mirror of
https://github.com/jart/cosmopolitan.git
synced 2025-07-21 18:10:30 +00:00
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.
This commit is contained in:
Justine Tunney
parent
c4eb838516
commit
ec480f5aa0
638 changed files with 7925 additions and 8282 deletions
|
|
@ -49,6 +49,6 @@ textwindows dontasan void WinSockInit(void) {
|
|||
NTTRACE("WSAStartup()");
|
||||
if ((rc = WSAStartup(VERSION, &kNtWsaData)) != 0 ||
|
||||
kNtWsaData.wVersion != VERSION) {
|
||||
ExitProcess(1 << 8);
|
||||
_Exit(1);
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -1,341 +0,0 @@
|
|||
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
|
||||
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
|
||||
╞══════════════════════════════════════════════════════════════════════════════╡
|
||||
│ Copyright 2022 Justine Alexandra Roberts Tunney │
|
||||
│ │
|
||||
│ Permission to use, copy, modify, and/or distribute this software for │
|
||||
│ any purpose with or without fee is hereby granted, provided that the │
|
||||
│ above copyright notice and this permission notice appear in all copies. │
|
||||
│ │
|
||||
│ THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL │
|
||||
│ WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED │
|
||||
│ WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE │
|
||||
│ AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL │
|
||||
│ DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR │
|
||||
│ PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER │
|
||||
│ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR │
|
||||
│ PERFORMANCE OF THIS SOFTWARE. │
|
||||
╚─────────────────────────────────────────────────────────────────────────────*/
|
||||
#include "libc/assert.h"
|
||||
#include "libc/calls/calls.h"
|
||||
#include "libc/calls/struct/bpf.internal.h"
|
||||
#include "libc/calls/struct/filter.internal.h"
|
||||
#include "libc/calls/struct/seccomp.internal.h"
|
||||
#include "libc/calls/struct/sigaction.h"
|
||||
#include "libc/calls/struct/sigset.h"
|
||||
#include "libc/calls/syscall_support-sysv.internal.h"
|
||||
#include "libc/dce.h"
|
||||
#include "libc/errno.h"
|
||||
#include "libc/intrin/describeflags.internal.h"
|
||||
#include "libc/intrin/kprintf.h"
|
||||
#include "libc/intrin/likely.h"
|
||||
#include "libc/macros.internal.h"
|
||||
#include "libc/runtime/runtime.h"
|
||||
#include "libc/sock/sock.h"
|
||||
#include "libc/sock/struct/msghdr.h"
|
||||
#include "libc/sock/struct/sockaddr.h"
|
||||
#include "libc/str/str.h"
|
||||
#include "libc/sysv/consts/af.h"
|
||||
#include "libc/sysv/consts/audit.h"
|
||||
#include "libc/sysv/consts/nr.h"
|
||||
#include "libc/sysv/consts/nrlinux.h"
|
||||
#include "libc/sysv/consts/pr.h"
|
||||
#include "libc/sysv/consts/ptrace.h"
|
||||
#include "libc/sysv/consts/sig.h"
|
||||
#include "libc/sysv/errfuns.h"
|
||||
#include "net/http/ip.h"
|
||||
#ifdef __x86_64__
|
||||
|
||||
#define ORIG_RAX 120
|
||||
#define RAX 80
|
||||
#define RDI 112
|
||||
#define RSI 104
|
||||
#define RDX 96
|
||||
#define R8 72
|
||||
#define R9 64
|
||||
#define __WALL 0x40000000
|
||||
|
||||
#define OFF(f) offsetof(struct seccomp_data, f)
|
||||
|
||||
#if 0
|
||||
#define DEBUG(...) kprintf(__VA_ARGS__)
|
||||
#else
|
||||
#define DEBUG(...) donothing
|
||||
#endif
|
||||
|
||||
#define ORDIE(x) \
|
||||
do { \
|
||||
if (UNLIKELY((x) == -1)) { \
|
||||
DEBUG("%s:%d: %s failed %m\n", __FILE__, __LINE__, #x); \
|
||||
notpossible; \
|
||||
} \
|
||||
} while (0)
|
||||
|
||||
static const struct sock_filter kInetBpf[] = {
|
||||
// cargo culted architecture assertion
|
||||
BPF_STMT(BPF_LD | BPF_W | BPF_ABS, OFF(arch)),
|
||||
BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, AUDIT_ARCH_X86_64, 1, 0),
|
||||
BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_KILL_PROCESS),
|
||||
// block system calls from the future
|
||||
BPF_STMT(BPF_LD + BPF_W + BPF_ABS, OFF(nr)),
|
||||
BPF_JUMP(BPF_JMP | BPF_JGE | BPF_K, __NR_linux_memfd_secret, 0, 1),
|
||||
BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_ERRNO | 38), // ENOSYS
|
||||
// only allow local and internet sockets
|
||||
BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, __NR_linux_socket, 0, 5),
|
||||
BPF_STMT(BPF_LD | BPF_W | BPF_ABS, OFF(args[0])),
|
||||
BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, 0x001, 2, 0), // AF_UNIX
|
||||
BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, 0x002, 1, 0), // AF_INET
|
||||
BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_ERRNO | 1), // EPERM
|
||||
BPF_STMT(BPF_LD | BPF_W | BPF_ABS, OFF(nr)),
|
||||
// support for these not implemented yet
|
||||
BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K, 0x133, 0, 1), // sendmmsg
|
||||
BPF_STMT(BPF_RET + BPF_K, SECCOMP_RET_ERRNO | 1), // EPERM
|
||||
// trace syscalls with struct sockaddr
|
||||
BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K, 0x02e, 3, 0), // sendmsg
|
||||
BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K, 0x02c, 2, 0), // sendto
|
||||
BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K, 0x031, 1, 0), // bind
|
||||
BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K, 0x02a, 0, 1), // connect
|
||||
BPF_STMT(BPF_RET + BPF_K, SECCOMP_RET_TRACE),
|
||||
// default course of action
|
||||
BPF_STMT(BPF_RET + BPF_K, SECCOMP_RET_ALLOW),
|
||||
};
|
||||
|
||||
static int PeekData(int pid, long addr, void *buf, size_t size) {
|
||||
long i, j, w;
|
||||
for (i = 0; i < size; i += sizeof(long)) {
|
||||
if (sys_ptrace(PTRACE_PEEKTEXT, pid, addr + i, &w) != -1) {
|
||||
for (j = 0; i + j < size && j < sizeof(long); ++j) {
|
||||
((char *)buf)[i + j] = w;
|
||||
w >>= 8;
|
||||
}
|
||||
} else {
|
||||
return -1;
|
||||
}
