MinGW: port POSIX socket compatibility layer #2046

kinkie · 2025-03-29T12:35:06Z

winsock's socket-related functions are very similar
but not identical to the POSIX standard.
Make the mswindows compatibility layer
available to MinGW, and modernize it.

Implement portable wrappers around
socket-related calls, named x, and
use them in all callsites.

Error highlighting the issue:

TcpAcceptor.cc: In member function
  'bool Comm::TcpAcceptor::acceptInto(Comm::ConnectionPointer&)':
TcpAcceptor.cc:352:17: error:
  comparison of unsigned expression in '< 0'
  is always false [-Werror=type-limits]
  352 |     if (rawSock < 0) {

kinkie · 2025-03-29T12:49:38Z

note: more socket-related functions will need to have this treatment. I wonder whether it's best
to have each have their own header and implementation or to group them all under a single xsocket.* file

yadij

I agree we should probably be planning for libcompat to present the socket.h API as compat/socket.h instead of a separate include file for each syscall.

compat/xaccept.cc

rousskov · 2025-03-31T20:47:40Z

more socket-related functions will need to have this treatment. I wonder whether it's best to have each have their own header and implementation or to group them all under a single xsocket.* file

We should group sys/socket.h wrappers (i.e. listen(2), accept(2), connect(2), shutdown(2), getsockopt(2), etc.), especially if they have similar implementations.

A more difficult question is whether introduction of xfoo() warrants conversion from legacy POSIX foo() to modern C++ API: If every foo() caller is a C++ caller and that caller has to be updated anyway, then it could be the right time to start doing proper memory management and error handling (instead of writing or preserving awkward and duplicated code). That requires a lot more/serious work though. It may also require to get rid of the remaining legacy C callers.

compat/xaccept.cc

@rousskov

okay, pending @rousskov review issues and final call on whether to use compat/xsocket.* filenames.

Windows' accept(2) has the same signature as POSIX', but returns a SOCKET (under the hood, a int *). This confuses error handling code in squid, which uses POSIX return values to detect error values. Implement xaccept() to wrap around it, taking the implementation from mswindows.h and using it on MinGW as well. Error highlighting the issue: ``` TcpAcceptor.cc: In member function 'bool Comm::TcpAcceptor::acceptInto(Comm::ConnectionPointer&)': TcpAcceptor.cc:352:17: error: comparison of unsigned expression in '< 0' is always false [-Werror=type-limits] 352 | if (rawSock < 0) { ```

Co-authored-by: Amos Jeffries <[email protected]>

compat/socket.cc

kinkie · 2025-04-30T11:47:03Z

We are almost done here. There are more methods (WSADuplicateSocket for instance) but I'd look into them separately.

The only point of questio for me is ioctl/ioctlsocket . Windows has separate versions for files and sockets, and doesn't support fcntl. But then, in practice, we only use ioctl for sockets. My plan here would be to name the x-function xioctlsocket, to highlight that it's not meant to be used on filedescriptors, and not to have an xioctl at all.

@yadij , @rousskov , I'd like advice before committing to a path and then having to change it

rousskov

What a fundamental effort!

Francesco: The only point of questio for me is ioctl/ioctlsocket . Windows has separate versions for files and sockets, and doesn't support fcntl. But then, in practice, we only use ioctl for sockets. My plan here would be to name the x-function xioctlsocket, to highlight that it's not meant to be used on filedescriptors, and not to have an xioctl at all.

watch_child() may call ioctl() for an open("/dev/tty") descriptor. Does that contradict the above "we only use ioctl for sockets" assertion? Similar concerns may apply to ModDevPoll.cc that uses "/dev/poll".

Win32 API: This ioctlsocket function performs only a subset of functions on a socket when compared to the ioctl function found in Berkeley sockets. The ioctlsocket function has no command parameter equivalent to the FIOASYNC of ioctl, and SIOCATMARK is the only socket-level command that is supported by ioctlsocket.

Can we use xioctl() (instead of xioctlsocket()) everywhere in high-level code? If some callers/use cases can only be supported on Windows via something other than ioctlsocket(), then perhaps our Windows-specific xioctl() implementation can detect those use cases by testing xioctl() parameters and activating special non-ioctlsocket() handling code as needed?

.github/workflows/mingw.yaml

rousskov · 2025-04-30T13:37:18Z

compat/netdb.cc

+
+#include "squid.h"
+
+#if (_SQUID_WINDOWS_ || _SQUID_MINGW_)


Please address the usual condition consistency and "empty line at the end of the file" concerns.

Done in c21a151

rousskov · 2025-04-30T13:38:07Z

compat/netdb.cc

+struct hostent *
+xgethostbyname(const char *name)
+{
+    auto result = ::gethostbyname(name);


Please add const to new variable declarations whenever possible.

done in c21a151

rousskov · 2025-04-30T13:45:09Z

compat/netdb.h

+#include <netdb.h>
+#endif
+
+/// Provide POSIX gethostbyname(2) API on MinGW and Visual Studio build environments


Please focus on describing the function (in all environments it is present). The reasons this function was added will be clear from the commit message. For example:

Suggested change

/// Provide POSIX gethostbyname(2) API on MinGW and Visual Studio build environments

/// POSIX gethostbyname(3) equivalent

or simply

Suggested change

/// Provide POSIX gethostbyname(2) API on MinGW and Visual Studio build environments

/// gethostbyname(3) equivalent

Please add or fix other descriptions in a similar way, paying attention to the correct man page section.

rousskov · 2025-04-30T13:49:28Z

compat/netdb.h

+#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)
+// Windows and MinGW implementations are in compat/netdb.cc


Please do not place a comment inside the if statement that is false for that comment context:

Suggested change

#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)

// Windows and MinGW implementations are in compat/netdb.cc

// Windows and MinGW implementations are in compat/netdb.cc

#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)

or

Suggested change

#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)

// Windows and MinGW implementations are in compat/netdb.cc

#if _SQUID_WINDOWS_ || _SQUID_MINGW_

// Windows and MinGW implementations are in compat/netdb.cc

#else

or simply remove this comment -- it is trivial to find those implementations, and we do not have to document the default "implementations are in foo.cc" case anyway!

Suggested change

#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)

// Windows and MinGW implementations are in compat/netdb.cc

#if !(_SQUID_WINDOWS_ || _SQUID_MINGW_)

removing the comment

rousskov · 2025-04-30T13:54:26Z

compat/netdb.h

+    return gethostbyname(name);
+}
+
+inline struct servent *
+xgetservbyname(const char *name, const char *proto)
+{
+    return ::getservbyname(name, proto);


Please keep these trivial/forwarding implementations consistent. For example, either all of them should use :: prefix when calling the primary function or none (by default). The "none" option feels better to me because that is how most similar functions are normally called in Squid code and because these are C (rather than C++) functions, but I may be missing some good reasons to use the global namespace prefix.

rousskov self-requested a review March 29, 2025 14:24

yadij previously requested changes Mar 30, 2025

View reviewed changes

compat/xaccept.cc Outdated Show resolved Hide resolved

compat/xaccept.cc Outdated Show resolved Hide resolved

yadij added the S-waiting-for-author author action is expected (and usually required) label Mar 30, 2025

rousskov requested changes Mar 31, 2025

View reviewed changes

compat/xaccept.cc Outdated Show resolved Hide resolved

compat/xaccept.cc Outdated Show resolved Hide resolved

compat/xaccept.cc Outdated Show resolved Hide resolved

kinkie added 7 commits April 17, 2025 08:22

fix return type on POSIX

8aec939

Source format

4c6995f

More source formatting

c85fe33

implement xbind and xaccept

009650d

xconnect() and xgethostbyname()

01efb92

port setsockopt

ed2c975

kinkie force-pushed the update-mingw-16 branch from 084949b to ed2c975 Compare April 18, 2025 18:54

Source maintenance

7359d54

kinkie changed the title ~~MinGW: implement xaccept()~~ MinGW: port POSIX socket compatibility layer Apr 18, 2025

This comment was marked as resolved.

