A partial solution to some of the 'pull the umass device with a

mounted FS' problems.  These are more along the lines of 'avoiding an
avoidable panic' than a complete solution to removable devices.  We
now close the barn door after the horse has gotten lose and has been
hit by a truck, as it were.  The barn no longer catches fire in this
case, but the horse is still dead :-).

The vfs_bio.c fix causes us not to put a failed write back into the
dirty pool if the error returned was ENXIO.  In that case, the buffer
is treated like any other clean buffer that's being retured.  ENXIO
means the device isn't there anymore and will never be there again in
the future, so retrying is futile.

The vfs_mount.c fix treats 'ENXIO' as success for unmounting a file
system.  If the device is gone, retrying later won't help and we'll
never be able to unmount the device.

These two are part of a larger patch set submitted by the author.  The
other patches will be forth coming.  I added comments to these two
patches.

Submitted by: Henrik Gulbrandsen
Reviewed by: phk@
PR: usb/46176 (partial)

2 files changed, 11 insertions, 2 deletions

diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 444c2bef2608..b02d65b384f8 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -1170,6 +1170,7 @@ brelse(struct buf *bp)
 
 	if (bp->b_iocmd == BIO_WRITE &&
 	    (bp->b_ioflags & BIO_ERROR) &&
+	    bp->b_error != ENXIO &&
 	    !(bp->b_flags & B_INVAL)) {
 		/*
 		 * Failed write, redirty.  Must clear BIO_ERROR to prevent
@@ -1177,6 +1178,9 @@ brelse(struct buf *bp)
 		 * this case is not run and the next case is run to 
 		 * destroy the buffer.  B_INVAL can occur if the buffer
 		 * is outside the range supported by the underlying device.
+		 * If the error is that the device went away (ENXIO), we
+		 * shouldn't redirty the buffer either, but discard the
+		 * data too.
 		 */
 		bp->b_ioflags &= ~BIO_ERROR;
 		bdirty(bp);
diff --git a/sys/kern/vfs_mount.c b/sys/kern/vfs_mount.c
index dfb46e29c8a9..6a615b1e795c 100644
--- a/sys/kern/vfs_mount.c
+++ b/sys/kern/vfs_mount.c
@@ -1286,8 +1286,13 @@ dounmount(mp, flags, td)
 		error = VFS_UNMOUNT(mp, flags, td);
 	}
 	vn_finished_write(mp);
-	if (error) {
-		/* Undo cdir/rdir and rootvnode changes made above. */
+	/*
+	 * If we failed to flush the dirty blocks for this mount point,
+	 * undo all the cdir/rdir and rootvnode changes we made above.
+	 * Unless we failed to do so because the device is reporting that
+	 * it doesn't exist anymore.
+	 */
+	if (error && error != ENXIO) {
 		if ((flags & MNT_FORCE) &&
 		    VFS_ROOT(mp, LK_EXCLUSIVE, &fsrootvp, td) == 0) {
 			if (mp->mnt_vnodecovered != NULL)