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
static void LogProcessEvent(int main, int pid, int ws) {
|
||||
DEBUG("trace: %s%06d%s 0x%06x", //
|
||||
pid == main ? "\e[31;1m" : "", //
|
||||
pid, //
|
||||
pid == main ? "\e[0m" : "", //
|
||||
ws);
|
||||
if (WIFEXITED(ws)) {
|
||||
DEBUG(" exit %d", WEXITSTATUS(ws));
|
||||
}
|
||||
if (WIFSIGNALED(ws)) {
|
||||
DEBUG(" sig %d", WTERMSIG(ws));
|
||||
}
|
||||
if (WIFSTOPPED(ws)) {
|
||||
DEBUG(" stop %s %s", strsignal(WSTOPSIG(ws)),
|
||||
DescribePtraceEvent((ws & 0xff0000) >> 16));
|
||||
}
|
||||
if (WIFCONTINUED(ws)) {
|
||||
DEBUG(" cont");
|
||||
}
|
||||
if (WCOREDUMP(ws)) {
|
||||
DEBUG(" core");
|
||||
}
|
||||
DEBUG("\n");
|
||||
}
|
||||
|
||||
static int Raise(int sig) {
|
||||
sigset_t mask;
|
||||
sigaction(sig, &(struct sigaction){0}, 0);
|
||||
sigfillset(&mask);
|
||||
sigprocmask(SIG_SETMASK, &mask, 0);
|
||||
kill(getpid(), sig);
|
||||
sigdelset(&mask, sig);
|
||||
sigprocmask(SIG_SETMASK, &mask, 0);
|
||||
_Exit(128 + sig);
|
||||
}
|
||||
|
||||
static bool IsSockaddrAllowed(struct sockaddr_storage *addr) {
|
||||
uint32_t ip;
|
||||
if (addr->ss_family == AF_UNIX) {
|
||||
return true;
|
||||
}
|
||||
if (addr->ss_family == AF_INET) {
|
||||
ip = ntohl(((struct sockaddr_in *)addr)->sin_addr.s_addr);
|
||||
if (IsPrivateIp(ip) || IsLoopbackIp(ip)) {
|
||||
return true;
|
||||
} else {
|
||||
kprintf("warning: attempted to communicate with public ip "
|
||||
"%hhd.%hhd.%hhd.%hhd\n",
|
||||
ip >> 24, ip >> 16, ip >> 8, ip);
|
||||
return false;
|
||||
}
|
||||
}
|
||||
DEBUG("bad family %d\n", addr->ss_family);
|
||||
return false;
|
||||
}
|
||||
|
||||
static void OnSockaddrSyscall(int pid, int r1, int r2) {
|
||||
long si, dx;
|
||||
uint32_t addrlen;
|
||||
struct sockaddr_storage addr = {0};
|
||||
ORDIE(sys_ptrace(PTRACE_PEEKUSER, pid, r1, &si));
|
||||
ORDIE(sys_ptrace(PTRACE_PEEKUSER, pid, r2, &dx));
|
||||
addrlen = dx;
|
||||
if (!si) {
|
||||
// if address isn't supplied, it's probably safe. for example,
|
||||
// send() is implemented in cosmo using sendto() with 0/0 addr
|
||||
return;
|
||||
}
|
||||
if (PeekData(pid, si, &addr, MIN(addrlen, sizeof(addr))) == -1) {
|
||||
DEBUG("failed to peek addr\n"); // probably an efault
|
||||
goto Deny;
|
||||
}
|
||||
if (IsSockaddrAllowed(&addr)) {
|
||||
return;
|
||||
} else {
|
||||
goto Deny;
|
||||
}
|
||||
Deny:
|
||||
ORDIE(sys_ptrace(PTRACE_POKEUSER, pid, ORIG_RAX, -1));
|
||||
}
|
||||
|
||||
static void OnSendmsg(int pid) {
|
||||
long si;
|
||||
struct msghdr msg = {0};
|
||||
struct sockaddr_storage addr = {0};
|
||||
ORDIE(sys_ptrace(PTRACE_PEEKUSER, pid, RSI, &si));
|
||||
if (PeekData(pid, si, &msg, sizeof(msg)) == -1) {
|
||||
DEBUG("failed to peek msg\n"); // probably an efault
|
||||
goto Deny;
|
||||
}
|
||||
if (!msg.msg_name) {
|
||||
// if address isn't supplied, it's probably fine.
|
||||
return;
|
||||
}
|
||||
if (PeekData(pid, (long)msg.msg_name, &addr,
|
||||
MIN(msg.msg_namelen, sizeof(addr))) == -1) {
|
||||
DEBUG("failed to peek msg name\n"); // probably an efault
|
||||
goto Deny;
|
||||
}
|
||||
if (IsSockaddrAllowed(&addr)) {
|
||||
return;
|
||||
} else {
|
||||
goto Deny;
|
||||
}
|
||||
Deny:
|
||||
ORDIE(sys_ptrace(PTRACE_POKEUSER, pid, ORIG_RAX, -1));
|
||||
}
|
||||
|
||||
static void HandleSeccompTrace(int pid) {
|
||||
long ax;
|
||||
ORDIE(sys_ptrace(PTRACE_PEEKUSER, pid, ORIG_RAX, &ax));
|
||||
switch (ax) {
|
||||
case 0x031: // bind
|
||||
case 0x02a: // connect
|
||||
OnSockaddrSyscall(pid, RSI, RDX);
|
||||
break;
|
||||
case 0x02c: // sendto
|
||||
OnSockaddrSyscall(pid, R8, R9);
|
||||
break;
|
||||
case 0x02e: // sendmsg
|
||||
OnSendmsg(pid);
|
||||
break;
|
||||
default:
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
static int WaitForTrace(int main) {
|
||||
int ws, pid;
|
||||
for (;;) {
|
||||
// waits for state change on any child process or thread
|
||||
// eintr isn't possible since we're blocking all signals
|
||||
ORDIE(pid = waitpid(-1, &ws, __WALL));
|
||||
LogProcessEvent(main, pid, ws);
|
||||
if (WIFEXITED(ws)) {
|
||||
if (pid == main) {
|
||||
_Exit(WEXITSTATUS(ws));
|
||||
}
|
||||
} else if (WIFSIGNALED(ws)) {
|
||||
if (pid == main) {
|
||||
Raise(WTERMSIG(ws));
|
||||
}
|
||||
} else if (WIFSTOPPED(ws)) {
|
||||
if ((ws >> 8) == (SIGTRAP | (PTRACE_EVENT_SECCOMP << 8))) {
|
||||
return pid;
|
||||
} else if ((ws >> 8) == (SIGTRAP | (PTRACE_EVENT_EXEC << 8))) {
|
||||
ORDIE(ptrace(PTRACE_CONT, pid, 0, 0));
|
||||
} else if ((ws >> 8) == (SIGTRAP | (PTRACE_EVENT_FORK << 8)) ||
|
||||
(ws >> 8) == (SIGTRAP | (PTRACE_EVENT_VFORK << 8)) ||
|
||||
(ws >> 8) == (SIGTRAP | (PTRACE_EVENT_CLONE << 8))) {
|
||||
ORDIE(ptrace(PTRACE_CONT, pid, 0, 0));
|
||||
} else {
|
||||
ORDIE(ptrace(PTRACE_CONT, pid, 0, WSTOPSIG(ws)));
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Disables internet access.
|
||||
*
|
||||
* Warning: This function uses ptrace to react to seccomp filter events.
|
||||
* This approach is effective, but it's not bulletproof, since a highly
|
||||
* motivated attacker could theoretically use threads to modify sockaddr
|
||||
* in the short time between it being monitored and the actual syscall.
|
||||
*/
|
||||
int nointernet(void) {
|
||||
int ws, act, main;
|
||||
sigset_t set, old;
|
||||
struct sock_fprog prog = {.filter = kInetBpf, .len = ARRAYLEN(kInetBpf)};
|
||||
|
||||
// seccomp bpf and ptrace are pretty much just linux for now.
|
||||
if (!IsLinux() || !__is_linux_2_6_23()) {
|
||||
return enosys();
|
||||
}
|
||||
|
||||
// prevent crash handlers from intercepting sigsegv
|
||||
ORDIE(sigfillset(&set));
|
||||
ORDIE(sigprocmask(SIG_SETMASK, &set, &old));
|
||||
|
||||
// create traced child that'll replace this program
|
||||
if ((main = fork()) == -1) {
|
||||
ORDIE(sigprocmask(SIG_SETMASK, &old, 0));
|
||||
return -1;
|
||||
}
|
||||
if (!main) {
|
||||
if (sys_ptrace(PTRACE_TRACEME, 0, 0, 0) == -1) {
|
||||
// there can be only one
|
||||
// throw sigsegv on eperm
|
||||
// we're already being traced
|
||||
asm("hlt");
|
||||
}
|
||||
prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
|
||||
ORDIE(prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog));
|
||||
ORDIE(kill(getpid(), SIGSTOP));
|
||||
ORDIE(sigprocmask(SIG_SETMASK, &old, 0));
|
||||
// return to caller from child
|
||||
return 0;
|
||||
}
|
||||
|
||||
// wait for child to stop itself
|
||||
ORDIE(waitpid(main, &ws, 0));
|
||||
if (WIFSIGNALED(ws)) {
|
||||
// child couldn't enable ptrace or seccomp
|
||||
sigprocmask(SIG_SETMASK, &old, 0);
|
||||
return eperm();
|
||||
}
|
||||
npassert(WIFSTOPPED(ws));
|
||||
|
||||
// parent process becomes monitor of subprocess tree. all signals
|
||||
// continue to be blocked since we assume they'll also be sent to
|
||||
// children, which will die, and then the monitor dies afterwards
|
||||
ORDIE(sys_ptrace(PTRACE_SETOPTIONS, main, 0,
|
||||
PTRACE_O_TRACESECCOMP | PTRACE_O_TRACEFORK |
|
||||
PTRACE_O_TRACEVFORK | PTRACE_O_TRACECLONE |
|
||||
PTRACE_O_TRACEEXEC));
|
||||
for (act = main;;) {
|
||||
ORDIE(sys_ptrace(PTRACE_CONT, act, 0, 0));
|
||||
act = WaitForTrace(main);
|
||||
HandleSeccompTrace(act);
|
||||
}
|
||||
}
|
||||
|
||||
#endif /* __x86_64__ */
|
||||
|
|
@ -40,19 +40,25 @@
|
|||
#include "libc/sock/sendfile.internal.h"
|
||||
#include "libc/str/str.h"
|
||||
#include "libc/sysv/errfuns.h"
|
||||
#include "libc/thread/posixthread.internal.h"
|
||||
|
||||
// sendfile() isn't specified as raising eintr
|
||||
static textwindows int SendfileBlock(int64_t handle,
|
||||
struct NtOverlapped *overlapped) {
|
||||
struct PosixThread *pt;
|
||||
uint32_t i, got, flags = 0;
|
||||
if (WSAGetLastError() != kNtErrorIoPending &&
|
||||
WSAGetLastError() != WSAEINPROGRESS) {
|
||||
NTTRACE("TransmitFile failed %lm");
|
||||
return __winsockerr();
|
||||
}
|
||||
pt = _pthread_self();
|
||||
pt->abort_errno = 0;
|
||||
pt->ioverlap = overlapped;
|
||||
pt->iohandle = handle;
|
||||
for (;;) {
|
||||
i = WSAWaitForMultipleEvents(1, &overlapped->hEvent, true,
|
||||
__SIG_POLLING_INTERVAL_MS, true);
|
||||
__SIG_IO_INTERVAL_MS, true);
|
||||
if (i == kNtWaitFailed) {
|
||||
NTTRACE("WSAWaitForMultipleEvents failed %lm");
|
||||
return __winsockerr();
|
||||
|
|
@ -65,9 +71,14 @@ static textwindows int SendfileBlock(int64_t handle,
|
|||
break;
|
||||
}
|
||||
}
|
||||
pt->ioverlap = 0;
|
||||
pt->iohandle = 0;
|
||||
if (WSAGetOverlappedResult(handle, overlapped, &got, false, &flags)) {
