mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-06-28 15:28:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Eliminate cyclic locks in runtime

Browse source 
This change introduces a new deadlock detector for Cosmo's POSIX threads
implementation. Error check mutexes will now track a DAG of nested locks
and report EDEADLK when a deadlock is theoretically possible. These will
occur rarely, but it's important for production hardening your code. You
don't even need to change your mutexes to use the POSIX error check mode
because `cosmocc -mdbg` will enable error checking on mutexes by default
globally. When cycles are found, an error message showing your demangled
symbols describing the strongly connected component are printed and then
the SIGTRAP is raised, which means you'll also get a backtrace if you're
using ShowCrashReports() too. This new error checker is so low-level and
so pure that it's able to verify the relationships of every libc runtime
lock, including those locks upon which the mutex implementation depends.
This commit is contained in:
Justine Tunney 2024-12-16 20:51:27 -08:00
parent 26c051c297
commit af7bd80430
No known key found for this signature in database
GPG key ID: BE714B4575D6E328
141 changed files with 2094 additions and 1601 deletions
Download patch file Download diff file Expand all files Collapse all files

139
libc/intrin/bzero.c
View file

@ -16,155 +16,18 @@
│ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR │
│ PERFORMANCE OF THIS SOFTWARE. │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "libc/assert.h"
#include "libc/dce.h"
#include "libc/nexgen32e/nexgen32e.h"
#include "libc/nexgen32e/x86feature.h"
#include "libc/str/str.h"
typedef char xmm_t __attribute__((__vector_size__(16), __aligned__(1)));
typedef long long xmm_a __attribute__((__vector_size__(16), __aligned__(16)));
static void bzero128(char *p, size_t n) {
xmm_t v = {0};
if (n <= 32) {
*(xmm_t *)(p + n - 16) = v;
*(xmm_t *)p = v;
} else {
do {
n -= 32;
*(xmm_t *)(p + n) = v;
*(xmm_t *)(p + n + 16) = v;
} while (n > 32);
*(xmm_t *)(p + 16) = v;
*(xmm_t *)p = v;
}
}
#if defined(__x86_64__) && !defined(__chibicc__)
_Microarchitecture("avx") static void bzero_avx(char *p, size_t n) {
xmm_t v = {0};
if (n <= 32) {
*(xmm_t *)(p + n - 16) = v;
*(xmm_t *)p = v;
} else if (n >= 1024 && X86_HAVE(ERMS)) {
asm("rep stosb" : "+D"(p), "+c"(n), "=m"(*(char(*)[n])p) : "a"(0));
} else {
if (n < kHalfCache3 || !kHalfCache3) {
do {
n -= 32;
*(xmm_t *)(p + n) = v;
*(xmm_t *)(p + n + 16) = v;
} while (n > 32);
} else {
while ((uintptr_t)(p + n) & 15) {
p[--n] = 0;
}
do {
n -= 32;
__builtin_ia32_movntdq((xmm_a *)(p + n), (xmm_a)v);
__builtin_ia32_movntdq((xmm_a *)(p + n + 16), (xmm_a)v);
} while (n > 32);
asm("sfence");
}
*(xmm_t *)(p + 16) = v;
*(xmm_t *)p = v;
}
}
#endif
/**
* Sets memory to zero.
*
* bzero n=0 661 picoseconds
* bzero n=1 661 ps/byte 1,476 mb/s
* bzero n=2 330 ps/byte 2,952 mb/s
* bzero n=3 220 ps/byte 4,428 mb/s
* bzero n=4 165 ps/byte 5,904 mb/s
* bzero n=7 94 ps/byte 10,333 mb/s
* bzero n=8 41 ps/byte 23,618 mb/s
* bzero n=15 44 ps/byte 22,142 mb/s
* bzero n=16 20 ps/byte 47,236 mb/s
* bzero n=31 21 ps/byte 45,760 mb/s
* bzero n=32 20 ps/byte 47,236 mb/s
* bzero n=63 10 ps/byte 92,997 mb/s
* bzero n=64 15 ps/byte 62,982 mb/s
* bzero n=127 15 ps/byte 62,490 mb/s
* bzero n=128 10 ps/byte 94,473 mb/s
* bzero n=255 14 ps/byte 68,439 mb/s
* bzero n=256 9 ps/byte 105 gb/s
* bzero n=511 15 ps/byte 62,859 mb/s
* bzero n=512 11 ps/byte 83,976 mb/s
* bzero n=1023 15 ps/byte 61,636 mb/s
* bzero n=1024 10 ps/byte 88,916 mb/s
* bzero n=2047 9 ps/byte 105 gb/s
* bzero n=2048 8 ps/byte 109 gb/s
* bzero n=4095 8 ps/byte 115 gb/s
* bzero n=4096 8 ps/byte 118 gb/s
* bzero n=8191 7 ps/byte 129 gb/s
* bzero n=8192 7 ps/byte 130 gb/s
* bzero n=16383 6 ps/byte 136 gb/s
* bzero n=16384 6 ps/byte 137 gb/s
* bzero n=32767 6 ps/byte 140 gb/s
* bzero n=32768 6 ps/byte 141 gb/s
* bzero n=65535 15 ps/byte 64,257 mb/s
* bzero n=65536 15 ps/byte 64,279 mb/s
* bzero n=131071 15 ps/byte 63,166 mb/s
* bzero n=131072 15 ps/byte 63,115 mb/s
* bzero n=262143 15 ps/byte 62,052 mb/s
* bzero n=262144 15 ps/byte 62,097 mb/s
* bzero n=524287 15 ps/byte 61,699 mb/s
* bzero n=524288 15 ps/byte 61,674 mb/s
* bzero n=1048575 16 ps/byte 60,179 mb/s
* bzero n=1048576 15 ps/byte 61,330 mb/s
* bzero n=2097151 15 ps/byte 61,071 mb/s
* bzero n=2097152 15 ps/byte 61,065 mb/s
* bzero n=4194303 16 ps/byte 60,942 mb/s
* bzero n=4194304 16 ps/byte 60,947 mb/s
* bzero n=8388607 16 ps/byte 60,872 mb/s
* bzero n=8388608 16 ps/byte 60,879 mb/s
*
* @param p is memory address
* @param n is byte length
* @return p
* @asyncsignalsafe
*/
void bzero(void *p, size_t n) {
char *b;
uint64_t x;
b = p;
#ifdef __x86_64__
asm("xorl\t%k0,%k0" : "=r"(x));
#else
if (1) {
memset(p, 0, n);
return;
}
x = 0;
#endif
if (n <= 16) {
if (n >= 8) {
__builtin_memcpy(b, &x, 8);
__builtin_memcpy(b + n - 8, &x, 8);
} else if (n >= 4) {
__builtin_memcpy(b, &x, 4);
__builtin_memcpy(b + n - 4, &x, 4);
} else if (n) {
do {
asm volatile("" ::: "memory");
b[--n] = x;
} while (n);
}
#if defined(__x86_64__) && !defined(__chibicc__)
} else if (IsTiny()) {
asm("rep stosb" : "+D"(b), "+c"(n), "=m"(*(char(*)[n])b) : "a"(0));
return;
} else if (X86_HAVE(AVX)) {
bzero_avx(b, n);
#endif
} else {
bzero128(b, n);
}
memset(p, 0, n);
}
__weak_reference(bzero, explicit_bzero);