Re: 4.9 Stable Crashes on SuperMicro with SMP

From: Peter Radcliffe (pir_at_pir.net)
Date: 11/27/03

  • Next message: Daniel O'Connor: "Re: fdformat doesn't work anymore?" 
    Date: Thu, 27 Nov 2003 13:45:41 -0500
To: freebsd-stable@freebsd.org

    Jonathan Gilpin <jonathan@fluent.ltd.uk> probably said:
    > Tonight I will enable the second CPU again and see if I can find out
    > more infomation... Could it be that a memory fault only occurr when
    > the second CPU is enabled (ie: when Option SMP and option APIC_IO is
    > enabled in the kernel)?

    Cache problem on the 2nd CPU ?
    Swap the CPUs and run on the other one in single processor mode ?

    P. 

    -- 
pir
_______________________________________________
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"
  • Next message: Daniel O'Connor: "Re: fdformat doesn't work anymore?"

    Relevant Pages

    • Re: Sundays image
      ... the VRM for the second CPU. ... The connector at the junction, ... have no VRM though I have two identical P2/450), 1GB RAM, Matrox Millenium ...
      (comp.sys.ibm.ps2.hardware)
    • Re: Win32 and Win64
      ... Most of the time used in downloading messages from an NNTP ... > is spent waiting for the server to respond with the next block of bytes, ... > little CPU time. ... second CPU at all, ...
      (borland.public.delphi.non-technical)
    • Re: I like SCHED_4BSD
      ... >> measure the scheduler using a tool that is itself run by the ... > That point is taken, however, on my old SMP box _ULE adds an extra ... > the second CPU so that all CPU time is occupied on both CPUs. ...
      (freebsd-current)
    • Re: numbers dont lie ...
      ... Note the effect of the second CPU on the user time. ... i.e. since the hyperthreading virtual CPUs are not actually real CPUs, ... performance depending on your workload. ...
      (freebsd-hackers)
    • Re: [PATCH 2.6] serialization of kernelcapi work
      ... >> of the first one and the second CPU goes on with its work. ... If just a 'flag' is used, it is possible to loose an event, the flag like in ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)