From owner-cvs-all  Mon Oct 19 00:36:55 1998
Return-Path: <owner-cvs-all@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id AAA29694
          for cvs-all-outgoing; Mon, 19 Oct 1998 00:36:55 -0700 (PDT)
          (envelope-from owner-cvs-all@FreeBSD.ORG)
Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id AAA29678;
          Mon, 19 Oct 1998 00:36:53 -0700 (PDT)
          (envelope-from dillon@apollo.backplane.com)
Received: (from dillon@localhost)
	by apollo.backplane.com (8.9.1/8.9.1) id AAA12327;
	Mon, 19 Oct 1998 00:36:30 -0700 (PDT)
	(envelope-from dillon)
Date: Mon, 19 Oct 1998 00:36:30 -0700 (PDT)
From: Matthew Dillon <dillon@apollo.backplane.com>
Message-Id: <199810190736.AAA12327@apollo.backplane.com>
To: cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG
Subject: Anyone got any ideas?  getblk/pgtblk/inode lockup
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk

    I've been trying to track this down for a month or two now... it's
    the only regular crash I get under -current these days.

    What happens is that a process gets stuck in pgtblk (kern/vfs_bio.c
    line 1800 or so).  This causes a cascade of processes locking up
    until the system goes crunch.  The lockup occurs about once a week on
    a heavily loaded machine with lots of processes that use mmap() heavily
    (i.e. running Diablo).

    The code in question is this:

	   } else if (m->flags & PG_BUSY) {
		    s = splvm();
		    if (m->flags & PG_BUSY) {
			    vm_page_flag_set(m, PG_WANTED);
			    tsleep(m, PVM, "pgtblk", 0);
		    }
		    splx(s);
		    goto doretry;
	    } ...

    Looking at the kernel dump, the page structure 'm' in the process locked 
    up in pgtblk has PG_WANTED and PG_TABLED set, but PG_BUSY cleared.

    This combined with the code fragment above leads me to believe that 
    PG_BUSY is somehow being cleared on the page without waking up the
    process.

(kgdb) print *m
$3 = {
  pageq = {
    tqe_next = 0xf0f46024, 
    tqe_prev = 0xf0ee7438
  }, 
  hashq = {
    tqe_next = 0xf0f46024, 
    tqe_prev = 0xf0c3df50
  }, 
  listq = {
    tqe_next = 0xf0f184e4, 
    tqe_prev = 0xf0eb0d14
  }, 
  object = 0xfc4e3440, 
  pindex = 0x90, 
  phys_addr = 0x35ed000, 
  queue = 0x82, 
  flags = 0x6, 				PG_WANTED, PG_TABLED
  pc = 0x2d, 
  wire_count = 0x0, 
  hold_count = 0x0, 
  act_count = 0x5, 
  busy = 0x1, 
  valid = 0x0, 
  dirty = 0x0
}

    I have other kernel dumps that show the same thing.   The page flags are
    similar, except I see 0x46 sometimes (PG_ZERO|PG_TABLED|PG_WANTED).

    It's virtually the only crash I've gotten with FreeBSD-current so far.

    If anyone has any ideas, I'm all ears.  I've been chasing this one for a
    few months.

						-Matt

    Matthew Dillon  Engineering, HiWay Technologies, Inc. & BEST Internet 
                    Communications & God knows what else.
    <dillon@backplane.com> (Please include original email in any response)    

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