Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit6051857

Browse files
committed
On Windows, close the client socket explicitly during backend shutdown.
It turns out that this is necessary to keep Winsock from dropping anynot-yet-sent data, such as an error message explaining the reason forprocess termination. It's pretty weird that the implicit close doneby the kernel acts differently from an explicit close, but it's hardto argue with experimental results.Independently submitted by Alexander Lakhin and Lars Kanis (commentsby me, though). Back-patch to all supported branches.Discussion:https://postgr.es/m/90b34057-4176-7bb0-0dbb-9822a5f6425b@greiz-reinsdorf.deDiscussion:https://postgr.es/m/16678-253e48d34dc0c376@postgresql.org
1 parenta7da419 commit6051857

File tree

1 file changed

+20
-7
lines changed

1 file changed

+20
-7
lines changed

‎src/backend/libpq/pqcomm.c

Lines changed: 20 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -277,15 +277,28 @@ socket_close(int code, Datum arg)
277277
secure_close(MyProcPort);
278278

279279
/*
280-
* Formerly we did an explicit close() here, but it seems better to
281-
* leave the socket open until the process dies. This allows clients
282-
* to perform a "synchronous close" if they care --- wait till the
283-
* transport layer reports connection closure, and you can be sure the
284-
* backend has exited.
280+
* On most platforms, we leave the socket open until the process dies.
281+
* This allows clients to perform a "synchronous close" if they care
282+
* --- wait till the transport layer reports connection closure, and
283+
* you can be sure the backend has exited. Saves a kernel call, too.
285284
*
286-
* We do set sock to PGINVALID_SOCKET to prevent any further I/O,
287-
* though.
285+
* However, that does not work on Windows: if the kernel closes the
286+
* socket it will invoke an "abortive shutdown" that discards any data
287+
* not yet sent to the client. (This is a flat-out violation of the
288+
* TCP RFCs, but count on Microsoft not to care about that.) To get
289+
* the spec-compliant "graceful shutdown" behavior, we must invoke
290+
* closesocket() explicitly.
291+
*
292+
* This code runs late enough during process shutdown that we should
293+
* have finished all externally-visible shutdown activities, so that
294+
* in principle it's good enough to act as a synchronous close on
295+
* Windows too. But it's a lot more fragile than the other way.
288296
*/
297+
#ifdefWIN32
298+
closesocket(MyProcPort->sock);
299+
#endif
300+
301+
/* In any case, set sock to PGINVALID_SOCKET to prevent further I/O */
289302
MyProcPort->sock=PGINVALID_SOCKET;
290303
}
291304
}

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp