mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-01-31 03:27:39 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
cosmopolitan/build/definitions.mk
Justine Tunney ec480f5aa0
Make improvements 
- Every unit test now passes on Apple Silicon. The final piece of this
  puzzle was porting our POSIX threads cancelation support, since that
  works differently on ARM64 XNU vs. AMD64. Our semaphore support on
  Apple Silicon is also superior now compared to AMD64, thanks to the
  grand central dispatch library which lets *NSYNC locks go faster.

- The Cosmopolitan runtime is now more stable, particularly on Windows.
  To do this, thread local storage is mandatory at all runtime levels,
  and the innermost packages of the C library is no longer being built
  using ASAN. TLS is being bootstrapped with a 128-byte TIB during the
  process startup phase, and then later on the runtime re-allocates it
  either statically or dynamically to support code using _Thread_local.
  fork() and execve() now do a better job cooperating with threads. We
  can now check how much stack memory is left in the process or thread
  when functions like kprintf() / execve() etc. call alloca(), so that
  ENOMEM can be raised, reduce a buffer size, or just print a warning.

- POSIX signal emulation is now implemented the same way kernels do it
  with pthread_kill() and raise(). Any thread can interrupt any other
  thread, regardless of what it's doing. If it's blocked on read/write
  then the killer thread will cancel its i/o operation so that EINTR can
  be returned in the mark thread immediately. If it's doing a tight CPU
  bound operation, then that's also interrupted by the signal delivery.
  Signal delivery works now by suspending a thread and pushing context
  data structures onto its stack, and redirecting its execution to a
  trampoline function, which calls SetThreadContext(GetCurrentThread())
  when it's done.

- We're now doing a better job managing locks and handles. On NetBSD we
  now close semaphore file descriptors in forked children. Semaphores on
  Windows can now be canceled immediately, which means mutexes/condition
  variables will now go faster. Apple Silicon semaphores can be canceled
  too. We're now using Apple's pthread_yield() funciton. Apple _nocancel
  syscalls are now used on XNU when appropriate to ensure pthread_cancel
  requests aren't lost. The MbedTLS library has been updated to support
  POSIX thread cancelations. See tool/build/runitd.c for an example of
  how it can be used for production multi-threaded tls servers. Handles
  on Windows now leak less often across processes. All i/o operations on
  Windows are now overlapped, which means file pointers can no longer be
  inherited across dup() and fork() for the time being.

- We now spawn a thread on Windows to deliver SIGCHLD and wakeup wait4()
  which means, for example, that posix_spawn() now goes 3x faster. POSIX
  spawn is also now more correct. Like Musl, it's now able to report the
  failure code of execve() via a pipe although our approach favors using
  shared memory to do that on systems that have a true vfork() function.

- We now spawn a thread to deliver SIGALRM to threads when setitimer()
  is used. This enables the most precise wakeups the OS makes possible.

- The Cosmopolitan runtime now uses less memory. On NetBSD for example,
  it turned out the kernel would actually commit the PT_GNU_STACK size
  which caused RSS to be 6mb for every process. Now it's down to ~4kb.
  On Apple Silicon, we reduce the mandatory upstream thread size to the
  smallest possible size to reduce the memory overhead of Cosmo threads.
  The examples directory has a program called greenbean which can spawn
  a web server on Linux with 10,000 worker threads and have the memory
  usage of the process be ~77mb. The 1024 byte overhead of POSIX-style
  thread-local storage is now optional; it won't be allocated until the
  pthread_setspecific/getspecific functions are called. On Windows, the
  threads that get spawned which are internal to the libc implementation
  use reserve rather than commit memory, which shaves a few hundred kb.

- sigaltstack() is now supported on Windows, however it's currently not
  able to be used to handle stack overflows, since crash signals are
  still generated by WIN32. However the crash handler will still switch
  to the alt stack, which is helpful in environments with tiny threads.

- Test binaries are now smaller. Many of the mandatory dependencies of
  the test runner have been removed. This ensures many programs can do a
  better job only linking the the thing they're testing. This caused the
  test binaries for LIBC_FMT for example, to decrease from 200kb to 50kb

- long double is no longer used in the implementation details of libc,
  except in the APIs that define it. The old code that used long double
  for time (instead of struct timespec) has now been thoroughly removed.

- ShowCrashReports() is now much tinier in MODE=tiny. Instead of doing
  backtraces itself, it'll just print a command you can run on the shell
  using our new `cosmoaddr2line` program to view the backtrace.

- Crash report signal handling now works in a much better way. Instead
  of terminating the process, it now relies on SA_RESETHAND so that the
  default SIG_IGN behavior can terminate the process if necessary.

