gmane.comp.gcc.devel

adding ability to scan few local variables in GGC?

Basile STARYNKEVITCH — 2008-10-12T20:15:30

Hello All, I am sometimes wishing to be able to scan some few local variables in GCC garbage collector, GGC. One could imagine, for instance, that some passes would prefer, instead of using static GTY-ed variables, to declare some local GTY-ed structure LS , and to explicitly invoke the GGC collector with a pointer to LS and its marking routine (as generated by gengtype). As a fictional example, suppose some pass is building a gimple_seq stored in a local gs. It could perhaps call the GGC garbage collector as ggc_collect_with_local(gs, gt_ggc_mx_gimple_seq_node_d); So I am suggesting that we could extend the API in gcc/ggc.h with an additional function ggc_collect_with_local which takes a pointer to a single local structure (inside the call stack) and a marking routine (usually generated by gengtype) for this pointer and do a collection with marking the additional stuff. We might restrict that the passed pointer is on the call stack (or on the contrary in the GGC heap). what do people think a

eamonn raymona deepak

bryce ramana — 2008-10-12T15:23:58

camille laura maryam

Can not use correct shared lib version.

廖旂湧 — 2008-10-12T16:50:33

Hi, I use cross-compiler to build my code in my PC Linux environment, and put the executable binary to another hardware board ( embedded Linux in this board). I get the message said"error while loadingshard libraries: libdl.so.0: cannot open shared object file: No such file or directory ..." I have add library path to environment, but I still encountered this error. Actually, my Linux only has libdl.so.2 file, not have libdl.so.0. However, if I build this code with the gcc tool chain of my PC Linux, the executable binary can be exectued in PC. Anything I need to do? Many thanks!! Chiyung Liao. ______________________________________________________________________________________________________ 付費才容量無上限？Yahoo!奇摩電子信箱2.0免費給你，信件永遠不必刪！ http://tw.mg0.mail.yahoo.com/dc/landing

Does GCC support segmented memory ?

Dong Phuong — 2008-10-12T16:40:35

I'm porting for a microcontroller which support segment memory. But I don't know how to porting GCC so that it can understand addresses in different memory segment. For example, I want to create seperate code segment and data segment. Is it possible in GCC ? and if the answer is "yes", how can I do that ? Thank you very much.

For some class F, I can declare a variable of type F::F, or F::F::F, etc.

Weiqi Gao — 2008-10-12T13:52:34

Hi, We come across what seems to be a bug in gcc. If a class F has a public zero argument constructor, then we can declare a variable of type F::F, F::F::F, etc. For example, the following source file: // foo.cpp class F {}; F::F::F::F::F f; compiles with out errors in g++. The result is as if f is declared with F f; This is the case with the stock/latest GCC in Debian GNU/Linux x86_64 (v.4.1.2) and in Cygwin (v.3.4.4). Is this the intended behavior?

question. type long long

Александр Струняшев — 2008-10-12T08:29:20

Good afternoon. I need some help. As from what versions your compiler understand that "long long" is 64 bits ? Best regards, Alexander P.S. Sorry for my mistakes, I know English bad.

dhiraj chi-yao jasho

cal chen — 2008-10-11T08:57:47

haibo eladio dave

gcc-4.4-20081010 is now available

2008-10-11T00:08:18

Snapshot gcc-4.4-20081010 is now available on ftp://gcc.gnu.org/pub/gcc/snapshots/4.4-20081010/ and on various mirrors, see http://gcc.gnu.org/mirrors.html for details. This snapshot has been generated from the GCC 4.4 SVN branch with the following options: svn://gcc.gnu.org/svn/gcc/trunk revision 141048 You'll find: gcc-4.4-20081010.tar.bz2 Complete GCC (includes all of below) gcc-core-4.4-20081010.tar.bz2 C front end and core compiler gcc-ada-4.4-20081010.tar.bz2 Ada front end and runtime gcc-fortran-4.4-20081010.tar.bz2 Fortran front end and runtime gcc-g++-4.4-20081010.tar.bz2 C++ front end and runtime gcc-java-4.4-20081010.tar.bz2 Java front end and runtime gcc-objc-4.4-20081010.tar.bz2 Objective-C front end and runtime gcc-testsuite-4.4-20081010.tar.bz2 The GCC testsuite Diffs from 4.4-20081003 are available in the diffs/ subdirectory. When a particular snapshot is ready for public consumption the LATEST-4.4 link is updated an

