![[BACK]](/icons/back.gif){kind=icon}
Up to [cvs.NetBSD.org] / src / external / gpl3 / gcc / dist / gcc / config / vax
Request diff between arbitrary revisions
Keyword substitution: kv
Default branch: MAIN
many fixes for native GCC 12 builds - default to dwarf4 since out ctf tools fail with dwarf5. i tried to do this via subtarget options as other ports do, but this would require adjusting several of the targets non-netbsd code, and ideally we'll fix our tools sooner than later, so just hack it here for now. - catch up config/rs6000/netbsd64.h (mostly delete things) - remove some more upstream diffs in config/vax, though it does not work currently - libc.so link fails due to weird relocations. - include some headers to ensure types are available - add missing sanitizer netbsd files to the 'dist' tree - remove no longer present TIOCRCVFRAME and TIOCXMTFRAME interceptors - expanding the sanitizer OS/cpu defs to cover more netbsd platforms - copy across a few minor changes from gcc.old's sanitizer's (eg, disabling linux-specific code, adding netbsd specific code.) - add (untested, but building) tsan support for arm64 - re-enable the sanitzers, at least simple -fsanitize=address builds a working application on amd64 now - fix the source lists (mostly .cc -> .cpp rename here) - add some missing new defines, -f*, and -W* - bump sanitizer shlib version - apply -fno-jump-tables in some more places on m68k - update the list in README.gcc12 to note many things now work or a few things to do not work (although there are many uncommited changes at this point across the rest of 'src' and 'xsrc'.) main points are that vax, hppa, ppc, and sh3 don't build yet, but most other things are either testing and working or at least building.
fix merge botch.
initial merge of GCC 12.3.0. this doesn't include any of the changes to the .c -> .cc files renamed, and reverts our local changes to the vax port and libsanitizer subdir. vax GCC was rewritten and our local fixes no longer are relevant, and the new libsanitizer is more updated than our old one, and merging via gcc10->gcc12 is not really possible. unfortunately, our local changes to libsanitizer that aren't related to the general update of those sources (ie, the netbsd code) will need to be re-checked and perhaps re-ported.
initial import of GCC 12.3.0.
major changes in GCC 11 included:
- The default mode for C++ is now -std=gnu++17 instead of -std=gnu++14.
- When building GCC itself, the host compiler must now support C++11,
rather than C++98.
- Some short options of the gcov tool have been renamed: -i to -j and
-j to -H.
- ThreadSanitizer improvements.
- Introduce Hardware-assisted AddressSanitizer support.
- For targets that produce DWARF debugging information GCC now defaults
to DWARF version 5. This can produce up to 25% more compact debug
information compared to earlier versions.
- Many optimisations.
- The existing malloc attribute has been extended so that it can be
used to identify allocator/deallocator API pairs. A pair of new
-Wmismatched-dealloc and -Wmismatched-new-delete warnings are added.
- Other new warnings:
-Wsizeof-array-div, enabled by -Wall, warns about divisions of two
sizeof operators when the first one is applied to an array and the
divisor does not equal the size of the array element.
-Wstringop-overread, enabled by default, warns about calls to string
functions reading past the end of the arrays passed to them as
arguments.
-Wtsan, enabled by default, warns about unsupported features in
ThreadSanitizer (currently std::atomic_thread_fence).
- Enchanced warnings:
-Wfree-nonheap-object detects many more instances of calls to
deallocation functions with pointers not returned from a dynamic
memory allocation function.
-Wmaybe-uninitialized diagnoses passing pointers or references to
uninitialized memory to functions taking const-qualified arguments.
-Wuninitialized detects reads from uninitialized dynamically
allocated memory.
-Warray-parameter warns about functions with inconsistent array forms.
-Wvla-parameter warns about functions with inconsistent VLA forms.
- Several new features from the upcoming C2X revision of the ISO C
standard are supported with -std=c2x and -std=gnu2x.
- Several C++20 features have been implemented.
- The C++ front end has experimental support for some of the upcoming
C++23 draft.
- Several new C++ warnings.
- Enhanced Arm, AArch64, x86, and RISC-V CPU support.
- The implementation of how program state is tracked within
-fanalyzer has been completely rewritten with many enhancements.
see https://gcc.gnu.org/gcc-11/changes.html for a full list.
major changes in GCC 12 include:
- An ABI incompatibility between C and C++ when passing or returning
by value certain aggregates containing zero width bit-fields has
been discovered on various targets. x86-64, ARM and AArch64
will always ignore them (so there is a C ABI incompatibility
between GCC 11 and earlier with GCC 12 or later), PowerPC64 ELFv2
always take them into account (so there is a C++ ABI
incompatibility, GCC 4.4 and earlier compatible with GCC 12 or
later, incompatible with GCC 4.5 through GCC 11). RISC-V has
changed the handling of these already starting with GCC 10. As
the ABI requires, MIPS takes them into account handling function
return values so there is a C++ ABI incompatibility with GCC 4.5
through 11.
