mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-02-07 23:13:34 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1
cosmopolitan/libc/sock/accept-nt.c

154 lines
6 KiB
C
Raw Normal View History

Initial import
2020-06-15 14:18:57 +00:00
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
flip et / noet in modelines `et` means `expandtab`. ```sh rg 'vi: .* :vi' -l -0 | \ xargs -0 sed -i '' 's/vi: \(.*\) et\(.*\) :vi/vi: \1 xoet\2:vi/' rg 'vi: .* :vi' -l -0 | \ xargs -0 sed -i '' 's/vi: \(.*\)noet\(.*\):vi/vi: \1et\2 :vi/' rg 'vi: .* :vi' -l -0 | \ xargs -0 sed -i '' 's/vi: \(.*\)xoet\(.*\):vi/vi: \1noet\2:vi/' ```
2023-12-08 03:11:56 +00:00
│ vi: set et ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi │
Initial import
2020-06-15 14:18:57 +00:00
╞══════════════════════════════════════════════════════════════════════════════╡
│ Copyright 2020 Justine Alexandra Roberts Tunney │
│ │
Change license
2020-12-28 01:18:44 +00:00
│ 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. │
Initial import
2020-06-15 14:18:57 +00:00
│ │
Change license
2020-12-28 01:18:44 +00:00
│ 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. │
Initial import
2020-06-15 14:18:57 +00:00
╚─────────────────────────────────────────────────────────────────────────────*/
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
#include "libc/assert.h"
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
#include "libc/atomic.h"
Initial import
2020-06-15 14:18:57 +00:00
#include "libc/calls/internal.h"
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
#include "libc/calls/struct/fd.internal.h"
#include "libc/calls/struct/sigset.internal.h"
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
#include "libc/cosmo.h"
Fix ctrl-c in redbean on Windows
2023-10-13 14:43:47 +00:00
#include "libc/errno.h"
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
#include "libc/nt/enum/wsaid.h"
Clean up more WIN32 APIs
2023-07-30 23:00:58 +00:00
#include "libc/nt/thunk/msabi.h"
Initial import
2020-06-15 14:18:57 +00:00
#include "libc/nt/winsock.h"
#include "libc/sock/internal.h"
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
#include "libc/sock/struct/sockaddr.h"
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
#include "libc/sock/wsaid.internal.h"
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
#include "libc/str/str.h"
Cleanup socket code
2021-02-06 07:45:34 +00:00
#include "libc/sysv/consts/o.h"
Initial import
2020-06-15 14:18:57 +00:00
#include "libc/sysv/consts/sock.h"
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
#include "libc/sysv/consts/sol.h"
#include "libc/thread/thread.h"
#ifdef __x86_64__
Initial import
2020-06-15 14:18:57 +00:00
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
__msabi extern typeof(__sys_setsockopt_nt) *const __imp_setsockopt;
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
__msabi extern typeof(__sys_closesocket_nt) *const __imp_closesocket;
Clean up more WIN32 APIs
2023-07-30 23:00:58 +00:00
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
union AcceptExAddr {
Fix some more socket bugs - The functions that return a sockaddr now do so the same way the Linux Kernel does across platforms, e.g. getpeername(), accept4() - Socket system calls on Windows will now only check for interrupts when a blocking operation needs to be performed. - Write tests for recvfrom() system call
2023-07-23 23:31:10 +00:00
struct sockaddr_storage addr;
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
char buf[sizeof(struct sockaddr_storage) + 16];
};
struct AcceptExBuffer {
union AcceptExAddr local;
union AcceptExAddr remote;
};
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
struct AcceptResources {
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
int64_t handle;
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
};
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
struct AcceptArgs {
int64_t listensock;
struct AcceptExBuffer *buffer;
};
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
static struct {
atomic_uint once;
bool32 (*__msabi lpAcceptEx)(
int64_t sListenSocket, int64_t sAcceptSocket,
void *out_lpOutputBuffer /*[recvlen+local+remoteaddrlen]*/,
uint32_t dwReceiveDataLength, uint32_t dwLocalAddressLength,
uint32_t dwRemoteAddressLength, uint32_t *out_lpdwBytesReceived,
struct NtOverlapped *inout_lpOverlapped);
} g_acceptex;
static void acceptex_init(void) {
static struct NtGuid AcceptExGuid = WSAID_ACCEPTEX;
g_acceptex.lpAcceptEx = __get_wsaid(&AcceptExGuid);
}
