i386/59898: pxe boot: BTX halted

[nkoch at demig.de]

>Number:         59898
>Category:       i386
>Synopsis:       pxe boot: BTX halted
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-i386
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue Dec 02 02:00:40 PST 2003
>Closed-Date:
>Last-Modified:
>Originator:     Norbert Koch
>Release:        FreeBSD 4.9-STABLE i386
>Organization:
demig Prozessautomatisierung GmbH
>Environment:
System: FreeBSD entw-pr.demig 4.9-STABLE FreeBSD 4.9-STABLE #0: Fri Nov 28 09:49:46 CET 2003 root at entw-pr.demig:/usr/obj/usr/src/sys/ENTW-PR i386
        isc-dhcp3-3.0.1.r12


>Description:
	I am trying to pxe boot a client machine with an Intel 8255xER chip
        installed. I configured dhcpd, tftpd and nfs/mountd. BTX _always_ halts
        with an int 6 or d or e. The cs:eip dump often - not always - shows
        a number of ff's. I built pxeboot with -DBTXLDR_VERBOSE. The last
        information from screen is:
          Relocated kernel (size=740) to 9000
          Client base address is a000
          Client format is ELF
          text segment: offset=1000 vaddr=0 filesz=1f7a0 memsz=1f7a0
          data segment: offset=207a0 vaddr=207a0 filesz=2cd4 memsz=8d40
          Loading complete
          Console: internal video/keyboard
          BIOS drive C: is disk0

          PXE version 2.1, real mode entry point @9da3:0106
          BIOS 639kb/14336kb available memory

          FreeBSD/i386 bootstrap loader, Revision 0.8
          (root at entw-pr.demig, Tue Dec 2 10:02:02 CET 2003)
          pxe_open: server addr: 10.47.11.1
          pxe_open: server path: /usr/local/diskless_root
          pxe_open: gateway ip: 10.47.11.1
          \
          int= ... (int/eip/dump differing)
          ...
          BTX halted

        The client machine is a PentiumMMX/166MHz which successfully
        boots FreeBSD from hard disk. It is connected to the server
        machine's second ethernet card xl0 (3com).

        When I run ethereal on the server machine I see successful
        dhcp, tftp and nfs requests. No obvious problems.

        The boot loader sometimes hangs immediately, sometimes I can
        press ESC and can enter the loader command line.
        I see _no_ kernel file being loaded.
        
        It makes no difference if mountd is running and serving or not.
        
        Configuration files:
        
          /usr/local/etc/dhcpd.conf:
          
            default-lease-time 600;
            max-lease-time 7200;
            ddns-update-style ad-hoc;
            log-facility local7;
            subnet 10.47.11.0 netmask 255.255.255.0 {
              range 10.47.11.10 10.47.11.99;
              option routers 10.47.11.1;
              filename "pxeboot";
              next-server 10.47.11.1;
              option root-path "10.47.11.1:/usr/local/diskless_root";
            }

          /etc/exports:
            /usr -ro -maproot=0 -alldirs -network 10.47.11.00 -mask 255.255.255.0

          /etc/inetd.conf:
            tftp dgram  udp wait root /usr/libexec/tftpd tftp /local/diskless_root
            tftp stream tcp wait root /usr/libexec/tftpd tftp /local/diskless_root

        The client machine comes up with "Intel UNDI PXE-2.0 (build 074).
        I know that pxeboot(8) says that there are problems with some
        2.0 boot proms but the board manufacturer tells me that he has no
        newer version available.

        I read in a newsgroup that someone with the same problem was
        successful after changing /sys/boot/i386/libi386/pxe.c.
        He re-inserted the delay() lines which are commented out and raised
        the time to 10000. But that does not work for me.

>How-To-Repeat:
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: