port-shark/43652: savecore does not work

[jmmv%netbsd.org@localhost]

>Number:         43652
>Category:       port-shark
>Synopsis:       savecore does not work
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    port-shark-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jul 22 10:55:00 +0000 2010
>Originator:     Julio Merino
>Release:        NetBSD 5.99.30
>Organization:
Julio Merino
>Environment:
        
        
System: NetBSD shark 5.99.30 NetBSD 5.99.30 (GENERIC) #0: Sat Jun 19 12:22:43 
IST 2010  
jmmv@...:/home/jmmv/os/netbsd/obj.shark/home/jmmv/os/netbsd/src/sys/arch/shark/compile/GENERIC
 shark
Architecture: i386
Machine: i386
>Description:
        Setting as the highest priority because non-working kernel dumps make
        crashes very hard to debug.

        Copying email from port-arm::

        http://mail-index.netbsd.org/port-arm/2010/06/20/msg000984.html

        -----
        I've just updated my DNARD machine with a -current build of shark and,
        when running the tests, the system panics and reboots.  I can see the
        text below in the console, and it looks like the kernel correctly
        dumped core... however, when savecore starts after the reboot, it does
        not recover the kernel core (and it doesn't show any messages at all).

        panic: pool_destroy: pool busy: still out: 1
        cpu0: Begin traceback...
        0xf1f78ef0
        scp=0xf1f78ef0 rlv=0xf1f78ebc (0xf1f78ebc)
        rsp=0xf1f78ec4 rfp=0xf0183488
        Bad frame pointer: 0xf0183488
        cpu0: End traceback...
        boot: howto=00000104 curlwp=0xf1d32ca0
        current_mask=0000c3b8
        ipl_bio=ffff3d5f ipl_net=ffff3d5f ipl_tty=ffff3d5f ipl_vm=ffff3d5f
        ipl_audio=ffff3d5f ipl_clock=ffff3c5f ipl_none=ffffffff
        spl_masks[0]=ffffffff
        spl_masks[1]=ffffffff
        spl_masks[2]=ffffffff
        spl_masks[3]=ffffffff
        spl_masks[4]=ffffffff
        spl_masks[5]=ffff3d5f
        spl_masks[6]=ffff3c5f
        spl_masks[7]=ffff3c47

        dumping to dev 16,1 offset 75287
        dump 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
        25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
        48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
        71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
        succeeded
        -----

        And the only answer:

        -----
        I thought I'd fixed this issue some time ago:
        http://mail-index.netbsd.org/source-changes/2008/01/01/0014.html
        http://mail-index.netbsd.org/source-changes/2008/01/01/0016.html

        Nick then fixed up gdb, and tweaked the stack layout:
        http://mail-index.netbsd.org/source-changes/2008/01/12/0018.html
        http://mail-index.netbsd.org/source-changes/2008/01/12/0019.html

        So it should work.

        My first thought is that perhaps shark isn't using something from the
        above, and has it's own version due to having to interact with OFW.

        Does savecore -v show anything useful?
        -----

>How-To-Repeat:
        Make the system crash (atf-run from /usr/tests should be enough).
        Let it dump core, reboot, and see that /var/crash is empty.
>Fix:
        

>Unformatted: