Fix a possible socket leak at Windows (src/backend/port/win32/socket.c)
Hi.
While there are plans to remove the sockets functions (Windows) [1]Re: Windows sockets (select missing events?) </messages/by-id/CA+hUKGKSLgxFhSP8+dqQqHsuZvrRRU3wZ6ytLOcno-mdGvckHg@mail.gmail.com>, I
believe it is worth fixing possible current bugs.
In the pgwin32_socket function (src/backend/port/win32/socket.c), there is
a possible socket leak if the socket cannot be made non-blocking.
Trivial patch attached.
Best regards,
Ranier Vilela
[1]: Re: Windows sockets (select missing events?) </messages/by-id/CA+hUKGKSLgxFhSP8+dqQqHsuZvrRRU3wZ6ytLOcno-mdGvckHg@mail.gmail.com>
</messages/by-id/CA+hUKGKSLgxFhSP8+dqQqHsuZvrRRU3wZ6ytLOcno-mdGvckHg@mail.gmail.com>
Attachments:
fix-socket-leak-windows.patchapplication/octet-stream; name=fix-socket-leak-windows.patchDownload
diff --git a/src/backend/port/win32/socket.c b/src/backend/port/win32/socket.c
index 6b60a9f220..0a819e07ae 100644
--- a/src/backend/port/win32/socket.c
+++ b/src/backend/port/win32/socket.c
@@ -303,6 +303,7 @@ pgwin32_socket(int af, int type, int protocol)
if (ioctlsocket(s, FIONBIO, &on))
{
TranslateSocketError();
+ closesocket(s);
return INVALID_SOCKET;
}
errno = 0;
On 13 Jan 2024, at 22:38, Ranier Vilela <ranier.vf@gmail.com> wrote:
In the pgwin32_socket function (src/backend/port/win32/socket.c), there is a possible socket leak if the socket cannot be made non-blocking.
I don't know Windows well enough to comment on the implications of not calling
closesocket here, but it definitely seems like a prudent thing to do
backpatched down to 12. Unless objections I'll do that.
--
Daniel Gustafsson
Em seg., 15 de jan. de 2024 às 09:43, Daniel Gustafsson <daniel@yesql.se>
escreveu:
On 13 Jan 2024, at 22:38, Ranier Vilela <ranier.vf@gmail.com> wrote:
In the pgwin32_socket function (src/backend/port/win32/socket.c), there
is a possible socket leak if the socket cannot be made non-blocking.
I don't know Windows well enough to comment on the implications of not
calling
closesocket here, but it definitely seems like a prudent thing to do
backpatched down to 12. Unless objections I'll do that.
Thanks for taking care of this.
Do you have plans or should I register for a commitfest?
Best regards,
Ranier Vilela
On Tue, Jan 16, 2024 at 05:25:39PM -0300, Ranier Vilela wrote:
Thanks for taking care of this.
Yeah, that's a good catch.
Do you have plans or should I register for a commitfest?
Daniel has stated that he would take care of it, so why not letting
him a few days? I don't think that a CF entry is necessary.
--
Michael
On 17 Jan 2024, at 07:26, Michael Paquier <michael@paquier.xyz> wrote:
On Tue, Jan 16, 2024 at 05:25:39PM -0300, Ranier Vilela wrote:
Do you have plans or should I register for a commitfest?
Daniel has stated that he would take care of it, so why not letting
him a few days? I don't think that a CF entry is necessary.
It isn't, I've now committed it backpatched down to 12.
--
Daniel Gustafsson
Em qua., 17 de jan. de 2024 09:54, Daniel Gustafsson <daniel@yesql.se>
escreveu:
On 17 Jan 2024, at 07:26, Michael Paquier <michael@paquier.xyz> wrote:
On Tue, Jan 16, 2024 at 05:25:39PM -0300, Ranier Vilela wrote:Do you have plans or should I register for a commitfest?
Daniel has stated that he would take care of it, so why not letting
him a few days? I don't think that a CF entry is necessary.It isn't, I've now committed it backpatched down to 12.
Thanks for the commit, Daniel.
Best regards,
Ranier Vilela