
2010-01-25OTP-8377 ETS bug causing the memory counter from ets:info forSverker Eriksson
ordered_set's to sometimes get out of synch and absurdly high.
2010-01-23Remove mutex lock around sys_gettimeofday() in os:timestamp/0Tuncer Ayaz
In the erlang-questions thread "os:timestamp() uses mutex lock" Jonas Falkevik questioned the need for a mutex lock in os:timestamp/0. The mutex lock *is* needed in erlang:now() to guarantee the uniqueness of the returned time, but serves no useful purpose in os:timestamp(). Signed-off-by: Tuncer Ayaz <[email protected]>
2010-01-22OTP-8386 Immediately repeated multi-scheduling block/unblock cycles usingRickard Green
erlang:system_flag(multi_scheduling, block | unblock) could deadlock the runtime system.
2010-01-20OTP-8375 A bug causing memory corruption in re:run() has been corrected.Patrik Nyblom
(Thanks to Yamashina Hio.)
2010-01-19OTP-8373 ERTS makefiles used to detect the use of a gcc C compiler byRickard Green
checking if CC equaled gcc. That is, the makefiles failed to detect gcc C compilers with other command line names than gcc. `configure' now substitute GCC into the makefiles. If CC is a gcc C compiler, GCC will have the value yes. (Thanks to Jean-S�bastien P�dron)
2010-01-19OTP-8342 Processes, and/or ports could get stuck on a blocked schedulerRickard Green
when erlang:system_flag(multi_scheduling, block) was used. Processes, and/or ports could get stuck on an offline scheduler when schedulers online were reduced using erlang:system_flag(schedulers_online, SchedulersOnline).
2010-01-19Merge branch 'bg/compiler-beam_validator' into ccase/r13b04_devErlang/OTP
* bg/compiler-beam_validator: beam_validator: fix incorrect assumptions about GC guard BIFs OTP-8378 In rare circumstances when using garbaging collecting guard BIFs, the validation pass (beam_validator) would signal that the code was unsafe, when it in fact was correct. (Thanks to Kiran Khaladkar.)
2010-01-18Fix annoying failing statistics_SUITE test casesBjörn Gustavsson
In an SMP emulator, two of the test cases in statistics_SUITE frequently fail. Remove the runtime_zero_update/1 test case, because you can't really expect an SMP emulator not to spend any CPU time at all even if no Erlang process is active. Adjust the runtime_update/1 test case to not fail if the measured times are note close enough. It will still fail if statistics(runtime) would fail or return the wrong type.
2010-01-15Merge branch 'egil/binary-gc' into ccase/r13b04_devErlang/OTP
* egil/binary-gc: Add documentation for binary heap size settings. Add tracing capabilities for binary virtual heap Add min heap size start options to beam and erl Improve binary garbage collection OTP-8370 The default settings for garbage collection of binaries has been adjusted to be less aggressive than in R13B03. It is now also possible configure the settings for binary GC. See the documentation for spawn_opt/2-5, erlang:system_info/1, erlang:system_flag/2, process_flag/2-3, erlang:trace/3, and the documenation for erl for the new command line options +hms and +hmbs.
2010-01-15beam_validator: fix incorrect assumptions about GC guard BIFsBjörn Gustavsson
The beam_validator pass incorrectly assumes that a GC guard BIF (such as length/1) may first do a garbage collection and then fail. That assumption is not correct (guards BIF only do garbage collection when it is known that the BIF call will succeed), and will cause the compiler to reject valid programs. Modify the beam_validator to assume that if the branch is taken for a gc_bif instruction, all registers are unchanged and no garbage collection has occurred. Also add a comment in the emulator about that assumption.
2010-01-13Add tracing capabilities for binary virtual heapBjörn-Egil Dahlberg
2010-01-13Add min heap size start options to beam and erlBjörn-Egil Dahlberg
The erl (and beam) start arguments are extended with the following options: * +hms Size, sets the default minimum heap size for processes. * +hmbs Size, sets the default minimum binary virtual heap size for processes. The previous +h Size argument can still be used for backward compatibility purposes.
2010-01-13Improve binary garbage collectionBjörn-Egil Dahlberg
The garbage collector in r13b03 is too aggressive in some cases. This commit raises the level of default initial allowed binary garbage (virtual heap for binaries) before collecting from 233 words to 46368 words (181 kB on 32-bit). A new option, min_bin_vheap_size, has been added to spawn_opt, system_flag and process_flag can be used to change the default values. The option can also be used with system_info and process_info to inspect the values. For symmetry the option min_heap_size has been added to the above functions where it was previously missing. Add testcases for min_bin_vheap_size and min_heap_size for functions process_flag/2, process_info/2, system_info/2 and spawn_opt/2.
2010-01-13OTP-8240 Improved GC performance after BIF/NIF call when a lot of heapSverker Eriksson
fragments was created. This will mainly benefit NIFs that return large compound terms.
2010-01-12Merge branch 'jv/binary_to_term-opts' into ccase/r13b04_devErlang/OTP
* jv/binary_to_term-opts: document ErtsExternalDist flags and CON_ID mask add options to binary_to_term OTP-8367 There is new erlang:binary_to_binary/2 BIF that takes an option list. The option safe can be used to prevent creation of resources that are not garbage collected (such as atoms). (Thanks to Jayson Vantuyl.)
2010-01-08document ErtsExternalDist flags and CON_ID maskJayson Vantuyl
In the ErtsExternalDist structure, the flags field holds a combination of flags (tagged into the high bits) and the connection ID (in the low bits). This wasn't clearing indicated anywhere. This patch adds a comment before the flags and mask that indicates their use and relation to each other. This will help guide people through the code and reduce the likelihood that someone will add a flag without adjusting the mask.
2010-01-08add options to binary_to_termJayson Vantuyl
term_to_binary and binary_to_term are powerful tools that can be used easily in lieu of a custom binary network protocol. Unfortunately, carefully crafted data can be used to exhaust the memory in an Erlang node by merely attempting to decode binaries. This makes it unsafe to receive data from untrusted sources. This is possible because binary_to_term/1 will allocate new atoms and new external function references. These data structures are not garbage collected. This patch implements the new form of binary_to_term that takes a list of options, and a simple option called 'safe'. If specified, this option will cause decoding to fail with a badarg error if an atom or external function reference would be allocated. In the general case, it will happily decode any Erlang term other than those containing new atoms or new external function references. However, fun, pid, and ref data types can embed atoms. They might fail to decode if one of these embedded atoms is new to the node. This may be an issue if encoded binaries are transferred between nodes or persisted between invocations of Erlang.
2009-12-17Send up AssocID as unsignedBjörn Gustavsson
It must be unsigned so that prim_inet will not reject when it is sent down again. (Suggested fix by Raimo for a bug reported by Simon Cornish.)
2009-12-15Merge branch 'sc/darwin_build_rm_dSYMN' into ccase/r13b04_devErlang/OTP
* sc/darwin_build_rm_dSYMN: Fix 'make clean' in erts with debug builds under Darwin
2009-12-14Fix 'make clean' in erts with debug builds under DarwinSimon Cornish
The Apple compiler creates DWARF debugging symbols in a bundle (ie. a directory) named <name>.dSYM In order for make clean to work correctly, the rm command needs -r