Make improvements

- Let OpenMP be usable via cosmocc
- Let libunwind be usable via cosmocc
- Make X86_HAVE(AVXVNNI) work correctly
- Avoid using MAP_GROWSDOWN on qemu-aarch64
- Introduce in6addr_any and in6addr_loopback
- Have thread stacks use MAP_GROWSDOWN by default
- Ask OpenMP to not use filesystem to manage threads
- Make NI_MAXHOST and NI_MAXSERV available w/o _GNU_SOURCE

third_party/libunwind/BUILD.mk

@@ -37,7 +37,8 @@ THIRD_PARTY_LIBUNWIND_A_SRCS_CC =					\
 THIRD_PARTY_LIBUNWIND_A_SRCS_C =					\
 	third_party/libunwind/Unwind-sjlj.c				\
 	third_party/libunwind/UnwindLevel1-gcc-ext.c			\
-	third_party/libunwind/UnwindLevel1.c
+	third_party/libunwind/UnwindLevel1.c				\
+	third_party/libunwind/gcc_personality_v0.c
 
 THIRD_PARTY_LIBUNWIND_A_SRCS =						\
 	$(THIRD_PARTY_LIBUNWIND_A_SRCS_C)				\
@@ -70,6 +71,7 @@ $(THIRD_PARTY_LIBUNWIND_A).pkg:						\
 
 $(THIRD_PARTY_LIBUNWIND_A_OBJS): private				\
 		CFLAGS +=						\
+			-fexceptions					\
 			-fno-sanitize=all				\
 			-ffunction-sections				\
 			-fdata-sections					\
@@ -77,6 +79,7 @@ $(THIRD_PARTY_LIBUNWIND_A_OBJS): private				\
 
 $(THIRD_PARTY_LIBUNWIND_A_OBJS): private				\
 		CXXFLAGS +=						\
+			-fexceptions					\
 			-fno-sanitize=all				\
 			-ffunction-sections				\
 			-fdata-sections					\

