m68000 is broken due to missing gmp (MKGCCCMDS = no).
clang build for everyone is also broken for the same reason.
We must reconsider to handle gmp in src/lib/Makefile.
Note that these two funcs, nbsd_ilp32_solib_svr4_fetch_link_map_offsets,
and nbsd_lp64_solib_svr4_fetch_link_map_offsets, merely fall back to
default ELF functions.
Not working yet:
----
Reading symbols from cat...
Reading symbols from /usr/libdata/debug//bin/cat.debug...
warning: Couldn't find general-purpose registers in core file.
Core was generated by `cat'.
Program terminated with signal SIGQUIT, Quit.
warning: Couldn't find general-purpose registers in core file.
#0 <unavailable> in ?? ()
(gdb)
For gdb 13, _init_foo_bar functions are registered into init.c by
newly-introduced ``make-init-c'' script. It does not recognize
functions declaration with void argument. As a result, these
constructors are not called for aarch64 and ia64.
XXX
mknative later
- vax builds with lots of newly added -O0 workarounds to avoid
ICE (not committed yet).
- On simh, it successfully boots into multi-user mode.
- C++ exception handling is broken. For GCC 10.5.0 and priors,
we have local change for libunwind:
http://www.execsw.org/netbsd/changeset.cgi?id=20140318T234341Z.c59ac4642e7d4b345f6dc5aa9887e9e1f9aa459a#src/gnu/dist/gcc4/gcc/config/vax/vax.c
But, mechanically applying it (adjusted-for-GCC-10.5.0 version)
does not help.
- Native gcc cannot compile hello.c due to ICE, even without
any optimization.
- Sanitizers do not build yet.
Otherwise, wired relocation occur for readonly section `.eh_frame',
by which libc.so cannot link:
----
warning: dynamic relocation to `?' in readonly section `.eh_frame'
warning: creating DT_TEXTREL in a shared object
.eh_frame_hdr refers to overlapping FDEs
final link failed: bad value
----
This workaround has been introduced by vax.c, v1.15:
http://cvsweb.netbsd.org/bsdweb.cgi/src/external/gpl3/gcc/dist/gcc/config/vax/Attic/vax.c#rev1.15
but accidentally lost during GCC 12.3.0 merge.
Also for ia64, riscv64, sh3e{bl}, and sparc for sure, but
nothing changed.
XXX
mknative adds something like
-I/usr/src/tools/gdb/obj.machine-arch/.native/include
to include path. This should be dropped...
earmv5{,eb}:
- No new regression for full ATF.
m68k:
- blake2_prov.c needs -fno-stack-protector.
- No new regression for full ATF on amiga.
sh3el:
- In addition to kernel, userland is completely broken.
- No interesting diff with upstream in gcc/config/sh.
- Probably, we need bisectioning for GCC...
this introduces 4 new warning disable flags:
CC_WNO_MISSING_TEMPLATE_KEYWORD
CC_WNO_REGISTER
CC_WNO_STRINGOP_OVERREAD
CC_WNO_ARRAY_BOUNDS
and documents them in README.warnings. of these, the string op
and array bounds are both problematic (real bugs) and also spurious
(not real bugs), and the other 2 are mostly temporary for older
3rd party code.
add some new uses of CC_WNO_STRINGOP_OVERFLOW.
fix m68k build for gallium and GCC 12.
aarch64{,eb}:
- Everything (without x11 yet) works just fine, except for libtsan
(not a regression, this is newly introduced to aarch64 for GCC12).
earmv5{,eb}:
- As far as I can see, everything (nox11) seems OK.
- Full ATF is on-going.
earmv[67]hf{,eb}:
- Everything (nox11) works just fine.
i386:
- lib*san requires 64-bit atomic ops. If they are built with -march=i586,
all tests in ATF pass. Alternatively, we can probably use 32-bit atomic
variables instead, with some restrictions (now examining).
- New kernel freeze observed during full ATF run.
m68k:
- Another -fstack-protector breakage.
- Full ATF is on-going for amiga.
powerpc:
- Kernel is just working for all subarchs: oea, booke, ibm4xx (405 and
strictly-alignment 403).
sh3el:
- Kernel is broken [20]. Seems like regressions for function inlining;
see also [8] for userland. But, -fno-inline does not work unfortunately...
- Kernel works just fine for sun2. Once-observed error was due to
misc/57562.
- Kernel and bootloader work just fine for amiga.
- Userland needs COPTS.strsuftoll.c+=-fno-stack-protector.
- For amiga, no regression for tests/lib/libc/{sys,gen}.
- 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.
- arm64 mostly works with some uncommited patches
- arm might work, but my testbox seems unstable
- most ports have basically working native-gcc now
- mipsel builds (pmax)
- for usr.bin, build libcody and link it into cc1*plus
- move makefile code for auto-build.h and version.h (was gcov-iov.h)
into makefile fragment to share them. (XXX probably best to only do
this in one place, but that's a future step.)
- many updates to handle the .c -> .cc renames
- build some x86 files with -msse2
- bump shlib versions for libgomp and libstdc++v3
- fix MKGCCCMDS=no builds
- updates for new rtl-ssa subdir
- use -std=gnu++11 for various places
- define TARGET_MACHINE and HOST_MACHINE for some places
- avoid a .BEGIN rule that writes to missing obj (ie src) sometimes
cleanups still needed:
- remove a bunch of commented -Wno* that don't seem to be needed anymore
- leave or remove various vax-specific -O0s that should no longer be
necessary for the vax-gcc port
- some "XXXGCC12" that should be may be fixable or converted to using our
standard ${CC_...} warnings (but may need to remain, in which case the
XXXGCC12 marker will be removed.)
with these, and a few other changes around the tree, i've able to build
gcc12 releases for several ports that seem to work (arm64, sparc, amd64,
i386 at least so far, have run, several others have built.)
- most platforms native-gcc works
- some platforms libgcc works
- amd64 kernels work
- i386 kernels build
- add info about libsanitizer re-port (all no.)
May 27th, 2023: GDB 13.2 Released!
The latest version of GDB, version 13.2, is available for download.
This is a minor corrective release over GDB 13.1, fixing the following issues:
PR testsuite/30158 (rustc testsuite fails with 13.1, apparently
worked before with trunk 20230114 on i686-linux-gnu and
powerpc64le-linux-gnu)
PR gdb/30214 (GDB 13.1 does not compile on FreeBSD 13.1)
PR gdb/30240 ((linux/aarch) thread.c:86: internal-error: inferior_thread:
Assertion `current_thread_ != nullptr' failed)
PR gdb/30249 ([13 regression] hookpost-extended-remote will not work)
PR exp/30271 (Addresses of static thread_local fields are badly
calculated sometimes)
PR symtab/30357 (Segmentation fault for the 'start' command)
PR symtab/30369 ([gdb/symtab] False match issue in
skip_prologue_using_linetable)
PR gdb/30423 (Build failures with clang 16)
PR build/30450 (Build failure (linux-low.cc:5393:45: error: expected
':' before ')' token) with musl-1.2.4)
See the NEWS file for a more complete and detailed list of what this release
includes.
Feb 19th, 2023: GDB 13.1 Released!
The latest version of GDB, version 13.1, is available for download.
This version of GDB includes the following changes and enhancements:
Support for the following new targets has been added in both GDB and GDBserver:
GNU/Linux/LoongArch (gdbserver) loongarch*-*-linux*
GNU/Linux/CSKY (gdbserver) csky*-*linux*
The Windows native target now supports target async.
FreeBSD:
Arm and AArch64: Support for Thread Local Storage (TLS) variables
Hardware watchpoint support on AArch64 FreeBSD
Floating-point support has now been added on LoongArch GNU/Linux.
New commands:
set print nibbles [on|off]
show print nibbles
This controls whether the 'print/t' command will display binary
values in groups of four bits, known as "nibbles". The default is
'off'.
Various styling-related commands. See the gdb/NEWS file for more details.
Various maintenance commands. These are normally aimed at GDB
experts or developers. See the gdb/NEWS file for more details.
Python API improvements:
New Python API for instruction disassembly.
The new attribute 'locations' of gdb.Breakpoint returns a list of
gdb.BreakpointLocation objects specifying the locations where the
breakpoint is inserted into the debuggee.
New Python type gdb.BreakpointLocation.
New function gdb.format_address(ADDRESS, PROGSPACE, ARCHITECTURE)
that formats ADDRESS as 'address '
New function gdb.current_language that returns the name of the
current language. Unlike gdb.parameter('language'), this will never
return 'auto'.
New function gdb.print_options that returns a dictionary of the
prevailing print options, in the form accepted by gdb.Value.format_string.
New method gdb.Frame.language that returns the name of the frame's language.
gdb.Value.format_string now uses the format provided by 'print',
if it is called during a 'print' or other similar operation.
gdb.Value.format_string now accepts the 'summary' keyword. This
can be used to request a shorter representation of a value, the
way that 'set print frame-arguments scalars' does.
The gdb.register_window_type method now restricts the set of
acceptable window names. The first character of a window's name
must start with a character in the set [a-zA-Z], every subsequent
character of a window's name must be in the set [-_.a-zA-Z0-9].
` GDB/MI changes: MI version 1 is deprecated, and will be removed
in GDB 14. The async record stating the stopped reason 'breakpoint-hit'
now contains an optional field locno.
Miscellaneous improvements:
gdb now supports zstd compressed debug sections (ELFCOMPRESS_ZSTD) for ELF.
New convenience variable $_inferior_thread_count contains the number
of live threads in the current inferior.
New convenience variables $_hit_bpnum and $_hit_locno, set to the
breakpoint number and the breakpoint location number of the breakpoint
last hit.
The "info breakpoints" now displays enabled breakpoint locations
of disabled breakpoints as in the "y-" state.
The format of 'disassemble /r' and 'record instruction-history /r'
has changed to match the layout of GNU objdump when disassembling.
A new format "/b" has been introduce to provide the old behavior of "/r".
The TUI no longer styles the source and assembly code highlighted
by the current position indicator by default. You can however
re-enable styling using the new "set style tui-current-position"
command.
It is now possible to use the "document" command to document
user-defined commands.
Support for memory tag data for AArch64 MTE.
Support Removal notices:
DBX mode has been removed.
Support for building against Python version 2 has been removed. It
is now only possible to build GDB against Python 3.
Support for the following commands has been removed:
set debug aix-solib on|off
show debug aix-solib
set debug solib-frv on|off
show debug solib-frv
Use the "set/show debug solib" commands instead.
See the NEWS file for a more complete and detailed list of what
this release includes.
Dec 18th, 2022: GDB 13 branch created
The GDB 13 branch (gdb-13-branch) has been created. To check out
a copy of the branch use:
git clone --branch gdb-13-branch https://sourceware.org/git/binutils-gdb.git
May 1st, 2022: GDB 12.1 Released!
The latest version of GDB, version 12.1, is available for download.
This version of GDB includes the following changes and enhancements:
New support for the following native configuration:
GNU/Linux/OpenRISC or1k*-*-linux*
New support for the following targets:
GNU/Linux/LoongArch loongarch*-*-linux*
New GDBserver support on the following configuration:
GNU/Linux/OpenRISC or1k*-*-linux*
Support for the following target has been removed:
S+core score-*-*
Multithreaded symbol loading is now enabled by default
Deprecation Notices:
GDB 12 is the last release of GDB that will support building against Python 2
DBX mode is deprecated, and will be removed in GDB 13
GDB/MI changes:
The '-add-inferior' with no option flags now inherits the connection
of the current inferior, this restores the behaviour of GDB as it
was prior to GDB 10.
The '-add-inferior' command now accepts a '--no-connection' option,
which causes the new inferior to start without a connection.
Python API enhancements:
It is now possible to add GDB/MI commands implemented in Python
New function gdb.Architecture.integer_type()
New gdb.events.gdb_exiting event
New 'gdb.events.connection_removed' event registry
New gdb.TargetConnection object
New gdb.Inferior.connection property
New read-only attribute gdb.InferiorThread.details
New gdb.RemoteTargetConnection.send_packet method
New read-only attributes gdb.Type.is_scalar and gdb.Type.is_signed
The gdb.Value.format_string method now takes a 'styling' argument
Various new function in the "gdb" module
Miscellaneous:
The FreeBSD native target now supports async mode
Improved C++ template support
Support for disabling source highlighting through GNU of the Pygments
library instead.
The "print" command has been changed so as to print floating-point
values with a base-modifying formats such as "/x" to display the
underlying bytes of the value in the desired base.
The "clone-inferior" command now ensures that the TTY, CMD and ARGS
settings are copied from the original inferior to the new one. All
modifications to the environment variables done using the 'set
environment' or 'unset environment' commands are also copied to
the new inferior.
Various new commands have been introduced
See the NEWS file for a more complete and detailed list of what
this release includes.
Mar 20th, 2022: GDB 12 branch created
The GDB 12 branch (gdb-12-branch) has been created. To check out a copy of the branch use:
git clone --branch gdb-12-branch https://sourceware.org/git/binutils-gdb.git
January 16th, 2022: GDB 11.2 Released!
The latest version of GDB, version 11.2, is available for download.
This is a minor corrective release over GDB 11.1, fixing the following issues:
PR sim/28302 (gdb fails to build with glibc 2.34)
PR build/28318 (std::thread support configure check does not use CXX_DIALECT)
PR gdb/28405 (arm-none-eabi: internal-error: ptid_t
remote_target::select_thread_for_ambiguous_stop_reply(const
target_waitstatus*): Assertion `first_resumed_thread != nullptr'
failed)
PR tui/28483 ([gdb/tui] breakpoint creation not displayed)
PR build/28555 (uclibc compile failure since commit
4655f8509fd44e6efabefa373650d9982ff37fd6)
PR rust/28637 (Rust characters will be encoded using DW_ATE_UTF)
PR gdb/28758 (GDB 11 doesn't work correctly on binaries with a
SHT_RELR (.relr.dyn) section)
PR gdb/28785 (Support SHT_RELR (.relr.dyn) section)
See the NEWS file for a more complete and detailed list of what
this release includes.
September 13th, 2021: GDB 11.1 Released!
The latest version of GDB, version 11.1, is available for download.
This version of GDB includes the following changes and enhancements:
Support for ARM Symbian (arm*-*-symbianelf*) has been removed.
Building GDB now requires GMP (The GNU Multiple Precision Arithmetic Library).
New command-line options "--early-init-command" (or "-eix") and
"--early-init-eval-command" (or "-eiex")
GDB/MI Changes:
New --qualified option for the '-break-insert' and '-dprintf-insert' commands.
New --force-condition option for the '-break-insert' and
'-dprintf-insert' commands.
New --force option for the '-break-condition' command.
The '-file-list-exec-source-files' now accepts an optional regular
expression to filter the source files included in the result.
The results from '-file-list-exec-source-files' now include a
'debug-fully-read' field to indicate if the corresponding source's
debugging information has been partially read (false) or has been
fully read (true).
TUI Improvements:
Mouse actions are now supported. The mouse wheel scrolls the
appropriate window. Key combinations that do not have a specific
action on the focused window are now passed to GDB.
Python enhancements:
Inferior objects now contain a read-only 'connection_num' attribute
that gives the connection number as seen in 'info connections' and
'info inferiors'.
New method gdb.Frame.level() which returns the stack level of the frame object.
New method gdb.PendingFrame.level() which returns the stack level
of the frame object.
When hitting a catchpoint, the Python API will now emit a
gdb.BreakpointEvent rather than a gdb.StopEvent. The gdb.Breakpoint
attached to the event will have type BP_CATCHPOINT.
Python TUI windows can now receive mouse click events. If the Window
object implements the click method, it is called for each mouse
click event in this window. New setting "python ignore-environment
on|off"; if "on", causes GDB's builtin Python to ignore any
environment variable that would otherwise affect how Python behaves
(needs to be set during "early initialization" (see above). New
setting "python dont-write-bytecode auto|on|off".
Guile API enhancements:
Improved support for rvalue reference values.
New procedures for obtaining value variants: value-reference-value,
value-rvalue-reference-value and value-const-value.
New "qMemTags" and "QMemTags" remote protocol packets (associated
with Memory Tagging).
GDB will now look for the .gdbinit file in a config directory before
looking for ~/.gdbinit. The file is searched for in the following
locations: $XDG_CONFIG_HOME/gdb/gdbinit, $HOME/.config/gdb/gdbinit,
$HOME/.gdbinit. On Apple hosts the search order is instead:
$HOME/Library/Preferences/gdb/gdbinit, $HOME/.gdbinit. The "break
[...] if CONDITION" command no longer returns an error when the
condition is invalid at one or more locations. Instead, if the
condition is valid at one or more locations, the locations where
the condition is not valid are disabled.
The behavior of the "condition" command is changed to match the
new behavior of the "break" command.
Support for general memory tagging functionality (currently limited
to AArch64 MTE)
Core file debugging now supported for x86_64 Cygwin programs.
New "org.gnu.gdb.riscv.vector" feature for RISC-V targets.
GDB now supports fixed point types which are described in DWARF as
base types with a fixed-point encoding. Additionally, support for
the DW_AT_GNU_numerator and DW_AT_GNU_denominator has also been
added.
Miscellaneous:
New "startup-quietly on|off" setting; when "on", behaves the same
as passing the "-silent" option on the command line.
New "print type hex on|off" setting; when 'on', the 'ptype' command
uses hexadecimal notation to print sizes and offsets of struct
members. When 'off', decimal notation is used.
The "inferior" command, when run without argument, prints information
about the current inferior.
The "ptype" command now supports "/x" and "/d", affecting the base
used to print sizes and offsets.
The output of the "info source" has been restructured.
New "style version foreground | background | intensity" commands
to control the styling of the GDB version number.
Various debug and maintenance commands (mostly useful for the GDB
developers) See the NEWS file for a more complete and detailed list
of what this release includes.
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.
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.
Use ${CC_WNO_MAYBE_UNINITIALIZED} instead of
the older style more complex expressions.
Remove workarounds if they were for a specific
version of gcc < 10.
Rename compiler-warning-disable variables from
GCC_NO_warning
to
CC_WNO_warning
where warning is the full warning name as used by the compiler.
GCC_NO_IMPLICIT_FALLTHRU is CC_WNO_IMPLICIT_FALLTHROUGH
Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.
Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER
Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.
autoconf relies upon GNU m4 emacs-style regexs, which
don't support intervals so don't escape these.
A fix to usr.bin/m4 -g (GNU m4 emulation) is incoming.
Add local fix to AC_FUNC_ALLOCA() to include <stdlib.h> on BSD:
# elif defined(__NetBSD__) || defined(__FreeBSD__) || defined(__DragonFly__) || defined(__OpenBSD__)
# include <stdlib.h>
This change avoids regen losing this change in tools/compat/configure
that was added in rev 1.75 on 2013-03-05 of tools/compat/configure,
and subsequent commits that keep re-adding these lines back
manually (e.g., rev 1.97 on 2019-06-22 of tools/compat/configure)
XXX: Arguably this could be fixed in a separate way to allow us to
use an unmodified modern autoconf in the future ?
Define HOSTPROG_CXX before .include anything that brings in bsd.own.mk.
This ensures that HOST_DBG (etc) gets assigned before HOST_CFLAGS
and HOST_CXXFLAGS is created.
backend: .include <bsd.init.mk> much earlier, as per the other directories.
Fixes backend build when using clang++ as the host compiler (e.g., macOS),
because backend host tools are now built with -O.
Inspired by https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=255760
Note that gcc.old/Makefile.prog may be mis-used because it defines HOSTPROG_CXX
and this Makefile fragment is included after bsd.*.mk in other Makefiles,
but they seem to build ok so leaving it alone for now.
Fixes PR toolchain/57014
Use a more restrictive pattern to extract -j NNN from MAKEFLAGS
into GMAKE_J_ARGS, to avoid false positives when the source directory
has "-j" in the path (e.g "amd64-job-000012" or "src-j9999").
Previously this could pass either -"-j" or "-j BIGNUM" to gmake
and result in "vfork: Resource temporarily unavailable" failures.
PR misc/54456
The lock gets set with atomic_exchange() -> __sync_lock_test_and_set()
which sets the value to 255 instead of 1. Check for a taken lock
with "!= 0" instead of "== 1". This should work on all architectures.
Ok: Matthew Green
The lock gets set with atomic_exchange() -> __sync_lock_test_and_set()
which sets the value to 255 instead of 1. Check for a taken lock
with "!= 0" instead of "== 1". This should work on all architectures.
Ok: Matthew Green
reduces the size of gdb.debug and gdbtui.debug by 100MB each on amd64,
and about 70MB total in the debug set. (across all builds, this may
be in the order of 3-4GB in releasedir output.)
rin