- Our pledge() functionality has now been fully ported to AARCH64 Linux.
2023-09-18 21:04:47 -07:00

383 lines
10 KiB
Makefile
Raw Blame History

#-*-mode:makefile-gmake;indent-tabs-mode:t;tab-width:8;coding:utf-8-*-┐
#───vi: set et ft=make ts=8 tw=8 fenc=utf-8 :vi───────────────────────┘
#
# SYNOPSIS
#
# Cosmopolitan Core Build Definitions
#
# DESCRIPTION
#
# Nearly all compiler flag tuning is done within this one file. Flags
# may be customized with environment variables. We also use this file
# to compute expensive values once per build.
#
# When tuning the variables below, please note they're interpreted in
# the strictest sense. For example, we don't pass CFLAGS to gcc if we
# know it's compiling a .S file. This enables our `make V=0` logging
# to be succinct and informative at the cost of being less forgiving.
#
# Further note that link order is equally unforgiving in repositories
# of this scale. We approach that by over-specifying dependencies, in
# packages that're broken down usually on a per-directory basis. This
# is aided by the checkdeps and functional programming tools which in
# most cases should be able to deduce correct ordering automatically.
#
# Concerning our approach to flag tuning, most of it is non-essential
# and largely serves to turn features off. Particularly features that
# would otherwise lock us in to a particular platform or legal terms.
# Other flags will usually be the ones that provide us marginal gains
# in terms of performance and code size, but the world won't end when
# they aren't used. Flags that play a critical role in source working
# correctly will usually be specified on a object-by-object basis, in
# their respective packages.
#
# VARIABLES
#
# Our configuration variables, ordered by increasing preference:
#
# CCFLAGS frontend flags (.i, .c, .cc, .f, .S, .lds, etc.)
# OFLAGS objectify flags (precludes -S and -E)
# CPPFLAGS preprocessor flags (.h, .c, .cc, .S, .inc, .lds, etc.)
# TARGET_ARCH microarchitecture flags (e.g. -march=native)
# COPTS c/c++ flags (.c, .cc)
# CFLAGS c flags (.c only)
# CXXFLAGS c++ flags (.cc only)
# LDFLAGS linker flags (don't use -Wl, frontend prefix)
# ASFLAGS assembler flags (don't use -Wa, frontend prefix)
#
# For each FOO above, there exists (by increasing preference)
#
# DEFAULT_FOO see build/definitions.mk
# CONFIG_FOO see build/config.mk
# FOO set ~/.cosmo.mk and target-specific
# OVERRIDE_FOO set ~/.cosmo.mk and target-specific (use rarely)
#
LC_ALL = C
SOURCE_DATE_EPOCH = 0
ARFLAGS = rcsD
ZFLAGS ?=
XARGS ?= xargs -P4 -rs8000
DOT ?= dot
CLANG = clang
TMPDIR = o/tmp
AR = build/bootstrap/ar.com
CP = build/bootstrap/cp.com
RM = build/bootstrap/rm.com -f
GZIP = build/bootstrap/gzip.com
ECHO = build/bootstrap/echo.com
CHMOD = build/bootstrap/chmod.com
TOUCH = build/bootstrap/touch.com
PKG = build/bootstrap/package.com
MKDEPS = build/bootstrap/mkdeps.com
ZIPOBJ = build/bootstrap/zipobj.com
ZIPCOPY = build/bootstrap/zipcopy.com
PECHECK = build/bootstrap/pecheck.com
FIXUPOBJ = build/bootstrap/fixupobj.com
MKDIR = build/bootstrap/mkdir.com -p
COMPILE = build/bootstrap/compile.com -V9 -P4096 $(QUOTA)
COMMA := ,
PWD := $(shell build/bootstrap/pwd.com)
IGNORE := $(shell $(MKDIR) $(TMPDIR))
ifneq ($(findstring aarch64,$(MODE)),)
ARCH = aarch64
HOSTS ?= pi silicon-wifi
else
ARCH = x86_64
HOSTS ?= freebsd rhel7 rhel5 xnu win10 openbsd netbsd
endif
ifeq ($(PREFIX),)
ifeq ($(USE_SYSTEM_TOOLCHAIN),)
ifneq ("$(wildcard o/third_party/gcc/bin/x86_64-pc-linux-gnu-*)","")
PREFIX = o/third_party/gcc/bin/x86_64-pc-linux-gnu-
else
IGNORE := $(shell build/bootstrap/unbundle.com)
PREFIX = o/third_party/gcc/bin/x86_64-linux-musl-
endif
ifeq ($(ARCH), aarch64)
PREFIX = o/third_party/gcc/bin/aarch64-linux-musl-
endif
endif
endif
AS = $(PREFIX)as
CC = $(PREFIX)gcc
CXX = $(PREFIX)g++
CXXFILT = $(PREFIX)c++filt
LD = $(PREFIX)ld.bfd
NM = $(PREFIX)nm
GCC = $(PREFIX)gcc
STRIP = $(PREFIX)strip
OBJCOPY = $(PREFIX)objcopy
OBJDUMP = $(PREFIX)objdump
ifneq ($(wildcard $(PWD)/$(PREFIX)addr2line), )
ADDR2LINE = $(PWD)/$(PREFIX)addr2line
else
ADDR2LINE = $(PREFIX)addr2line
endif
export ADDR2LINE
export LC_ALL
export MKDIR
export MODE
export SOURCE_DATE_EPOCH
export TMPDIR
ifeq ($(LANDLOCKMAKE_VERSION),)
TMPSAFE = $(join $(TMPDIR),$(subst /,_,$@)).tmp
else
TMPSAFE = $(TMPDIR)/
endif
BACKTRACES = \
-fno-optimize-sibling-calls \
-mno-omit-leaf-frame-pointer
ifneq ($(ARCH), aarch64)
BACKTRACES += -fno-schedule-insns2
endif
SANITIZER = \
-fsanitize=address
NO_MAGIC = \
-ffreestanding \
-fno-stack-protector \
-fwrapv \
-fno-sanitize=all \
-fpatchable-function-entry=0,0
OLD_CODE = \
-fno-strict-aliasing \
-fno-strict-overflow
TRADITIONAL = \
-Wno-implicit-int \
-Wno-return-type \
-Wno-pointer-sign
DEFAULT_CCFLAGS += \
-Wall \
-Werror \
-fno-omit-frame-pointer \
-frecord-gcc-switches
DEFAULT_COPTS ?= \
-fno-math-errno \
-fno-ident \
-fno-common \
-fno-gnu-unique \
-fstrict-aliasing \
-fstrict-overflow \
-fno-semantic-interposition \
-fno-dwarf2-cfi-asm \
-fno-unwind-tables \
-fno-asynchronous-unwind-tables
ifeq ($(ARCH), x86_64)
# Microsoft says "[a]ny memory below the stack beyond the red zone
# [note: Windows defines the x64 red zone size as 0] is considered
# volatile and may be modified by the operating system at any time."
# https://devblogs.microsoft.com/oldnewthing/20190111-00/?p=100685
DEFAULT_COPTS += \
-mno-red-zone \
-mno-tls-direct-seg-refs
endif
ifeq ($(ARCH), aarch64)
#
# - Apple says in "Writing ARM64 code for Apple platforms" that we're
# not allowed to use the x18 register.
#
# - Cosmopolitan Libc uses x28 for thread-local storage because Apple
# forbids us from using tpidr_el0 too.
#
# - Cosmopolitan currently lacks an implementation of the runtime
# libraries needed by the -moutline-atomics flag
#
DEFAULT_COPTS += \
-ffixed-x18 \
-ffixed-x28 \
-mno-outline-atomics
endif
MATHEMATICAL = \
-O3 \
-fwrapv
DEFAULT_CPPFLAGS += \
-D_COSMO_SOURCE \
-DMODE='"$(MODE)"' \
-nostdinc \
-iquote .
DEFAULT_CFLAGS = \
-std=gnu2x
DEFAULT_CXXFLAGS = \
-fno-rtti \
-fno-exceptions \
-fuse-cxa-atexit \
-Wno-int-in-bool-context \
-Wno-narrowing \
-Wno-literal-suffix
DEFAULT_ASFLAGS = \
-W \
-I. \
--noexecstack
DEFAULT_LDFLAGS = \
-static \
-nostdlib \
-znorelro \
--gc-sections \
--build-id=none \
--no-dynamic-linker
# # generate linker report files
# DEFAULT_LDFLAGS += --cref -Map=$@.map
ifeq ($(ARCH), aarch64)
DEFAULT_LDFLAGS += \
-zmax-page-size=0x4000 \
-zcommon-page-size=0x4000 \
-znorelro
else
DEFAULT_LDFLAGS += \