|
||||
return got;
|
||||
} else {
|
||||
if (WSAGetLastError() == kNtErrorOperationAborted) {
|
||||
errno = pt->abort_errno;
|
||||
}
|
||||
return -1;
|
||||
}
|
||||
}
|
||||
|
|
@ -99,7 +110,7 @@ static dontinline textwindows ssize_t sys_sendfile_nt(
|
|||
return ebadf();
|
||||
}
|
||||
struct NtOverlapped ov = {
|
||||
.Pointer = (void *)(intptr_t)offset,
|
||||
.Pointer = offset,
|
||||
.hEvent = WSACreateEvent(),
|
||||
};
|
||||
if (TransmitFile(oh, ih, uptobytes, 0, &ov, 0, 0)) {
|
||||
|
|
|
|||
|
|
@ -2,9 +2,6 @@
|
|||
#define COSMOPOLITAN_LIBC_SOCK_SOCK_H_
|
||||
#if !(__ASSEMBLER__ + __LINKER__ + 0)
|
||||
COSMOPOLITAN_C_START_
|
||||
/*───────────────────────────────────────────────────────────────────────────│─╗
|
||||
│ cosmopolitan § system api » berkeley sockets ─╬─│┼
|
||||
╚────────────────────────────────────────────────────────────────────────────│*/
|
||||
|
||||
#define INET_ADDRSTRLEN 22
|
||||
#define IFHWADDRLEN 6
|
||||
|
|
@ -27,7 +24,6 @@ uint32_t inet_addr(const char *);
|
|||
int parseport(const char *);
|
||||
uint32_t *GetHostIps(void);
|
||||
|
||||
int nointernet(void);
|
||||
int socket(int, int, int);
|
||||
int listen(int, int);
|
||||
int shutdown(int, int);
|
||||
|
|
|
|||
|
|
@ -20,6 +20,7 @@
|
|||
#include "libc/calls/bo.internal.h"
|
||||
#include "libc/calls/internal.h"
|
||||
#include "libc/calls/sig.internal.h"
|
||||
#include "libc/calls/struct/timespec.h"
|
||||
#include "libc/errno.h"
|
||||
#include "libc/nt/enum/wait.h"
|
||||
#include "libc/nt/enum/wsa.h"
|
||||
|
|
@ -35,66 +36,91 @@
|
|||
#include "libc/str/str.h"
|
||||
#include "libc/sysv/consts/o.h"
|
||||
#include "libc/sysv/errfuns.h"
|
||||
#include "libc/thread/posixthread.internal.h"
|
||||
#include "libc/thread/tls.h"
|
||||
|
||||
static textwindows void __wsablock_abort(int64_t handle,
|
||||
struct NtOverlapped *overlapped) {
|
||||
unassert(CancelIoEx(handle, overlapped) ||
|
||||
GetLastError() == kNtErrorNotFound);
|
||||
}
|
||||
|
||||
textwindows int __wsablock(struct Fd *fd, struct NtOverlapped *overlapped,
|
||||
textwindows int __wsablock(struct Fd *f, struct NtOverlapped *overlapped,
|
||||
uint32_t *flags, int sigops, uint32_t timeout) {
|
||||
bool nonblock;
|
||||
int e, rc, err;
|
||||
uint32_t i, got;
|
||||
int rc, abort_errno;
|
||||
uint32_t waitfor;
|
||||
struct PosixThread *pt;
|
||||
struct timespec now, remain, interval, deadline;
|
||||
|
||||
if (WSAGetLastError() != kNtErrorIoPending) {
|
||||
// our i/o operation never happened because it failed
|
||||
return __winsockerr();
|
||||
}
|
||||
BEGIN_BLOCKING_OPERATION;
|
||||
|
||||
// our i/o operation is in flight and it needs to block
|
||||
abort_errno = EAGAIN;
|
||||
if (fd->flags & O_NONBLOCK) {
|
||||
__wsablock_abort(fd->handle, overlapped);
|
||||
nonblock = !!(f->flags & O_NONBLOCK);
|
||||
pt = _pthread_self();
|
||||
pt->abort_errno = EAGAIN;
|
||||
interval = timespec_frommillis(__SIG_IO_INTERVAL_MS);
|
||||
deadline = timeout