Branch created for transactional memory

Richard Henderson — 2008-10-11T00:30:35

I've created svn://gcc.gnu.org/svn/gcc/branches/transactional-memory/ for the development of support for transactional memory within gcc. This branch will track mainline gcc. There are some folks that have already been working on this project; their page is at http://www.hipeac.net/node/2419 So far, I've applied their patch (which is vs 4.3) against mainline, fixed up *.rej, and committed the result. I've not yet gone through their patch to convert it to tuples, or anything else that would let it compile. That I'll begin doing next week. Please mark anything related to this branch with [trans-mem]. Cheers, r~

install path in libgcc Makefile.in

Zhang Le — 2008-10-10T19:17:31

Hi, Another problem when cross building the native mips compiler. I.e. build=x86, host=target=mipsel I have done some search, but haven't found any related discussion. The install path of libgcc contains gcc version. Currently in libgcc/Makefile.in, gcc version is get like this: version := $(shell $(CC) -dumpversion) This will lead to a problem that I have only experienced when cross building the native compiler. The problem is that crtbegin.o/crtend.o/etc. will be installed into old gcc's path, overwrite existing ones. New gcc will have no crtbegin.o... There is no problem when natively building native compiler. I don't understand why the problem only happens when cross building native compiler. I know I need to study how gcc is built. Just want make you aware of this problem. And I have a suggestion to this problem. I have observed that in other components, the version is get like this: version := $(shell cat $(srcdir)/../gcc/BASE-VER) So what about we do the same in libgcc/Makefile.in? Because t

Compiler error w/ templates and default argument value

Peter A. Felvegi — 2008-10-10T19:16:10

Hello All, I've run into this: ----8<----8<----8<---- template class B { protected: static const int i = 42; }; template class D : protected B { public: D(int n_ = B::i); // line 12 }; ----8<----8<----8<---- gcc-4.1 and 4.3 give the same error message: t.cpp:12: error: expected ‘,’ or ‘...’ before ‘>’ token t.cpp:12: error: wrong number of template arguments (1, should be 2) t.cpp:2: error: provided for ‘template class B’ t.cpp:12: error: default argument missing for parameter 2 of ‘D::D(int, A)’ If I write D(int n_ = (B::i) ), then it compiles. Imho it should compile w/o the extra parentheses. VC++ 2008 Express compiled it. Wanted to test gcc 4.4, too, but I wasn't able to compile the 20081003 snapshot (/usr/include/gnu/stubs.h:7:27: error: gnu/stubs-32.h: No such file or directory). I have a debian lenny/amd64 system. Regards, P

[graphite] Cleanup of command line parameters

Tobias Grosser — 2008-10-10T18:26:45

Hi graphities, graphite consists of four flags "-floop-block", "-floop-interchange", "-floop-stripmine" and "-fgraphite". If any of these flags is set, we enable the graphite pass and we search for SCoPs. For every SCoP we try to apply transformations specified with "-floop-block", "-floop-interchange" or "-floop-stripmine". But only if we could apply a transformation, we replace the SCoP with new code generated from the GRAPHITE data structures. Otherwise we do not touch the GIMPLE representation. If we only specify "-fgraphite", we never generate code for any SCoP, as we never tried any transformation. So just using "-fgraphite" is useless. What is missing is a way to make GRAPHITE always generate code, even if we did not apply any transformation. This has two benefits: - We can check the overhead the GIMPLE -> GRAPHITE -> GIMPLE transformation costs. - Wider test coverage, as we are able to run the code generator for every SCoP, not only the SCoPs, we are able to transform. Now: ---- -fgraphite:

gcc moving memory reference across call

Andrew Haley — 2008-10-10T17:55:02

I have some broken code, compiled from Java source. It looks like: D.843 = &java.text.Collator.class$$; _Jv_InitClass (D.843); D.845 = &_CD_java_text_Collator; is being turned into: D.843 = &java.text.Collator.class$$; D.845 = &_CD_java_text_Collator; _Jv_InitClass (D.843); i.e. the memory reference is moved to before the call to _Jv_InitClass. To get to my question: I presume that loop2 is supposed to be able to move loads across calls, but it can only do so if the memory is unreachable from the callee. But how does loop2 know if the memory is reachable? Is there a full data flow graph computed, or does it depend one the scope of the memory decl? Thanks, Andrew. in the loop2_invariant dump we have: (call_insn 31 30 33 3 /home/aph/gcc/classpath-098-merge-branch/libjava/java/text/Collator.java:304 (call (mem:QI (symbol_ref:DI ("_Jv_InitClass") [flags 0x41] ) [0 S1 A8]) (const_int 0 [0x0])) 646 {*call_0} (expr_list:REG_DE