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
static void sys_accept_nt_unwind(void *arg) {
struct AcceptResources *resources = arg;
if (resources->handle != -1) {
__imp_closesocket(resources->handle);
}
}
static int sys_accept_nt_start(int64_t handle, struct NtOverlapped *overlap,
uint32_t *flags, void *arg) {
struct AcceptArgs *args = arg;
Improve threading and i/o routines - On Windows connect() can now be interrupted by a signal; connect() w/ O_NONBLOCK will now raise EINPROGRESS; and connect() with SO_SNDTIMEO will raise ETIMEDOUT after the interval has elapsed. - We now get the AcceptEx(), ConnectEx(), and TransmitFile() functions from the WIN32 API the officially blessed way, using WSAIoctl(). - Do nothing on Windows when fsync() is called on a directory handle. This was raising EACCES earlier becaues GENERIC_WRITE is required on the handle. It's possible to FlushFileBuffers() a directory handle if it's opened with write access but MSDN doesn't document what it does. If you have any idea, please let us know! - Prefer manual reset event objects for read() and write() on Windows. - Do some code cleanup on our dlmalloc customizations. - Fix errno type error in Windows blocking routines. - Make the futex polyfill simpler and faster.
2023-10-13 01:53:17 +00:00
cosmo_once(&g_acceptex.once, acceptex_init);
if (g_acceptex.lpAcceptEx(args->listensock, handle, args->buffer, 0,
sizeof(args->buffer->local),
sizeof(args->buffer->remote), 0, overlap)) {
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
// inherit properties of listening socket
unassert(!__imp_setsockopt(args->listensock, SOL_SOCKET,
kNtSoUpdateAcceptContext, &handle,
sizeof(handle)));
return 0;
} else {
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
return -1;
}
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
}
textwindows int sys_accept_nt(struct Fd *f, struct sockaddr_storage *addr,
int accept4_flags) {
int client = -1;
sigset_t m = __sig_block();
struct AcceptResources resources = {-1};
pthread_cleanup_push(sys_accept_nt_unwind, &resources);
// creates resources for child socket
// inherit the listener configuration
if ((resources.handle = WSASocket(f->family, f->type, f->protocol, 0, 0,
kNtWsaFlagOverlapped)) == -1) {
client = __winsockerr();
Put more thought into i/o polyfills wait4() is now solid enough to run `make -j100` on Windows. You can now use MSG_DONTWAIT on Windows. There was a handle leak in accept() that's been fixed. Our WIN32 overlapped i/o code has been simplified. Priority class now inherits into subprocesses, so the verynice command will work and the signal mask will now be inherited by execve() and posix_spawn()
2023-11-07 00:38:44 +00:00
goto Finish;
Initial import
2020-06-15 14:18:57 +00:00
}
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
// accept network connection
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
// this operation can re-enter, interrupt, cancel, block, timeout, etc.
struct AcceptExBuffer buffer;
ssize_t bytes_received = __winsock_block(
resources.handle, 0, !!(f->flags & O_NONBLOCK), f->rcvtimeo, m,
sys_accept_nt_start, &(struct AcceptArgs){f->handle, &buffer});
Put more thought into i/o polyfills wait4() is now solid enough to run `make -j100` on Windows. You can now use MSG_DONTWAIT on Windows. There was a handle leak in accept() that's been fixed. Our WIN32 overlapped i/o code has been simplified. Priority class now inherits into subprocesses, so the verynice command will work and the signal mask will now be inherited by execve() and posix_spawn()
2023-11-07 00:38:44 +00:00
if (bytes_received == -1) {
__imp_closesocket(resources.handle);
goto Finish;
}
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
// create file descriptor for new socket
// don't inherit the file open mode bits
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
int oflags = 0;
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
if (accept4_flags & SOCK_CLOEXEC) oflags |= O_CLOEXEC;
if (accept4_flags & SOCK_NONBLOCK) oflags |= O_NONBLOCK;
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
client = __reservefd(-1);
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
g_fds.p[client].flags = oflags;
g_fds.p[client].mode = 0140666;
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
g_fds.p[client].family = f->family;
g_fds.p[client].type = f->type;
g_fds.p[client].protocol = f->protocol;
g_fds.p[client].sndtimeo = f->sndtimeo;
g_fds.p[client].rcvtimeo = f->rcvtimeo;
g_fds.p[client].handle = resources.handle;
resources.handle = -1;
Fix some more socket bugs - The functions that return a sockaddr now do so the same way the Linux Kernel does across platforms, e.g. getpeername(), accept4() - Socket system calls on Windows will now only check for interrupts when a blocking operation needs to be performed. - Write tests for recvfrom() system call
2023-07-23 23:31:10 +00:00
memcpy(addr, &buffer.remote.addr, sizeof(*addr));
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
g_fds.p[client].kind = kFdSocket;
Put more thought into i/o polyfills wait4() is now solid enough to run `make -j100` on Windows. You can now use MSG_DONTWAIT on Windows. There was a handle leak in accept() that's been fixed. Our WIN32 overlapped i/o code has been simplified. Priority class now inherits into subprocesses, so the verynice command will work and the signal mask will now be inherited by execve() and posix_spawn()
2023-11-07 00:38:44 +00:00
Finish:
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
pthread_cleanup_pop(false);
__sig_unblock(m);
Fix ctrl-c in redbean on Windows
2023-10-13 14:43:47 +00:00
if (client == -1 && errno == ECONNRESET) {
errno = ECONNABORTED;
}
Support non-blocking i/o across platforms This change introduces new tests for `O_NONBLOCK` and `SOCK_NONBLOCK` to confirm that non-blocking i/o is now working on all supported platforms, including Windows. For example, you can now say on Windows, MacOS, etc.: socket(AF_INET, SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP); To create a non-blocking IPv4 TCP socket. Or you can enable non-blocking i/o on an existing socket / pipe / etc. file descriptor by calling fcntl fcntl(fd, F_SETFL, fcntl(fd, F_GETFL) | O_NONBLOCK); This functionality is polyfilled on older Linux kernels too, e.g. RHEL5. Now that fcntl() support is much better the FIOCLEX / FIONCLEX polyfills for ioctl() have been removed since they're ugly non-POSIX diameond APIs This change fixes a weakness in kprintf() that was causing Windows trace tools to frequently crash.
2023-07-23 09:56:47 +00:00
return client;
Initial import
2020-06-15 14:18:57 +00:00
}
Make improvements - We now serialize the file descriptor table when spawning / executing processes on Windows. This means you can now inherit more stuff than just standard i/o. It's needed by bash, which duplicates the console to file descriptor #255. We also now do a better job serializing the environment variables, so you're less likely to encounter E2BIG when using your bash shell. We also no longer coerce environ to uppercase - execve() on Windows now remotely controls its parent process to make them spawn a replacement for itself. Then it'll be able to terminate immediately once the spawn succeeds, without having to linger around for the lifetime as a shell process for proxying the exit code. When process worker thread running in the parent sees the child die, it's given a handle to the new child, to replace it in the process table. - execve() and posix_spawn() on Windows will now provide CreateProcess an explicit handle list. This allows us to remove handle locks which enables better fork/spawn concurrency, with seriously correct thread safety. Other codebases like Go use the same technique. On the other hand fork() still favors the conventional WIN32 inheritence approach which can be a little bit messy, but is *controlled* by guaranteeing perfectly clean slates at both the spawning and execution boundaries - sigset_t is now 64 bits. Having it be 128 bits was a mistake because there's no reason to use that and it's only supported by FreeBSD. By using the system word size, signal mask manipulation on Windows goes very fast. Furthermore @asyncsignalsafe funcs have been rewritten on Windows to take advantage of signal masking, now that it's much more pleasant to use. - All the overlapped i/o code on Windows has been rewritten for pretty good signal and cancelation safety. We're now able to ensure overlap data structures are cleaned up so long as you don't longjmp() out of out of a signal handler that interrupted an i/o operation. Latencies are also improved thanks to the removal of lots of "busy wait" code. Waits should be optimal for everything except poll(), which shall be the last and final demon we slay in the win32 i/o horror show. - getrusage() on Windows is now able to report RUSAGE_CHILDREN as well as RUSAGE_SELF, thanks to aggregation in the process manager thread.
2023-10-08 12:36:18 +00:00
#endif /* __x86_64__ */
Reference in a new issue Copy permalink