third_party/libunwind/gcc_personality_v0.c

@@ -0,0 +1,236 @@
+//===-- gcc_personality_v0.c - Implement __gcc_personality_v0 -------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "third_party/libunwind/include/unwind.h"
+#include "third_party/compiler_rt/int_lib.h"
+
+// Pointer encodings documented at:
+//   http://refspecs.freestandards.org/LSB_1.3.0/gLSB/gLSB/ehframehdr.html
+
+#define DW_EH_PE_omit 0xff // no data follows
+
+#define DW_EH_PE_absptr 0x00
+#define DW_EH_PE_uleb128 0x01
+#define DW_EH_PE_udata2 0x02
+#define DW_EH_PE_udata4 0x03
+#define DW_EH_PE_udata8 0x04
+#define DW_EH_PE_sleb128 0x09
+#define DW_EH_PE_sdata2 0x0A
+#define DW_EH_PE_sdata4 0x0B
+#define DW_EH_PE_sdata8 0x0C
+
+#define DW_EH_PE_pcrel 0x10
+#define DW_EH_PE_textrel 0x20
+#define DW_EH_PE_datarel 0x30
+#define DW_EH_PE_funcrel 0x40
+#define DW_EH_PE_aligned 0x50
+#define DW_EH_PE_indirect 0x80 // gcc extension
+
+// read a uleb128 encoded value and advance pointer
+static size_t readULEB128(const uint8_t **data) {
+  size_t result = 0;
+  size_t shift = 0;
+  unsigned char byte;
+  const uint8_t *p = *data;
+  do {
+    byte = *p++;
+    result |= (byte & 0x7f) << shift;
+    shift += 7;
+  } while (byte & 0x80);
+  *data = p;
+  return result;
+}
+
+// read a pointer encoded value and advance pointer
+static uintptr_t readEncodedPointer(const uint8_t **data, uint8_t encoding) {
+  const uint8_t *p = *data;
+  uintptr_t result = 0;
+
+  if (encoding == DW_EH_PE_omit)
+    return 0;
+
+  // first get value
+  switch (encoding & 0x0F) {
+  case DW_EH_PE_absptr:
+    result = *((const uintptr_t *)p);
+    p += sizeof(uintptr_t);
+    break;
+  case DW_EH_PE_uleb128:
+    result = readULEB128(&p);
+    break;
+  case DW_EH_PE_udata2:
+    result = *((const uint16_t *)p);
+    p += sizeof(uint16_t);
+    break;
+  case DW_EH_PE_udata4:
+    result = *((const uint32_t *)p);
+    p += sizeof(uint32_t);
+    break;
+  case DW_EH_PE_udata8:
+    result = *((const uint64_t *)p);
+    p += sizeof(uint64_t);
+    break;
+  case DW_EH_PE_sdata2:
+    result = *((const int16_t *)p);
+    p += sizeof(int16_t);
+    break;
+  case DW_EH_PE_sdata4:
+    result = *((const int32_t *)p);
+    p += sizeof(int32_t);
+    break;
+  case DW_EH_PE_sdata8:
+    result = *((const int64_t *)p);
+    p += sizeof(int64_t);
+    break;
+  case DW_EH_PE_sleb128:
+  default:
+    // not supported
+    compilerrt_abort();
+    break;
+  }
+
+  // then add relative offset
+  switch (encoding & 0x70) {
+  case DW_EH_PE_absptr:
+    // do nothing
+    break;
+  case DW_EH_PE_pcrel:
+    result += (uintptr_t)(*data);
+    break;
+  case DW_EH_PE_textrel:
+  case DW_EH_PE_datarel:
+  case DW_EH_PE_funcrel:
+  case DW_EH_PE_aligned:
+  default:
+    // not supported
+    compilerrt_abort();
+    break;
+  }
+
+  // then apply indirection
+  if (encoding & DW_EH_PE_indirect) {
+    result = *((const uintptr_t *)result);
+  }
+
+  *data = p;
+  return result;
+}
+
+#if defined(__arm__) && !defined(__USING_SJLJ_EXCEPTIONS__) &&                 \
+    !defined(__ARM_DWARF_EH__) && !defined(__SEH__)
+#define USING_ARM_EHABI 1
+_Unwind_Reason_Code __gnu_unwind_frame(struct _Unwind_Exception *,
+                                       struct _Unwind_Context *);
+#endif
+
+static inline _Unwind_Reason_Code
+continueUnwind(struct _Unwind_Exception *exceptionObject,
+               struct _Unwind_Context *context) {
+#if USING_ARM_EHABI
+  // On ARM EHABI the personality routine is responsible for actually
+  // unwinding a single stack frame before returning (ARM EHABI Sec. 6.1).
+  if (__gnu_unwind_frame(exceptionObject, context) != _URC_OK)
+    return _URC_FAILURE;
+#endif
+  return _URC_CONTINUE_UNWIND;
+}
+
+// The C compiler makes references to __gcc_personality_v0 in
+// the dwarf unwind information for translation units that use
+// __attribute__((cleanup(xx))) on local variables.
+// This personality routine is called by the system unwinder
+// on each frame as the stack is unwound during a C++ exception
+// throw through a C function compiled with -fexceptions.
+#if __USING_SJLJ_EXCEPTIONS__
+// the setjump-longjump based exceptions personality routine has a
+// different name
+COMPILER_RT_ABI _Unwind_Reason_Code __gcc_personality_sj0(
+    int version, _Unwind_Action actions, uint64_t exceptionClass,
+    struct _Unwind_Exception *exceptionObject, struct _Unwind_Context *context)
+#elif USING_ARM_EHABI
+// The ARM EHABI personality routine has a different signature.
+COMPILER_RT_ABI _Unwind_Reason_Code __gcc_personality_v0(
+    _Unwind_State state, struct _Unwind_Exception *exceptionObject,
+    struct _Unwind_Context *context)
+#elif defined(__SEH__)
+static _Unwind_Reason_Code __gcc_personality_imp(
+    int version, _Unwind_Action actions, uint64_t exceptionClass,
+    struct _Unwind_Exception *exceptionObject, struct _Unwind_Context *context)
+#else
+COMPILER_RT_ABI _Unwind_Reason_Code __gcc_personality_v0(
+    int version, _Unwind_Action actions, uint64_t exceptionClass,
+    struct _Unwind_Exception *exceptionObject, struct _Unwind_Context *context)
+#endif
+{
+  // Since C does not have catch clauses, there is nothing to do during
+  // phase 1 (the search phase).
+#if USING_ARM_EHABI
+  // After resuming from a cleanup we should also continue on to the next
+  // frame straight away.
+  if ((state & _US_ACTION_MASK) != _US_UNWIND_FRAME_STARTING)
+#else
+  if (actions & _UA_SEARCH_PHASE)
+#endif
+    return continueUnwind(exceptionObject, context);
+
+  // There is nothing to do if there is no LSDA for this frame.
+  const uint8_t *lsda = (uint8_t *)_Unwind_GetLanguageSpecificData(context);
+  if (lsda == (uint8_t *)0)
+    return continueUnwind(exceptionObject, context);
+
+  uintptr_t pc = (uintptr_t)_Unwind_GetIP(context) - 1;
+  uintptr_t funcStart = (uintptr_t)_Unwind_GetRegionStart(context);
+  uintptr_t pcOffset = pc - funcStart;
+
+  // Parse LSDA header.
+  uint8_t lpStartEncoding = *lsda++;
+  if (lpStartEncoding != DW_EH_PE_omit) {
+    readEncodedPointer(&lsda, lpStartEncoding);
+  }
+  uint8_t ttypeEncoding = *lsda++;
+  if (ttypeEncoding != DW_EH_PE_omit) {
+    readULEB128(&lsda);
+  }
+  // Walk call-site table looking for range that includes current PC.
+  uint8_t callSiteEncoding = *lsda++;
+  size_t callSiteTableLength = readULEB128(&lsda);
+  const uint8_t *callSiteTableStart = lsda;
+  const uint8_t *callSiteTableEnd = callSiteTableStart + callSiteTableLength;
+  const uint8_t *p = callSiteTableStart;
+  while (p < callSiteTableEnd) {
+    uintptr_t start = readEncodedPointer(&p, callSiteEncoding);
+    size_t length = readEncodedPointer(&p, callSiteEncoding);
+    size_t landingPad = readEncodedPointer(&p, callSiteEncoding);
+    readULEB128(&p); // action value not used for C code
+    if (landingPad == 0)
+      continue; // no landing pad for this entry
+    if ((start <= pcOffset) && (pcOffset < (start + length))) {
+      // Found landing pad for the PC.
+      // Set Instruction Pointer to so we re-enter function
+      // at landing pad. The landing pad is created by the compiler
+      // to take two parameters in registers.
+      _Unwind_SetGR(context, __builtin_eh_return_data_regno(0),
+                    (uintptr_t)exceptionObject);
+      _Unwind_SetGR(context, __builtin_eh_return_data_regno(1), 0);
+      _Unwind_SetIP(context, (funcStart + landingPad));
+      return _URC_INSTALL_CONTEXT;
+    }
+  }
+
+  // No landing pad found, continue unwinding.
+  return continueUnwind(exceptionObject, context);
+}
+
+#if defined(__SEH__) && !defined(__USING_SJLJ_EXCEPTIONS__)
+COMPILER_RT_ABI EXCEPTION_DISPOSITION
+__gcc_personality_seh0(PEXCEPTION_RECORD ms_exc, void *this_frame,
+                       PCONTEXT ms_orig_context, PDISPATCHER_CONTEXT ms_disp) {
+  return _GCC_specific_handler(ms_exc, this_frame, ms_orig_context, ms_disp,
+                               __gcc_personality_imp);
+}
+#endif

third_party/musl/netdb.h

@@ -99,6 +99,9 @@ struct protoent *getprotoent (void);
 struct protoent *getprotobyname (const char *);
 struct protoent *getprotobynumber (int);
 
+#define NI_MAXHOST 255
+#define NI_MAXSERV 32
+
 #if defined(_COSMO_SOURCE) || defined(_GNU_SOURCE) || defined(_BSD_SOURCE) || defined(_POSIX_SOURCE) \
  || (defined(_POSIX_C_SOURCE) && _POSIX_C_SOURCE+0 < 200809L) \
  || (defined(_XOPEN_SOURCE) && _XOPEN_SOURCE+0 < 700)
@@ -130,8 +133,6 @@ int getservbyname_r(const char *, const char *, struct servent *, char *, size_t
 #define EAI_ALLDONE    -103
 #define EAI_INTR       -104
 #define EAI_IDN_ENCODE -105
-#define NI_MAXHOST 255
-#define NI_MAXSERV 32
 #endif
 
 COSMOPOLITAN_C_END_

third_party/openmp/README.cosmo

@@ -14,3 +14,4 @@ LOCAL CHANGES
   - Ran third_party/openmp/generate.sh
   - Removed usage of syscall() function
   - Turned off quad floating point support (why does openmp have it?)
+  - Remove bloat for checking if multiple OpenMP libraries are linked

third_party/openmp/kmp_config.h

@@ -18,12 +18,14 @@
 #include "libc/stdio/syscall.h"
 #endif
 
+#if IsModeDbg()
+#define KMP_DEBUG 1
+#endif
+
 #define KMP_USE_FUTEX 0
 #define KMP_FTN_ENTRIES KMP_FTN_PLAIN
 #define syscall {{openmp_shall_not_use_syscall}}
 
-#define DEBUG_BUILD IsModeDbg()
-#define RELWITHDEBINFO_BUILD (IsOptimized() && !IsTiny())
 #define LIBOMP_USE_ITT_NOTIFY 0
 #define USE_ITT_NOTIFY LIBOMP_USE_ITT_NOTIFY
 #if ! LIBOMP_USE_ITT_NOTIFY
@@ -152,9 +154,6 @@
 #if STUBS_LIBRARY
 # define KMP_STUB 1
 #endif
-#if DEBUG_BUILD || RELWITHDEBINFO_BUILD
-# define KMP_DEBUG 1
-#endif
 
 #if KMP_OS_WINDOWS
 # define KMP_WIN_CDECL
@@ -169,8 +168,4 @@
 #define KMP_USE_SHM
 #endif
 
-#ifdef __COSMOPOLITAN__
-#define KMP_USE_SHM
-#endif
-
 #endif // KMP_CONFIG_H

third_party/openmp/kmp_runtime.cpp

@@ -6699,10 +6699,12 @@ void __kmp_internal_end_thread(int gtid_req) {
 // -----------------------------------------------------------------------------
 // Library registration stuff.
 
+#ifndef __COSMOPOLITAN__
 static long __kmp_registration_flag = 0;
 // Random value used to indicate library initialization.
 static char *__kmp_registration_str = NULL;
 // Value to be saved in env var __KMP_REGISTERED_LIB_<pid>.
+#endif // __COSMOPOLITAN__
 
 static inline char *__kmp_reg_status_name() {
 /* On RHEL 3u5 if linked statically, getpid() returns different values in
@@ -6726,6 +6728,7 @@ char *temp_reg_status_file_name = nullptr;
 #endif
 
 void __kmp_register_library_startup(void) {
+#ifndef __COSMOPOLITAN__
 
   char *name = __kmp_reg_status_name(); // Name of the environment variable.
   int done = 0;
@@ -6937,9 +6940,11 @@ void __kmp_register_library_startup(void) {
   } // while
   KMP_INTERNAL_FREE((void *)name);
 
+#endif // __COSMOPOLITAN__
 } // func __kmp_register_library_startup
 
 void __kmp_unregister_library(void) {
+#ifndef __COSMOPOLITAN__
 
   char *name = __kmp_reg_status_name();
   char *value = NULL;
@@ -7006,6 +7011,7 @@ void __kmp_unregister_library(void) {
   __kmp_registration_flag = 0;
   __kmp_registration_str = NULL;
 
+#endif // __COSMOPOLITAN__
 } // __kmp_unregister_library
 
 // End of Library registration stuff.

third_party/openmp/util1.cpp

@@ -2054,6 +2054,7 @@ void __kmp_initialize_system_tick() {
 }
 #endif
 
+#ifndef __COSMOPOLITAN__
 /* Determine whether the given address is mapped into the current address
    space. */
 
@@ -2062,12 +2063,7 @@ int __kmp_is_address_mapped(void *addr) {
   int found = 0;
   int rc;
 
-#if defined(__COSMOPOLITAN__)
-
-  (void)rc;
-  found = kisdangerous(addr);
-
-#elif KMP_OS_LINUX || KMP_OS_HURD
+#if KMP_OS_LINUX || KMP_OS_HURD
 
   /* On GNUish OSes, read the /proc/<pid>/maps pseudo-file to get all the
      address ranges mapped into the address space. */
@@ -2236,6 +2232,7 @@ int __kmp_is_address_mapped(void *addr) {
   return found;
 
 } // __kmp_is_address_mapped
+#endif // __COSMOPOLITAN__
 
 #ifdef USE_LOAD_BALANCE
 

third_party/openmp/util2.S

@@ -11,6 +11,14 @@
 ////===----------------------------------------------------------------------===//
 //
 
+#ifdef __COSMOPOLITAN__
+	.macro	no.comm name:req size:req align:req
+	.globl	\name
+\name:	.align	\align
+	.byte	\size
+	.endm
+#endif
+
 // -----------------------------------------------------------------------
 // macros
 // -----------------------------------------------------------------------
@@ -180,11 +188,11 @@ KMP_PREFIX_UNDERSCORE(\proc):
 
 .macro COMMON name, size, align_power
 #if KMP_OS_DARWIN
-	.comm \name, \size
+	no.comm \name, \size
 #elif KMP_OS_WINDOWS
-	.comm \name, \size, \align_power
+	no.comm \name, \size, \align_power
 #else // !KMP_OS_DARWIN && !KMP_OS_WINDOWS
-	.comm \name, \size, (1<<(\align_power))
+	no.comm \name, \size, (1<<(\align_power))
 #endif
 .endm
 
@@ -202,14 +210,14 @@ KMP_PREFIX_UNDERSCORE(\proc):
 # if KMP_ARCH_X86
 #  if KMP_OS_DARWIN
         .data
-        .comm .gomp_critical_user_,32
+        no.comm .gomp_critical_user_,32
         .data
         .globl ___kmp_unnamed_critical_addr
 ___kmp_unnamed_critical_addr:
         .long .gomp_critical_user_
 #  else /* Linux* OS */
         .data
-        .comm .gomp_critical_user_,32,8
+        no.comm .gomp_critical_user_,32,8
         .data
 	ALIGN 4
         .global __kmp_unnamed_critical_addr
@@ -223,21 +231,14 @@ __kmp_unnamed_critical_addr:
 # if KMP_ARCH_X86_64
 #  if KMP_OS_DARWIN
         .data
-        .comm .gomp_critical_user_,32
+        no.comm .gomp_critical_user_,32
         .data
         .globl ___kmp_unnamed_critical_addr
 ___kmp_unnamed_critical_addr:
         .quad .gomp_critical_user_
 #  else /* Linux* OS */
         .data
-#ifdef __COSMOPOLITAN__
-	.globl	.gomp_critical_user_
-".gomp_critical_user_":
-	.align	8
-	.byte	32
-#else
-        .comm .gomp_critical_user_,32,8
-#endif
+        no.comm .gomp_critical_user_,32,8
         .data
 	ALIGN 8
         .global __kmp_unnamed_critical_addr