Skip to content
Snippets Groups Projects
  1. Oct 19, 2007
  2. Oct 17, 2007
  3. Feb 21, 2007
  4. Feb 12, 2007
  5. Jan 23, 2007
  6. Jan 12, 2007
  7. Oct 03, 2006
  8. Jun 26, 2006
  9. Jun 25, 2006
  10. Jan 12, 2006
  11. Jan 10, 2006
  12. Sep 13, 2005
  13. Sep 09, 2005
  14. Jun 25, 2005
    • Vivek Goyal's avatar
      [PATCH] kdump documentation update to introduce use of irqpoll · d58831e4
      Vivek Goyal authored
      
      o Specify "irqpoll" command line option which loading second kernel. This
        helps in reducing driver initialization failures in second kernel due
        to shared interrupts.
      o Enabled LAPIC/IOAPIC support for UP kernels in second kernel. This reduces
        the chances of devices sharing the irq and hence reduces the chances of
        driver initialization failures in second kernel.
      o Build a UP capture kernel and disabled SMP support.
      
      Signed-off-by: default avatarVivek Goyal <vgoyal@in.ibm.com>
      Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@osdl.org>
      d58831e4
    • Vivek Goyal's avatar
      [PATCH] kdump: Documentation for Kdump · b089f4a6
      Vivek Goyal authored
      
      This patch contains the documentation for the kexec based crash dump tool.
      
      Quick kdump-howto
      ================================================================
      
      1) Download and build kexec-tools.
      
      2) Download and build the latest kexec/kdump (-mm) kernel patchset.
         Two kernels need to be built in order to get this feature working.
      
        A) First kernel:
         a) Enable "kexec system call" feature:
      	CONFIG_KEXEC=y
         b) Physical load address (use default):
      	CONFIG_PHYSICAL_START=0x100000
         c) Enable "sysfs file system support":
      	CONFIG_SYSFS=y
         d) Boot into first kernel with the command line parameter "crashkernel=Y@X":
            For example: "crashkernel=64M@16M".
      
        B) Second kernel:
         a) Enable "kernel crash dumps" feature:
      	CONFIG_CRASH_DUMP=y
         b) Physical load addreess, use same load address as X in "crashkernel"
            kernel parameter in d) above, e.g., 16 MB or 0x1000000.
      	CONFIG_PHYSICAL_START=0x1000000
         c) Enable "/proc/vmcore support" (Optional, in Pseudo filesystems).
      	CONFIG_PROC_VMCORE=y
      
      3) Boot into the first kernel.
      
      4) Load the second kernel to be booted using:
      
         kexec -p <second-kernel> --crash-dump --args-linux --append="root=<root-dev>
         maxcpus=1 init 1"
      
      5) System reboots into the second kernel when a panic occurs. A module can be
         written to force the panic, for testing purposes.
      
      6) See Documentation/kdump.txt for how to read the first kernel's
         memory image and how to analyze it.
      
      Signed-off-by: default avatarHariprasad Nellitheertha <hari@in.ibm.com>
      Signed-off-by: default avatarEric Biederman <ebiederm@xmission.com>
      Signed-off-by: default avatarVivek Goyal <vgoyal@in.ibm.com>
      Signed-off-by: default avatarrandy_dunlap <rdunlap@xenotime.net>
      Signed-off-by: default avatarManeesh Soni <maneesh@in.ibm.com>
      Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@osdl.org>
      b089f4a6
Loading