Re: git: 454212b9718b - main - sctp: fix so_proto when peeling off a socket

In reply to: Gleb Smirnoff : "Re: git: 454212b9718b - main - sctp: fix so_proto when peeling off a socket"
Go to: [ bottom of page ] [ top of archives ] [ this month ]

From: Michael Tuexen <tuexen_at_FreeBSD.org>
Date: Sat, 28 Feb 2026 06:50:21 UTC

> On 28. Feb 2026, at 02:44, Gleb Smirnoff <glebius@FreeBSD.org> wrote:
> 
> On Fri, Feb 27, 2026 at 08:11:39AM +0100, Michael Tuexen wrote:
> M> >> - so = sopeeloff(head);
> M> >> + so = sopeeloff(head, &sctp_stream_protosw);
> M> > 
> M> > This part of the change broke LINT-NOINET because sctp_stream_protosw is only defined in the INET case. I didn’t dig into the SCTP syscalls too far, but I think that the right protocol needs to be picked per the address family, right?
> 
> btw, I tried to move the function to sctp_syscalls.c and this required exposing
> soattach() and soalloc() outside.  After weighting in and out decided to leave
> it the way it is.
I tried that also and came to the same conclusion. That is why I added the
parameter to sopeeloff().

Best regards
Michael
> 
> -- 
> Gleb Smirnoff