|
||||
? timespec_add(timespec_real(), timespec_frommillis(timeout))
|
||||
: timespec_max;
|
||||
e = errno;
|
||||
BlockingOperation:
|
||||
if (!nonblock) {
|
||||
pt->ioverlap = overlapped;
|
||||
pt->iohandle = f->handle;
|
||||
}
|
||||
if (nonblock) {
|
||||
CancelIoEx(f->handle, overlapped);
|
||||
} else if (_check_interrupts(sigops)) {
|
||||
Interrupted:
|
||||
abort_errno = errno; // EINTR or ECANCELED
|
||||
__wsablock_abort(fd->handle, overlapped);
|
||||
pt->abort_errno = errno; // EINTR or ECANCELED
|
||||
CancelIoEx(f->handle, overlapped);
|
||||
} else {
|
||||
for (;;) {
|
||||
i = WSAWaitForMultipleEvents(1, &overlapped->hEvent, true,
|
||||
__SIG_POLLING_INTERVAL_MS, true);
|
||||
if (i == kNtWaitFailed || i == kNtWaitTimeout) {
|
||||
now = timespec_real();
|
||||
if (timespec_cmp(now, deadline) >= 0) {
|
||||
CancelIoEx(f->handle, overlapped);
|
||||
nonblock = true;
|
||||
break;
|
||||
}
|
||||
remain = timespec_sub(deadline, now);
|
||||
if (timespec_cmp(remain, interval) >= 0) {
|
||||
waitfor = __SIG_IO_INTERVAL_MS;
|
||||
} else {
|
||||
waitfor = timespec_tomillis(remain);
|
||||
}
|
||||
i = WSAWaitForMultipleEvents(1, &overlapped->hEvent, true, waitfor, true);
|
||||
if (i == kNtWaitFailed) {
|
||||
// Failure should be an impossible condition, but MSDN lists
|
||||
// WSAENETDOWN and WSA_NOT_ENOUGH_MEMORY as possible errors.
|
||||
pt->abort_errno = WSAGetLastError();
|
||||
CancelIoEx(f->handle, overlapped);
|
||||
nonblock = true;
|
||||
break;
|
||||
} else if (i == kNtWaitTimeout) {
|
||||
if (_check_interrupts(sigops)) {
|
||||
goto Interrupted;
|
||||
}
|
||||
if (i == kNtWaitFailed) {
|
||||
// Failure should be an impossible condition, but MSDN lists
|
||||
// WSAENETDOWN and WSA_NOT_ENOUGH_MEMORY as possible errors,
|
||||
// which we're going to hope are ephemeral.
|
||||
SleepEx(__SIG_POLLING_INTERVAL_MS, false);
|
||||
}
|
||||
if (timeout) {
|
||||
if (timeout <= __SIG_POLLING_INTERVAL_MS) {
|
||||
__wsablock_abort(fd->handle, overlapped);
|
||||
break;
|
||||
}
|
||||
timeout -= __SIG_POLLING_INTERVAL_MS;
|
||||
}
|
||||
continue;
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
pt->ioverlap = 0;
|
||||
pt->iohandle = 0;
|
||||
|
||||
// overlapped is allocated on stack by caller, so it's important that
|
||||
// we wait for win32 to acknowledge that it's done using that memory.
|
||||
if (WSAGetOverlappedResult(fd->handle, overlapped, &got, true, flags)) {
|
||||
if (WSAGetOverlappedResult(f->handle, overlapped, &got, nonblock, flags)) {
|
||||
rc = got;
|
||||
} else {
|
||||
rc = -1;
|
||||
if (WSAGetLastError() == kNtErrorOperationAborted) {
|
||||
errno = abort_errno;
|
||||
err = WSAGetLastError();
|
||||
if (err == kNtErrorOperationAborted) {
|
||||
errno = pt->abort_errno;
|
||||
} else if (err == kNtErrorIoIncomplete) {
|
||||
errno = e;
|
||||
goto BlockingOperation;
|
||||
}
|
||||
}
|
||||
END_BLOCKING_OPERATION;
|
||||
return rc;
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue