PERFORCE change 18572 for review

[Robert Watson]

http://people.freebsd.org/~peter/p4db/chv.cgi?CH=18572

Change 18572 by rwatson at rwatson_tislabs on 2002/10/02 21:13:55

	Back out apparent incorrect solution to problems with extended
	attributes on specfs vnodes in UFS2.  Vnode operation vectors
	default to eopnotsupp unless overridden, so the change in
	question was essentially a noop.  It also didn't fix the
	problem with EAs, which relates to VOP_WRITE() being implemented
	by specfs not ffs, and therefore the IO_EXT components of the
	write not getting directed to the right implementation.  This
	is one potential problem with choosing to implement extended
	attribute operations using the existing ffs_readwrite
	implementation, as it composes improperly when you have
	pseudo-file systems like specfs.

Affected files ...

.. //depot/projects/trustedbsd/mac/sys/fs/specfs/spec_vnops.c#12 edit

Differences ...

==== //depot/projects/trustedbsd/mac/sys/fs/specfs/spec_vnops.c#12 (text+ko) ====

@@ -107,10 +107,6 @@
 	{ &vop_lock_desc,		(vop_t *) vop_nolock },
 	{ &vop_unlock_desc,		(vop_t *) vop_nounlock },
 	{ &vop_islocked_desc,		(vop_t *) vop_noislocked },
-	{ &vop_getextattr_desc,		(vop_t *) vop_eopnotsupp },
-	{ &vop_setextattr_desc,		(vop_t *) vop_eopnotsupp },
-	{ &vop_openextattr_desc,	(vop_t *) vop_eopnotsupp },
-	{ &vop_closeextattr_desc,	(vop_t *) vop_eopnotsupp },
 	{ NULL, NULL }
 };
 static struct vnodeopv_desc spec_vnodeop_opv_desc =
To Unsubscribe: send mail to majordomo at trustedbsd.org
with "unsubscribe trustedbsd-cvs" in the body of the message