-zmax-page-size=0x4000 \
-zcommon-page-size=0x1000
endif
ASONLYFLAGS = \
-c \
-g
DEFAULT_LDLIBS =
MCA = llvm-mca-10 \
-mtriple=x86_64-pc-linux-gnu \
-instruction-info \
-iterations=3 \
-all-stats \
-all-views \
-timeline
cc.flags = \
$(DEFAULT_CCFLAGS) \
$(CONFIG_CCFLAGS) \
$(CCFLAGS) \
$(OVERRIDE_CCFLAGS)
o.flags = \
$(DEFAULT_OFLAGS) \
$(CONFIG_OFLAGS) \
$(OFLAGS) \
$(OVERRIDE_OFLAGS)
cpp.flags = \
$(DEFAULT_CPPFLAGS) \
$(CONFIG_CPPFLAGS) \
$(CPPFLAGS) \
$(OVERRIDE_CPPFLAGS) \
-include libc/integral/normalize.inc
copt.flags = \
$(TARGET_ARCH) \
$(DEFAULT_COPTS) \
$(CONFIG_COPTS) \
$(COPTS) \
$(OVERRIDE_COPTS)
f.flags = \
$(DEFAULT_FFLAGS) \
$(CONFIG_FFLAGS) \
$(FFLAGS) \
$(OVERRIDE_FFLAGS)
c.flags = \
$(DEFAULT_CFLAGS) \
$(CONFIG_CFLAGS) \
$(CFLAGS) \
$(OVERRIDE_CFLAGS)
cxx.flags = \
$(DEFAULT_CXXFLAGS) \
$(CONFIG_CXXFLAGS) \
$(CXXFLAGS) \
$(OVERRIDE_CXXFLAGS)
s.flags = \
$(DEFAULT_ASFLAGS) \
$(CONFIG_ASFLAGS) \
$(ASFLAGS) \
$(OVERRIDE_ASFLAGS)
S.flags = $(addprefix -Wa$(COMMA),$(s.flags))
LD.libs = \
$(LDLIBS) \
$(LOADLIBES) \
$(DEFAULT_LDLIBS) \
$(CONFIG_LDLIBS) \
$(LDLIBS) \
$(DEFAULT_LIBS) \
$(CONFIG_LIBS) \
$(LIBS)
COMPILE.c.flags = $(cc.flags) $(copt.flags) $(cpp.flags) $(c.flags)
COMPILE.cxx.flags = $(cc.flags) $(copt.flags) $(cpp.flags) $(cxx.flags)
COMPILE.i.flags = $(cc.flags) $(copt.flags) $(c.flags)
COMPILE.ii.flags = $(cc.flags) $(copt.flags) $(cxx.flags)
LINK.flags = $(DEFAULT_LDFLAGS) $(CONFIG_LDFLAGS) $(LDFLAGS)
OBJECTIFY.c.flags = $(cc.flags) $(o.flags) $(S.flags) $(cpp.flags) $(copt.flags) $(c.flags)
OBJECTIFY.cxx.flags = $(cc.flags) $(o.flags) $(S.flags) $(cpp.flags) $(copt.flags) $(cxx.flags)
OBJECTIFY.s.flags = $(ASONLYFLAGS) $(s.flags)
OBJECTIFY.S.flags = $(cc.flags) $(o.flags) $(S.flags) $(cpp.flags)
PREPROCESS.flags = -E $(copt.flags) $(cc.flags) $(cpp.flags)
PREPROCESS.lds.flags = -D__LINKER__ $(filter-out -g%,$(PREPROCESS.flags)) -P -xc
COMPILE.c = $(CC) -S $(COMPILE.c.flags)
COMPILE.cxx = $(CXX) -S $(COMPILE.cxx.flags)
COMPILE.i = $(CC) -S $(COMPILE.i.flags)
OBJECTIFY.s = $(AS) $(OBJECTIFY.s.flags)
OBJECTIFY.S = $(CC) $(OBJECTIFY.S.flags) -c
OBJECTIFY.c = $(CC) $(OBJECTIFY.c.flags) -c
OBJECTIFY.cxx = $(CXX) $(OBJECTIFY.cxx.flags) -c
PREPROCESS = $(CC) $(PREPROCESS.flags)
PREPROCESS.lds = $(CC) $(PREPROCESS.lds.flags)
LINK = $(LD) $(LINK.flags)
ELF = o/libc/elf/elf.lds
ELFLINK = $(COMPILE) -ALINK.elf $(LINK) $(LINKARGS) $(OUTPUT_OPTION) && $(COMPILE) -AFIXUP.ape -T$@ $(FIXUPOBJ) $@
LINKARGS = $(patsubst %.lds,-T %.lds,$(call uniqr,$(LD.libs) $(filter-out %.pkg,$^)))
# The compiler won't generate %xmm code for sources extensioned .greg.c,
# which is needed for C modules wanting to run at the executive level or
# during privileged runtime states, e.g. code morphing.
OBJECTIFY.greg.c = \
$(CC) \
$(filter-out -pg,$(OBJECTIFY.c.flags)) \
-D__MGENERAL_REGS_ONLY__ \
-mgeneral-regs-only \
-fno-stack-protector \
-fno-instrument-functions \
-fno-optimize-sibling-calls \
-fno-sanitize=all \
-ffreestanding \
-fwrapv \
-c
TAGSFLAGS = \
-e \
-a \
--if0=no \
--langmap=c:.c.h.i \
--line-directives=yes \
--exclude=libc/nt/struct/imagefileheader.internal.h \
--exclude=libc/nt/struct/filesegmentelement.h
Reference in a new issue View git blame Copy permalink