summaryrefslogtreecommitdiff
path: root/gnu/egcs/gcc
AgeCommit message (Collapse)Author
2003-04-17Change gcc to generate ELF code for i386.Dale Rahn
2003-04-14standalone/stand alone -> stand-alone;Jason McIntyre
from NetBSD (Igor Sobrado); httpd stuff passed to apache people; ok millert@
2003-04-09fix the readonly objects placement into thee .rodata proper; espie@ okMichael Shalayeff
2003-04-09xmalloc called with the wrong sizeof in sentinel attribute setupAnil Madhavapeddy
espie@ ok
2003-04-08fix another non-POSIX chownTodd C. Miller
2003-03-28PIC does not mix well w/ nopic (yet); espie@ okMichael Shalayeff
2003-03-24Recognize -soname foo, so that ld can properly ignore it.Marc Espie
ok millert@, pval@.
2003-03-18spelling typos/errorsDavid Krause
ok miod@ jmc@
2003-03-15New frame addressing (plus (subreg (reg fp)) const) is supported by ↵Hiroaki Etoh
push_frame_in_operand.
2003-03-12typo; Ross.Richardson@utas.edu.auTheo de Raadt
2003-03-11The variable cse_not_expected must be initialized.Hiroaki Etoh
2003-03-11The loop optimization using frame pointer is omitted.Hiroaki Etoh
2003-03-11not to combine the register related to frame addressing for the case ↵Hiroaki Etoh
FRAME_GROWS upward
2003-03-09Originally from gcc:Henric Jungheim
2002-06-14 Jeff Sturm <jsturm@one-point.com> * config/sparc/sparc.h (DYNAMIC_CHAIN_ADDRESS): Add SPARC_STACK_BIAS. 2002-03-12 Andrew MacLeod <amacleod@redhat.com> * config/sparc/sparc.h (RETURN_ADDR_RTX): Include v9 stack bias in address calculation. through Hiroaki Etoh (thanks). This gets "__builtin_return_address()" to work on sparc64. ok deraadt@
2003-03-06broken .Xr's;Jason McIntyre
typos, rather than dead references ok deraadt@
2003-02-28changes to fold the frame address of a setjmp insnHiroaki Etoh
2003-02-27add -trace-ctors-dtors option to collect2.Marc Espie
Very useful for figuring out what's wrong with C++ shared libraries on a.out platforms.
2003-02-27typo; Sebastian StarkTodd C. Miller
2003-02-26preserve the addressing using frame pointer at the cse2 phase.Hiroaki Etoh
2003-02-26fix the bison problem on alpha, it uses the pointer that initialzes the ↵Hiroaki Etoh
outside address of a character array.
2003-02-21push_frame_in_operand: add the setup and restore code for setjmp on powerpc ↵Hiroaki Etoh
and alpha
2003-02-21sweep_string_in_operand: properly sweep inline function frame.Hiroaki Etoh
2003-02-21mark the top address of inline function frame with integrated flag for the ↵Hiroaki Etoh
propolice protection
2003-02-21In the case of no optimization, fix the problem that the first argument ↵Hiroaki Etoh
doesn't copy to the region of local variables.
2003-02-19oopsMarc Espie
2003-02-17fix trampoline support on hppa. Tested by mickey@Marc Espie
(does not solve the C++ ctor problem, sorry)
2003-02-16place ro data into .rodata and not .text (now that it's handled special. 10x ↵Michael Shalayeff
to dale); espie@ ok
2003-02-16a few missing man pages converted from info documents; work done by jmcTheo de Raadt
2003-02-16typoTheo de Raadt
2003-02-15use real registers to propagete frame offset for powerpc.Hiroaki Etoh
2003-02-03disable a stack protection when the arch doesn't have stack_grows_downward.Hiroaki Etoh
2003-02-03add contains_fp declaration to examine whether the input contains frame ↵Hiroaki Etoh
pointer addressing.
2003-02-03prepare_stack_protection: add validate_insns_of_varrefs to validate insns of ↵Hiroaki Etoh
inline function for processors, which doesn't have long integer addition. push_frame_of_insns: validate insns to use powerpc specific addressing and alpha specific addressing.
2003-02-03modify general_induction_var not to generate the address of frame pointer ↵Hiroaki Etoh
plus minus constant as an induction variable.
2003-02-03delete unused "protector.h" inclusion.Hiroaki Etoh
2003-02-03modify assign_stack_local to move the addressing of frame pointer plus zero ↵Hiroaki Etoh
to the function plus_constant_wide. modify assign_stack_temp_for_type for allocating a space to return a structure.
2003-02-03modify expand_expr not to modify the addressing form of local variables: ↵Hiroaki Etoh
frame pointer plus constant.
2003-02-03modify plus_constant_wide to provide the address of frame pointer plus zero ↵Hiroaki Etoh
constant.
2003-02-03modify simplify_plus_minus not to assign the address of frame pointer plus ↵Hiroaki Etoh
no constant.
2003-02-03allocate a space (keep=5) to return a structure, not protecting it as a ↵Hiroaki Etoh
character string.
2003-02-02I've been guaranteed that guaranteed is better than guarenteed.Miod Vallat
2003-01-18protoize(1) man page, cobbled from various places by jmc@prioris.mini.pw.edu.plTheo de Raadt
2003-01-14Bye-bye libc_r. libc_r and libnpthread go away. libpthreadMarco S Hyman
takes their place. The -pthread option to gcc uses libpthread. ppc portion of patch from drahn@ (thanks). gcc portions of patch OK espie@. Ports testing (no bad effects!) by naddy@ best build sequence is: build gcc, build system, build ports remove /usr/lib/lib{c_r,npthread}.*. Note, once libc_r and libnpthread is removed threaded ports will stop working until they have been re-built using the libpthread and the updated gcc. That's why the above build sequence is recommended.
2003-01-12Revert last change, causes issues on macppc and alpha.Miod Vallat
2003-01-07attribute(sentinel) improvements:Marc Espie
- move __null node to c-common.c so that sentinel can check for it (since g++ __null is `special') - rework diagnostics to be more clear. In particular, distinguish `not terminated' from `terminated with a 0 of the wrong type' ok miod@, millert@
2003-01-07prepare_stack_protection: add validate_insns_of_varrefs to validate insns of ↵Hiroaki Etoh
inline function for processors, which doesn't have long integer addition push_frame_of_insns: validate insns not to use long integer addition for powerpc
2002-12-25set variables of inline function with DECL_INLINEHiroaki Etoh
2002-12-25In order to determine an inlinable function, pass the condition to ↵Hiroaki Etoh
prepare_stack_protection.
2002-12-25search_string_def is declared as public.Hiroaki Etoh
2002-12-25arrange_var_order: not to weep the string of inlined function.Hiroaki Etoh