Copyright notices during assignment limbo

Joern Rennecke — 2008-10-10T16:43:31

For code that we have written, what should the copyright notices read during the period where we have given the FSF a copyright assignemnt, but they haven't yet acknowledged that it is on file? Obviously when we wrote the code, we've put our own copyright notices on it because we need to have a copyright in it before we can assign it to someone else. Eventually, for the code that we have assigned, they should be changed to say Copyright FSF. But at what point in time should the Copyright notices actually be changed?

G.C.C. should support more Ciphers.

Rüdiger Müller — 2008-10-10T08:32:13

Hello, how to support UNICODE with G.C.C.? How to support e.G. XML-FILE to translate "reserved Words" in G.C.C.? Is the only possibility to support e.g. "Umlauts" in Source-Text a Pre-Translator for the G.C.C. Translator? For Example: 1. C-Text in Unicode 2. Own Pre-Translator that gives "G.C.C-ready" Bookstabes with no "Umlauts" or other ciphers that are not supported by G.C.C.. How many ciphers do G.C.C. already support? Why does G.C.C. does not more ciphers support? Where is the technical challenge in supporting more ciphers? Does a Pre-Translator, from e.G. Unicode to G.C.C.-Ciphers already exist? Where to get this Pre-Translator? With friendly wishes, R. Müller 09.10.08 11:31 Hallo zurück, vielen Dank für die Antworten. Mit z.B. "UNICODE" Unterstützung könnte man ja nicht nur lokalisierte Sprachversionen anbieten, sondern auch eigene "mathematische Beschreibungssprachen". Mit z.B. xml-Karteien, die die einzelnen reservierten Wörter eigens einstellen können, könnte man dann j

gcc-4.3-20081009 is now available

2008-10-09T22:44:57

Snapshot gcc-4.3-20081009 is now available on ftp://gcc.gnu.org/pub/gcc/snapshots/4.3-20081009/ and on various mirrors, see http://gcc.gnu.org/mirrors.html for details. This snapshot has been generated from the GCC 4.3 SVN branch with the following options: svn://gcc.gnu.org/svn/gcc/branches/gcc-4_3-branch revision 141011 You'll find: gcc-4.3-20081009.tar.bz2 Complete GCC (includes all of below) gcc-core-4.3-20081009.tar.bz2 C front end and core compiler gcc-ada-4.3-20081009.tar.bz2 Ada front end and runtime gcc-fortran-4.3-20081009.tar.bz2 Fortran front end and runtime gcc-g++-4.3-20081009.tar.bz2 C++ front end and runtime gcc-java-4.3-20081009.tar.bz2 Java front end and runtime gcc-objc-4.3-20081009.tar.bz2 Objective-C front end and runtime gcc-testsuite-4.3-20081009.tar.bz2 The GCC testsuite Diffs from 4.3-20081002 are available in the diffs/ subdirectory. When a particular snapshot is ready for public consumption the LATEST-4.3

IRA compile time on gcc-c-torture/compile/limits-fnargs.c

2008-10-09T21:12:57

Is anyone else seeing a long compile time on gcc-c-torture/compile/limits-fnargs.c with -O2 optimization? limits-fnargs.c contains a single call to a function with 10,000 arguments. There is PR 31850 which talks about compile time on limits-fnargs.c, but this was fixed (at least on IA64, if not for spu-elf) a while back and this new time out problem only affects IA64 when -O2 or higher optimization is used. I tracked the time down to the main loop in coalesce_spill_loops in ira-colors.c. The main IRA pass calls reload which calls ira_sort_regnos_for_alter_reg with n = 28977. This in turn calls coalesce_spill_slots with num = 10061. It is in the main loop of coalesce_spill_slots where I seem to spend a lot of time compiling limits-fnargs.c. Vladimir, is this something you are aware of? Would you like me to submit a Bugzilla report? Steve Ellcey sje< at >cup.hp.com

