Re: Regression 7.0R -> 7-stable? - FreeBSD

This is a discussion on Re: Regression 7.0R -> 7-stable? - FreeBSD ; On Tue, 7 Oct 2008 13:38:01 +0200 Gerrit Kühn wrote about Re: Regression 7.0R -> 7-stable?: [...] GK> However, yesterday I upgraded the system to a recent 7-stable GK> codebase, and now it locks again hard after probing the CPU ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: Re: Regression 7.0R -> 7-stable?

  1. Re: Regression 7.0R -> 7-stable?

    On Tue, 7 Oct 2008 13:38:01 +0200 Gerrit Kühn
    wrote about Re: Regression 7.0R -> 7-stable?:

    [...]

    GK> However, yesterday I upgraded the system to a recent 7-stable
    GK> codebase, and now it locks again hard after probing the CPU cores. My
    GK> setup remained exactly the same, the new code just does not work.
    GK> Please let me know what I have do to provide further information.

    Well, comparing the kernel setups again I found one thing that makes a
    difference: The scheduler (somewhen SCHED_ULE has been declared the
    default).
    SCHED_ULE is crashing the system, SCHED_4BSD works fine...


    cu
    Gerrit
    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/lis...freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"


  2. Re: Regression 7.0R -> 7-stable?

    On Tuesday 07 October 2008 09:37:10 am Jeremy Chadwick wrote:
    > On Tue, Oct 07, 2008 at 02:37:24PM +0200, Gerrit Kühn wrote:
    > > On Tue, 7 Oct 2008 13:38:01 +0200 Gerrit Kühn
    > > wrote about Re: Regression 7.0R -> 7-stable?:
    > >
    > > [...]
    > >
    > > GK> However, yesterday I upgraded the system to a recent 7-stable
    > > GK> codebase, and now it locks again hard after probing the CPU cores. My
    > > GK> setup remained exactly the same, the new code just does not work.
    > > GK> Please let me know what I have do to provide further information.
    > >
    > > Well, comparing the kernel setups again I found one thing that makes a
    > > difference: The scheduler (somewhen SCHED_ULE has been declared the
    > > default).
    > > SCHED_ULE is crashing the system, SCHED_4BSD works fine...

    >
    > This is of great concern if in fact that's true.
    >
    > I'm adding Jeff Roberson, author of the ULE/SMP2.0 code, to the CC.


    Do you have more details about the crash? Are you getting an actual panic
    with messages on the console, or are you still seeing hangs? When you get a
    hang, can you break into the debugger and get a crash dump?

    --
    John Baldwin
    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/lis...freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"


  3. Re: Regression 7.0R -> 7-stable?

    On Tue, 7 Oct 2008 10:02:56 -0400 John Baldwin wrote
    about Re: Regression 7.0R -> 7-stable?:

    JB> Do you have more details about the crash? Are you getting an actual
    JB> panic with messages on the console, or are you still seeing hangs?

    Like it was before: system just hangs after displaying the probing
    messages about the CPU cores; next step for a working kernel would be
    mounting of the file systems (and changing from white kernel output to grey
    system output).

    JB> When you get a hang, can you break into the debugger and get a crash
    JB> dump?

    Is there a documentation somewhere how to do this?


    cu
    Gerrit
    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/lis...freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"


  4. Re: Regression 7.0R -> 7-stable?

    On Tue, Oct 07, 2008 at 04:15:34PM +0200, Gerrit Kühn wrote:
    > On Tue, 7 Oct 2008 10:02:56 -0400 John Baldwin wrote
    > about Re: Regression 7.0R -> 7-stable?:
    >
    > JB> Do you have more details about the crash? Are you getting an actual
    > JB> panic with messages on the console, or are you still seeing hangs?
    >
    > Like it was before: system just hangs after displaying the probing
    > messages about the CPU cores; next step for a working kernel would be
    > mounting of the file systems (and changing from white kernel output to grey
    > system output).


    Actually, I think you mean mounting of the root filesystem, do you not?

    If so, others have recently reported this problem (hard lock-ups before
    or after printing "Mounting root from...").

    > JB> When you get a hang, can you break into the debugger and get a crash
    > JB> dump?
    >
    > Is there a documentation somewhere how to do this?


    John can probably help you with the commands you need to type, but the
    FreeBSD Handbook goes over the general commands.

    As far as getting into the debugger, it's Control-Alt-Esc from the
    console.

    --
    | Jeremy Chadwick jdc at parodius.com |
    | Parodius Networking http://www.parodius.com/ |
    | UNIX Systems Administrator Mountain View, CA, USA |
    | Making life hard for others since 1977. PGP: 4BD6C0CB |

    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/lis...freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"


  5. Re: Regression 7.0R -> 7-stable?

    On Tue, 7 Oct 2008 07:25:42 -0700 Jeremy Chadwick
    wrote about Re: Regression 7.0R -> 7-stable?:

    JC> > Like it was before: system just hangs after displaying the probing
    JC> > messages about the CPU cores; next step for a working kernel would
    JC> > be mounting of the file systems (and changing from white kernel
    JC> > output to grey system output).

    JC> Actually, I think you mean mounting of the root filesystem, do you not?
    JC> If so, others have recently reported this problem (hard lock-ups
    JC> before or after printing "Mounting root from...").

    Yes, of course (sorry for being vague).

    JC> > JB> When you get a hang, can you break into the debugger and get a
    JC> > JB> crash dump?

    JC> > Is there a documentation somewhere how to do this?

    JC> John can probably help you with the commands you need to type, but the
    JC> FreeBSD Handbook goes over the general commands.
    JC> As far as getting into the debugger, it's Control-Alt-Esc from the
    JC> console.

    Ok, I added options KDB and DDB to my kernel configuration and compiled
    with SCHED_ULE. However, after hanging the system does not react on
    Ctrl-Alt-Esc. Am I missing something?


    cu
    Gerrit
    _______________________________________________
    freebsd-stable@freebsd.org mailing list
    http://lists.freebsd.org/mailman/lis...freebsd-stable
    To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"


+ Reply to Thread