Sun Jul 31 04:05:32 2016 UTC ()

In bwrite, add assertion that vp != NULL. (vp is the vnode from the
buffer being written.)

There's some logic here that carefully checks for vp being null, and
other logic that will crash if it is. It appears that it's all
needless paranoia. See tech-kern for more info.

Unless someone sees the assertion go off (in which case a lot more
investigation is needed) I or someone will clean out the logic at some
future point.

Spotted by coypu.


(dholland)

diff -r1.259 -r1.260 src/sys/kern/vfs_bio.c

--- src/sys/kern/vfs_bio.c 2016/02/01 05:05:43	1.259
+++ src/sys/kern/vfs_bio.c 2016/07/31 04:05:32	1.260

 @@ -1,14 +1,14 @@
-/*	$NetBSD: vfs_bio.c,v 1.259 2016/02/01 05:05:43 riz Exp $	*/
+/*	$NetBSD: vfs_bio.c,v 1.260 2016/07/31 04:05:32 dholland Exp $	*/
 /*-
  * Copyright (c) 2007, 2008, 2009 The NetBSD Foundation, Inc.
  * All rights reserved.
+ *
  * This code is derived from software contributed to The NetBSD Foundation
  * by Andrew Doran, and by Wasabi Systems, Inc.
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
 @@ -113,27 +113,27 @@
  * - bufcache_lock: protects global buffer cache state.
  * - BC_BUSY: a long term per-buffer lock.
  * - buf_t::b_objlock: lock on completion (biowait vs biodone).
+ *
  * For buffers associated with vnodes (a most common case) b_objlock points
  * to the vnode_t::v_interlock.  Otherwise, it points to generic buffer_lock.
+ *
  * Lock order:
  *	bufcache_lock ->
  *		buf_t::b_objlock
  */
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: vfs_bio.c,v 1.259 2016/02/01 05:05:43 riz Exp $");
+__KERNEL_RCSID(0, "$NetBSD: vfs_bio.c,v 1.260 2016/07/31 04:05:32 dholland Exp $");
 #ifdef _KERNEL_OPT
 #include "opt_bufcache.h"
 #include "opt_dtrace.h"
 #endif
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/kernel.h>
 #include <sys/proc.h>
 #include <sys/buf.h>
 #include <sys/vnode.h>
 #include <sys/mount.h>
 @@ -790,26 +790,34 @@ breadn(struct vnode *vp, daddr_t blkno,
  * Block write.  Described in Bach (p.56)
  */
 int
 bwrite(buf_t *bp)
+{
 	int rv, sync, wasdelayed;
 	struct vnode *vp;
 	struct mount *mp;
 	KASSERT(ISSET(bp->b_cflags, BC_BUSY));
 	KASSERT(!cv_has_waiters(&bp->b_done));
 	vp = bp->b_vp;
 	/*
 	 * dholland 20160728 AFAICT vp==NULL must be impossible as it
 	 * will crash upon reaching VOP_STRATEGY below... see further
 	 * analysis on tech-kern.
 	 */
 	KASSERTMSG(vp != NULL, "bwrite given buffer with null vnode");
 	if (vp != NULL) {
 		KASSERT(bp->b_objlock == vp->v_interlock);
 		if (vp->v_type == VBLK)
 			mp = spec_node_getmountedfs(vp);
 		else
 			mp = vp->v_mount;
 	} else {
 		mp = NULL;
+	}
 	if (mp && mp->mnt_wapbl) {
 		if (bp->b_iodone != mp->mnt_wapbl_op->wo_wapbl_biodone) {
 			bdwrite(bp);