On 11/12/17(Mon) 08:59, Martin Pieuchot wrote:
> This chunk has been #ifdef since r1.1 in 1995. I'd like to revisit
> SB_LOCK so this is in my way. Ok to remove it?
jsg@ pointed out this #if notdef is there since 1990, so 27 years should
be ok to die?
> Index: kern/uipc_usrreq.c
> ===================================================================
> RCS file: /cvs/src/sys/kern/uipc_usrreq.c,v
> retrieving revision 1.120
> diff -u -p -r1.120 uipc_usrreq.c
> --- kern/uipc_usrreq.c 2 Nov 2017 14:01:18 -0000 1.120
> +++ kern/uipc_usrreq.c 11 Dec 2017 07:57:48 -0000
> @@ -963,22 +963,6 @@ unp_gc(void *arg __unused)
> unp->unp_flags |= UNP_GCMARK;
>
> so = unp->unp_socket;
> -#ifdef notdef
> - if (so->so_rcv.sb_flags & SB_LOCK) {
> - /*
> - * This is problematical; it's not clear
> - * we need to wait for the sockbuf to be
> - * unlocked (on a uniprocessor, at least),
> - * and it's also not clear what to do
> - * if sbwait returns an error due to receipt
> - * of a signal. If sbwait does return
> - * an error, we'll go into an infinite
> - * loop. Delete all of this for now.
> - */
> - (void) sbwait(&so->so_rcv);
> - goto restart;
> - }
> -#endif
> unp_scan(so->so_rcv.sb_mb, unp_mark);
> }
> } while (unp_defer);
>