Sign in to view

kinkie force-pushed the update-mingw-16 branch from d96076c to 32ca454 Compare April 19, 2025 17:28

Fix files missing headers

a6dd98b

kinkie force-pushed the update-mingw-16 branch from 32ca454 to a6dd98b Compare April 19, 2025 17:48

kinkie and others added 8 commits April 19, 2025 17:55

do not include windows headers in socket.cc

b4b7bcb

Implement code review suggestion

21dd497

Accept proposal for SOCKET_ERROR handling

f78d5cb

Co-authored-by: Amos Jeffries <[email protected]>

make xclose more readable

5aec29e

increase xconnect readibility

ce61630

AAA for xgethostbyname

737edb3

improve xsetsockopt readability

60c88de

formatting

7942f2a

yadij requested changes Apr 20, 2025

View reviewed changes

kinkie added 19 commits April 28, 2025 22:51

Fix missing include in negotiate_kerberos_auth

1261eed

Add mingw github workflow

db0440a

add correct header

9413735

implement xrecvfrom

352d94c

fix typo

6835c94

Source maintenance

c898f1d

implement xsend

f1f5214

extract socket handle validation into heper function

fe1ae15

implement xsendto

ecd3532

minor formatting fix

ff9d81f

port xselect

d568548

remove some more wrappers

924bf6e

source formatting

4860967

more cleanups

4983a52

move xselect to compat/select.h

9663279

add missing windows guard

2c3cf6e

change parameter format to make linter happy

1eda212

try again to make linter happy

bbd7043

remove unused WSAAsyncSelect

49fec78

kinkie added S-waiting-for-reviewer ready for review: Set this when requesting a (re)review using GitHub PR Reviewers box and removed S-waiting-for-author author action is expected (and usually required) labels Apr 30, 2025

rousskov requested changes Apr 30, 2025

View reviewed changes

rousskov added the S-waiting-for-author author action is expected (and usually required) label Apr 30, 2025

kinkie added 6 commits April 30, 2025 18:42

adopt review feedback in compat/netdb.cc

c21a151

adopt review suggestions

0b455af

apply more review feedback

795ca82

Improve code consistency

9a2c7c6

soruce maintenance

b9892fa

remove ioctl.{cc,h}

8f61f72

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MinGW: port POSIX socket compatibility layer #2046

MinGW: port POSIX socket compatibility layer #2046

kinkie commented Mar 29, 2025 •

edited

Loading

kinkie commented Mar 29, 2025

yadij left a comment

rousskov commented Mar 31, 2025

This comment was marked as resolved.

kinkie commented Apr 30, 2025 •

edited

Loading

rousskov left a comment

rousskov Apr 30, 2025

kinkie Apr 30, 2025

rousskov Apr 30, 2025

kinkie Apr 30, 2025

rousskov Apr 30, 2025

rousskov Apr 30, 2025

kinkie Apr 30, 2025

rousskov Apr 30, 2025

	/// Provide POSIX gethostbyname(2) API on MinGW and Visual Studio build environments
	/// POSIX gethostbyname(3) equivalent

	/// Provide POSIX gethostbyname(2) API on MinGW and Visual Studio build environments
	/// gethostbyname(3) equivalent

		#if !(_SQUID_WINDOWS_ \|\| _SQUID_MINGW_)
		// Windows and MinGW implementations are in compat/netdb.cc

MinGW: port POSIX socket compatibility layer #2046

Are you sure you want to change the base?

MinGW: port POSIX socket compatibility layer #2046

Conversation

kinkie commented Mar 29, 2025 • edited Loading

kinkie commented Mar 29, 2025

yadij left a comment

Choose a reason for hiding this comment

rousskov commented Mar 31, 2025

This comment was marked as resolved.

kinkie commented Apr 30, 2025 • edited Loading

rousskov left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kinkie commented Mar 29, 2025 •

edited

Loading

kinkie commented Apr 30, 2025 •

edited

Loading