credit miller@hamnavoe.com from the patch of the same name on sources

note: much of this had already been done.  we'd missed just one before.

In ibrk and segattach there are two different expressions which check for
expanding or newly created segments overlapping an existing segment.  Neither
version was general enough to catch all possible cases.

This patch also includes a minimal correction for a bug reported
in 9fans by cinap_lenrek on 2014-06-08 (Re: duppage) which can
lead to resource starvation & deadlock:

	duppage() causes the freelist to be shuffled differently. without
	stuffing cached pages at the freelist tail, the tail accumulates
	a uncached "stopper" page which breaks the invariant of imagereclaim
	which just scans from the tail backwards as long as the pages are
	cached.

	imagereclaim does not move the pages to the head after uncaching them!
	so by default imagereclaim prevents the cached pages before the ones
	it reclaimed from being reclaimed ever.

Reference: /n/atom/patch/applied/segment-overlap
Date: Mon Jun 20 00:52:11 CES 2016
Signed-off-by: quanstro@quanstro.net

--- /sys/src/9/port/segment.c	Mon Jun 20 00:51:10 2016
+++ /sys/src/9/port/segment.c	Mon Jun 20 00:51:11 2016
@@ -44,7 +44,7 @@
 {
 	Image *i, *ie;
 
-	imagealloc.free = xalloc(conf.nimage*sizeof(Image));
+	imagealloc.free = malloc(conf.nimage*sizeof(Image));
 	if (imagealloc.free == nil)
 		panic("initseg: no memory");
 	ie = &imagealloc.free[conf.nimage-1];
@@ -624,7 +624,7 @@
 		ns = p->seg[i];
 		if(ns == 0)
 			continue;
-		if((newtop > ns->base && newtop <= ns->top) ||
+		if((newtop >= ns->base && newtop <= ns->top) ||
 		   (va >= ns->base && va < ns->top))
 			return ns;
 	}
--- /sys/src/nix/port/segment.c	Mon Jun 20 00:51:13 2016
+++ /sys/src/nix/port/segment.c	Mon Jun 20 00:51:14 2016
@@ -329,7 +329,7 @@
 		ns = p->seg[i];
 		if(ns == 0)
 			continue;
-		if((newtop > ns->base && newtop <= ns->top) ||
+		if((newtop >= ns->base && newtop <= ns->top) ||
 		   (va >= ns->base && va < ns->top))
 			return ns;
 	}