- STABS: Support for emitting the STABS debugging format is
deprecated and will be removed in the next release. All ports now
default to emit DWARF (version 2 or later) debugging info or are
obsoleted.
- Vectorization is enabled at -O2 which is now equivalent to the
original -O2 -ftree-vectorize -fvect-cost-model=very-cheap.
- GCC now supports the ShadowCallStack sanitizer.
- Support for __builtin_shufflevector compatible with the clang
language extension was added.
- Support for attribute unavailable was added.
- Support for __builtin_dynamic_object_size compatible with the
clang language extension was added.
- New warnings:
-Wbidi-chars warns about potentially misleading UTF-8
bidirectional control characters.
-Warray-compare warns about comparisons between two operands of
array type.
- Some new features from the upcoming C2X revision of the ISO C
standard are supported with -std=c2x and -std=gnu2x.
- Several C++23 features have been implemented.
- Many C++ enhancements across warnings and -f options.
see https://gcc.gnu.org/gcc-12/changes.html for a full list.
initial merge of GCC 10.3.0. these three files are not yet finished merging: gcc/config/rs6000/rs6000.c libsanitizer/lsan/lsan_allocator.h libsanitizer/sanitizer_common/sanitizer_platform_limits_posix.h our current GCC 9 sanitizers are newer than GCC 10's still, so that may take a little to untease, so it's probable that more than these 2 will need more changes. rs6000.c has some changes related to ABI and supported functionality that need to be merged forward.
initial import of GCC 10.3.0. main changes include: caveats: - ABI issue between c++14 and c++17 fixed - profile mode is removed from libstdc++ - -fno-common is now the default new features: - new flags -fallocation-dce, -fprofile-partial-training, -fprofile-reproducible, -fprofile-prefix-path, and -fanalyzer - many new compile and link time optimisations - enhanced drive optimisations - openacc 2.6 support - openmp 5.0 features - new warnings: -Wstring-compare and -Wzero-length-bounds - extended warnings: -Warray-bounds, -Wformat-overflow, -Wrestrict, -Wreturn-local-addr, -Wstringop-overflow, -Warith-conversion, -Wmismatched-tags, and -Wredundant-tags - some likely C2X features implemented - more C++20 implemented - many new arm & intel CPUs known hundreds of reported bugs are fixed. full list of changes can be found at: https://gcc.gnu.org/gcc-10/changes.html
merge GCC 9.3.0.
initial import of GCC 9.3.0. changes include: - live patching support - shell completion help - generally better diagnostic output (less verbose/more useful) - diagnostics and optimisation choices can be emitted in json - asan memory usage reduction - many general, and specific to switch, inter-procedure, profile and link-time optimisations. from the release notes: "Overall compile time of Firefox 66 and LibreOffice 6.2.3 on an 8-core machine was reduced by about 5% compared to GCC 8.3" - OpenMP 5.0 support - better spell-guesser - partial experimental support for c2x and c++2a - c++17 is no longer experimental - arm AAPCS GCC 6-8 structure passing bug fixed, may cause incompatibility (restored compat with GCC 5 and earlier.) - openrisc support
re-import GCC 8.4.0.
import GCC 7.5.0. doing this here so that the vendor branch has the code we'll merge into gcc.old and the netbsd-9 tree gcc tree. GCC 8.4.0 will be imported immediately on top of this again, restoring the current status. these PRs in the GCC bugzilla are fixed with this update: 89869 80693 89795 84272 85593 86669 87148 87647 87895 88103 88107 88563 88870 88976 89002 89187 89195 89234 89303 89314 89354 89361 89403 89412 89512 89520 89590 89621 89663 89679 89704 89734 89872 89933 90090 90208 87075 85870 89009 89242 88167 80864 81933 85890 86608 87145 88857 89024 89119 89214 89511 89612 89705 89400 81740 82186 84552 86554 87609 88105 88149 88415 88739 88903 89135 89223 89296 89505 89572 89677 89698 89710 90006 90020 90071 90328 90474 91126 91162 91812 91887 90075 88998 89945 87047 87506 88074 88656 88740 91137 89008 84010 89349 91136 91347 91995 89397 87030 60702 78884 85594 87649 87725 88181 88470 88553 88568 88588 88620 88644 88906 88949 89246 89587 89726 89768 89796 89998 90108 90756 90950 91704 88825 88983 86538 51333 89446 90220 91308 92143 89392 90213 90278 91131 91200 91510 89037 91481 87673 88418 88938 88948 90547 27221 58321 61250 67183 67958 77583 83531 86215 88648 88720 88726 89091 89466 89629 90105 90329 90585 90760 90924 91087 89222 81956 71861 35031 69455 81849 82993 85798 88138 88155 88169 88205 88206 88228 88249 88269 88376 77703 80260 82077 86248 88393 90786 57048 66089 66695 67679 68009 71723 72714 84394 85544 87734 88298 90937 91557 63891 64132 65342 68649 68717 71066 71860 71935 77746 78421 78645 78865 78983 79485 79540 85953 88326 89651 90744
Mostly merge changes from HEAD upto 20200411
merge GCC 8.3.0 pass 1. will not compile yet. thare are still about 25 files with merge issues.
import GCC 8.3. it includes these new features: - many optimisations improved: inter-procedural, profile-directed, LTO, loops including user-controllable unroll support, and more. - columns numbers added to line numbers in dwarf - gcov extended significantly - many sanitizer updates - many new warning messages - many better hints and more useful error messages - minor ABI changes on x86-64 libstdc++, and some c++17 modes - draft c++2a features - better c++17 experimental support - Armv8.4-A supported, better 8.2-A and 8.3-A support, including 32 bit arm port. cortex a-55, a-75 and a-55.a-75 combo support. - in the GCC bugzilla, 8.1 shows 1149 bugs fixed, 8.2 shows 100, and 8.3 shows 158.
Sync with HEAD
Sync with HEAD
first pass at merging GCC 7.4.0. unlikely to compile..
import GCC 7.4.0. main changes include: The non-standard C++0x type traits has_trivial_default_constructor, has_trivial_copy_constructor and has_trivial_copy_assign have been removed. On ARM targets (arm*-*-*), a bug introduced in GCC 5 that affects conformance to the procedure call standard (AAPCS) has been fixed. Many optimiser improvements DWARF-5 support. Many new and enhanced warnings. Warnings about format strings now underline the pertinent part of the string, and can offer suggested fixes. Several new warnings related to buffer overflows and buffer truncation. New __builtin_add_overflow_p, __builtin_sub_overflow_p, __builtin_mul_overflow_p built-ins added that test for overflow. The C++ front end has experimental support for all of the current C++17 draft. The -fverbose-asm option has been expanded to prints comments showing the source lines that correspond to the assembly. The gcc and g++ driver programs will now provide suggestions for misspelled arguments to command-line options. AArch64 specific: GCC has been updated to the latest revision of the procedure call standard (AAPCS64) to provide support for parameter passing when data types have been over-aligned. The ARMv8.2-A and ARMv8.3-A architecture are now supported. ARM specific: Support for the ARMv5 and ARMv5E architectures has been deprecated (which have no known implementations). A new command-line option -mpure-code has been added. It does not allow constant data to be placed in code sections. x86 specific: Support for the AVX-512 4FMAPS, 4VNNIW, VPOPCNTDQ and Software Guard Extensions (SGX) ISA extensions has been added. PPC specific: GCC now diagnoses inline assembly that clobbers register r2. RISC-V specific: Support for the RISC-V instruction set has been added. SH specific: Support for SH5/SH64 has been removed. Support for SH2A has been enhanced.
merge GCC 6.4.0. sanitizer stuff is probably busted, but most other changes merged easily. docs need to be regenerated with modern versions still.
import GCC 6.4.0. see this url for details which are too large to include here: http://gcc.gnu.org/gcc-6/changes.html the main visible changes appear to be: - The default mode for C++ is now -std=gnu++14 instead of -std=gnu++98. - The C and C++ compilers now support attributes on enumerators. - Diagnostics can now contain "fix-it hints" - more warnings (some added to -Wall)
Revert previous changes until we can test them better.
From Jake Hamby
For several years I've been eager to find the time to fix the bugs
in C++ exceptions on VAX to get them working on NetBSD, because
theyāve been broken for many years and it looked like only a few
changes were needed to get them working. Without C++ exceptions,
the NetBSD test suite canāt be run. The good news is that I was
able to fix all the bugs in the VAX machine description to make
C++ exceptions work in GCC 4.8.5 (version unimportant). I wrote a
blog post explaining the bugs, with patches:
Here's a short summary, with the diffs in text form at the end of this email.
1) Replace #define FRAME_POINTER_CFA_OFFSET(FNDECL) 0 with #define
ARG_POINTER_CFA_OFFSET(FNDECL) 0 in gcc/config/vax/elf.h and
gcc/config/vax/vax.h. This changes the definition of __builtin_dwarf_cfa()
to return %ap instead of %fp, which correctly points to CFA.
Previously, the stack unwinder was crashing in _Unwind_RaiseException()
trying to follow bad pointers from the initial CFA.
2) Define EH_RETURN_DATA_REGNO(N) to include only R2 and R3 (instead
of R2-R5) and add code to vax_expand_prologue() in gcc/config/vax/vax.c
to add R2-R3 to the procedure entry mask but only if crtl->calls_eh_return
is set. This fixes a crash when the stack unwinder tried to write
values to R2 and R3 in the previous stack frame via
__builtin_eh_return_data_regno (0) and __builtin_eh_return_data_regno (1).
3) Removed definitions of EH_RETURN_STACKADJ_RTX and STARTING_FRAME_OFFSET
from gcc/config/vax/elf.h. It's not necessary to remember the stack
adjustment or to waste four bytes on every stack frame for a value
that's not needed. Also remove the suspicious changes in
gcc/config/vax/vax.md to the definitions of call_pop and call_value
regarding DW_CFA_GNU_args_size and EH unwinding. I reverted to the
previous versions from an older version of GCC, adding a few useful
comments that had been removed.
4) The last bug is the one I understand the least. I'm hoping
someone reading this can implement a correct fix. What I was seeing
after making all the previous changes to fix the other bugs is that
my test program failed to catch any exceptions, but instead returned
normally to the original return path.
Investigation revealed that GCC was correctly generating the
necessary move instruction to copy the second parameter passed to
__builtin_eh_return() into the return address, because
EH_RETURN_HANDLER_RTX had been defined correctly in config/vax/elf.h.
Hereās what the call looks like in gcc/except.c:
#ifdef EH_RETURN_HANDLER_RTX
rtx insn = emit_move_insn (EH_RETURN_HANDLER_RTX, crtl->eh.ehr_handler);
#else
error ("__builtin_eh_return not supported on this target");
#endif
The problem was that the optimizer is deleting the final move
instruction when I compile with -O or higher. The assembly code at
-O0 (no optimization) generated for the __builtin_eh_return() call
at the end of _Unwind_RaiseException() looked like:
calls $2,_Unwind_DebugHook
movl -12(%fp),%r1
movl %r1,16(%fp)
ret
.cfi_endproc
But then when I compiled with -O1 or -O2, all I saw was:
calls $2,_Unwind_DebugHook
ret
.cfi_endproc
This was a mystery for me and I donāt know enough about how the
final peephole optimizer works to really track down why it thinks
it can remove the move call to store the previous return address.
My workaround was to add a call to RTX_FRAME_RELATED_P (insn) = 1;
after the emit_move_insn() in gcc/except.c, which was used in
vax_expand_prologue() to mark the procedure entry mask.
By making this change, the optimizer no longer removes the call to
write the value to the previous stack pointer, but it adds an extra
line of .cfi exception info, which seems unnecessary since the code
is immediately going to return from the call and any adjustment
made by the DWARF stack unwinder will already have been done. Hereās
what the optimized code looks like with the patch (%r6 had been
loaded earlier):
calls $2,_Unwind_DebugHook
movl %r6,16(%fp)
.cfi_offset 6, -36
ret
.cfi_endproc
With that final change, C++ exception handling now finally works
on NetBSD/vax, and I was able to successfully run the vast majority
of the tests in the ATF testsuite, which had been completely
inaccessible when I started due to both atf-run and atf-report
immediately dumping core due to the bad pointers that I fixed. Now
I have a bunch of new bugs to track down fixes for, but I think
this was the hardest set of problems that needed to be solved to
bring NetBSD on VAX up to the level of the other NetBSD ports.
Here are the diffs I have so far. They should apply to any recent
version of GCC (tested on GCC 4.8.5). With the exception of the
hack to gcc/except.c, the other diffs are ready to submit to NetBSD
as well as to upstream GCC. The fix Iād like to see for the final
problem I discovered of the emit_move_insn() being deleted by the
optimizer would be another patch to one of the files in the
gcc/config/vax directory to explain to the optimizer that writing
to 16(%fp) is important and not something to be deleted from the
epilogue (perhaps it thinks itās writing to a local variable in
the frame that's about to be destroyed?).
I didn't see any indication that any other GCC ports required
anything special to tell the optimizer not to delete the move
instruction to EH_RETURN_HANDLER_RTX, so the other suspicion I have
is that there may be a bug specific to VAX's peephole optimizer or
related functions. Any ideas?
Fix vax build, now we fail in: /usr/src/lib/csu/common/crt0-common.c: In function '___start': /usr/src/lib/csu/common/crt0-common.c:184:1: internal compiler error: in dwf_regno, at dwarf2cfi.c:988 } ^
merge GCC 5.3. round 1. this probably won't work :-) libsanitizer likely needs to be re-ported, though i've done most of the updating we'll need i think already.
import GCC 5.3.0. see these urls for details which are too large to include here: http://gcc.gnu.org/gcc-4.9/changes.html http://gcc.gnu.org/gcc-5/changes.html (note that GCC 5.x is a release stream like GCC 4.9.x, 4.8.x, etc.) the main issues we will have are: The default mode for C is now -std=gnu11 instead of -std=gnu89. ARM: The deprecated option -mwords-little-endian has been removed. The options -mapcs, -mapcs-frame, -mtpcs-frame and -mtpcs-leaf-frame which are only applicable to the old ABI have been deprecated. MIPS: The o32 ABI has been modified and extended. The o32 64-bit floating-point register support is now obsolete and has been removed. It has been replaced by three ABI extensions FPXX, FP64A, and FP64. The meaning of the -mfp64 command-line option has changed. It is now used to enable the FP64A and FP64 ABI extensions.
Rebase to HEAD as of a few days ago.
Rebase.
Add LEGITIMATE_PIC_OPERAND_P(X) to prevent global symbol + offset from being a legitimate pic operand (because it's not).
sync with head.
for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.
this commit was splitted into small chunks to avoid
a limitation of cvs. ("Protocol error: too many arguments")
Annotation for the Processor Status Word. Sensible DWARF frame layout.
import GCC 4.8 branch at r206687.
highlights from: http://gcc.gnu.org/gcc-4.6/changes.html
GCC now has stricter checks for invalid command-line options
New -Wunused-but-set-variable and -Wunused-but-set-parameter
warnings
Many platforms have been obsoleted
Link-time optimization improvements
A new switch -fstack-usage has been added
A new function attribute leaf was introduced
A new warning, enabled by -Wdouble-promotion
Support for selectively enabling and disabling warnings via
#pragma GCC diagnostic has been added
There is now experimental support for some features from the
upcoming C1X revision of the ISO C standard
Improved experimental support for the upcoming C++0x ISO C++
standard
G++ now issues clearer diagnostics in several cases
Updates for ARM, x86, MIPS, PPC/PPC64, SPARC
Darwin, FreeBSD, Solaris 2, MinGW and Cygwin now all support
__float128 on 32-bit and 64-bit x86 targets. [*1]
highlights from: http://gcc.gnu.org/gcc-4.7/changes.html
The -fconserve-space flag has been deprecated
Support for a new parameter --param case-values-threshold=n
was added
Interprocedural and Link-time optimization improvements
A new built-in, __builtin_assume_aligned, has been added
A new warning option -Wunused-local-typedefs was added
A new experimental command-line option -ftrack-macro-expansion
was added
Support for atomic operations specifying the C++11/C11 memory
model has been added
There is support for some more features from the C11 revision
of the ISO C standard
Improved experimental support for the new ISO C++ standard,
C++11
Updates for ARM, x86, MIPS, PPC/PPC64, SH, SPARC, TILE*
A new option (-grecord-gcc-switches) was added
highlights from: http://gcc.gnu.org/gcc-4.8/changes.html
GCC now uses C++ as its implementation language. This means
that to build GCC from sources, you will need a C++
compiler that understands C++ 2003
DWARF4 is now the default when generating DWARF debug
information
A new general optimization level, -Og, has been introduced
A new option -ftree-partial-pre was added
The option -fconserve-space has been removed
The command-line options -fipa-struct-reorg and
-fipa-matrix-reorg have been removed
Interprocedural and Link-time optimization improvements
AddressSanitizer, a fast memory error detector, has been
added [*2]
A new -Wsizeof-pointer-memaccess warning has been added
G++ now supports a -std=c++1y option for experimentation
with features proposed for the next revision of the
standard, expected around 2014
Improved experimental support for the new ISO C++ standard,
C++11
A new port has been added to support AArch64
Updates for ARM, x86, MIPS, PPC/PPC64, SH, SPARC, TILE*
[*1] we should support this too!
[*2] we should look into this.
https://code.google.com/p/address-sanitizer/
initial import of GCC 4.5.3 sources. changes since 4.1 are way too numerous to review, please see http://gcc.gnu.org/gcc-4.5/changes.html (and the 4.2, 4.3 and 4.4 versions, too.) this includes the core, c++, objc and the non java/ada/fortran parts of the testsuite.
Initial revision