From owner-freebsd-current  Wed Feb 11 14:20:54 1998
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id OAA02502
          for current-outgoing; Wed, 11 Feb 1998 14:20:54 -0800 (PST)
          (envelope-from owner-freebsd-current@FreeBSD.ORG)
Received: from lamb.sas.com (root@lamb.sas.com [192.35.83.8])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA02497
          for <freebsd-current@freebsd.org>; Wed, 11 Feb 1998 14:20:49 -0800 (PST)
          (envelope-from jwd@unx.sas.com)
Received: from mozart (markham.southpeak.com [192.58.185.8])
	by lamb.sas.com (8.8.7/8.8.7) with SMTP id RAA20368
	for <freebsd-current@freebsd.org>; Wed, 11 Feb 1998 17:20:19 -0500 (EST)
Received: from iluvatar.unx.sas.com by mozart (5.65c/SAS/Domains/5-6-90)
	id AA01189; Wed, 11 Feb 1998 17:20:19 -0500
From: "John W. DeBoskey" <jwd@unx.sas.com>
Received: by iluvatar.unx.sas.com (5.65c/SAS/Generic 9.01/3-26-93)
	id AA20790; Wed, 11 Feb 1998 17:20:19 -0500
Message-Id: <199802112220.AA20790@iluvatar.unx.sas.com>
Subject: NFS V3 page caching & vm_page_is_valid()
To: freebsd-current@FreeBSD.ORG
Date: Wed, 11 Feb 1998 17:20:18 -0500 (EST)
Cc: jwd@unx.sas.com (John W. DeBoskey)
X-Mailer: ELM [version 2.4 PL23]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Hi,

   For this test, I am cat'ing a 16385 byte file from an NFS V3
fileserver. ie: 2 8k blocks and a 1byte frag.

   The completely full pages are cached correctly.

   The last partial page is not. As I trace nfs_bioread() into brelse()
and then into vm_page_is_valid(), I find that m->valid is 0.

(kgdb) p m->valid
$12 = 0 '\000'
(kgdb) p bits
$13 = 1

   Thus, the following statement:

   if (m->valid && ((m->valid & bits) == bits))
      return 1;
   else
      return 0

   is false, and we are returning 0. Unfortunately, there is 1 single
solitary valid byte in this buffer. 

   Due to the above, the buffer is tossed causing an nfs readrpc to
the server to read the tail of the file everytime. Yuck.


   I am continuing to trace this problem, but any hints about where
to look for how the decision to set m->valid is made  would be much
appreciated.

Thanks,
John

ps: Thanks for pushing the logical vs bitwise and fix.
   
-- 
jwd@unx.sas.com       (w) John W. De Boskey          (919) 677-8000 x6915

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe current" in the body of the message