divmodsi4

Omar Torres — 2008-10-09T18:04:02

Hi All, I am having trouble distinguishing div vs mod while implementing the divmodsi4 instruction. The gccint documentation states: "If an instruction that just produces a quotient or just a remainder exists and is more efficient than the instruction that produces both, write the output routine of 'divmodm4' to call find_reg_note and look for a REG_UNUSED note on the quotient or remainder and generate the appropriate instruction." The problem is that both, the quotient and reminder, registers are getting marked with a REG_UNUSED note: (insn 12 11 17 (parallel [ (set (reg:SI 1 %r3 [33]) (div:SI (reg:SI 1 %r3 [30]) (reg:SI 5 %iph [orig:31 current ] [31]))) (set (reg:SI 5 %iph [34]) (mod:SI (reg:SI 1 %r3 [30]) (reg:SI 5 %iph [orig:31 current ] [31]))) ]) 56 {*divmodsi4} (insn_list:REG_DEP_TRUE 10 (insn_list:REG_DEP_TRUE 11 (nil))) (expr_list:REG_UNUSED (reg:SI 5 %iph [34]) (expr_list:REG_UNU

Support for NT based OS on ARM.

Farlie A — 2008-10-09T12:56:15

Hi, Would you be willing to consider supporting the PE object formats on the ARM based port of ReactOS? (ReactOS is an attempt to reimplement the NT kernel architecture and API as free software based on publicly available specifications and API documentation)

"UNICODE"-Unterstützung, XML-Reservierte-Wörter-Unterstützung / "UNICODE"-support, XML-FILE-WITH-RESERVED-WORDS Support

Rüdiger Müller — 2008-10-09T09:30:25

Hallo zurück, vielen Dank für die Antworten. Mit z.B. "UNICODE" Unterstützung könnte man ja nicht nur lokalisierte Sprachversionen anbieten, sondern auch eigene "mathematische Beschreibungssprachen". Mit z.B. xml-Karteien, die die einzelnen reservierten Wörter eigens einstellen können, könnte man dann ja auch mehrere mathematische Operatoren, wie z.B. "mathematisches UND /\" oder "mathematisches ODER \/" anstelle von "&&" oder "||" ersetzen. Wie kommt man zu einer solchen (Universeller Zeichensatz) beispielsweise "UNICODE" Unterstützung des G.-C-Übersetzers? Wie kann man xml-Karteien für "Reservierte-Wörter" Ersetzungen mit dem G.C-Übersetzer nutzen? M.f.G. R. Müller ---- Hello, how to support UNICODE with G.C.C.? How to support XML-FILE to translate "reserved Words" in G.C.C.? R. Müller ________________________________________________________________________ Schon gehört? Bei WEB.DE gibt' s viele kostenlose Spiele: http://games.entertainment.web.de/de/entertainment/games/free/index.h

gcc-4.2-20081008 is now available

2008-10-08T22:40:55

Snapshot gcc-4.2-20081008 is now available on ftp://gcc.gnu.org/pub/gcc/snapshots/4.2-20081008/ and on various mirrors, see http://gcc.gnu.org/mirrors.html for details. This snapshot has been generated from the GCC 4.2 SVN branch with the following options: svn://gcc.gnu.org/svn/gcc/branches/gcc-4_2-branch revision 140989 You'll find: gcc-4.2-20081008.tar.bz2 Complete GCC (includes all of below) gcc-core-4.2-20081008.tar.bz2 C front end and core compiler gcc-ada-4.2-20081008.tar.bz2 Ada front end and runtime gcc-fortran-4.2-20081008.tar.bz2 Fortran front end and runtime gcc-g++-4.2-20081008.tar.bz2 C++ front end and runtime gcc-java-4.2-20081008.tar.bz2 Java front end and runtime gcc-objc-4.2-20081008.tar.bz2 Objective-C front end and runtime gcc-testsuite-4.2-20081008.tar.bz2 The GCC testsuite Diffs from 4.2-20081001 are available in the diffs/ subdirectory. When a particular snapshot is ready for public consumption the LATEST-4.2