0 Replies Latest reply: Jul 12, 2011 5:14 PM by 871961 RSS

    lsimega: redzone violation, then panic as kernel heap corruption detected

    871961
      Has anyone encountered this ? Upon reboot after setting kmem_flags=0xf :

      ...
      sgen19 at lsimega0: target 45 lun 0
      sgen19 is /pci@1,0/pci1028,518@6/sgen@45,0
      sgen20 at lsimega0: target 46 lun 0
      sgen20 is /pci@1,0/pci1028,518@6/sgen@46,0
      st20 at lsimega0: target 46 lun 0
      st20 is /pci@1,0/pci1028,518@6/st@46,0
      /pci@1,0/pci1028,518@6/st@46,0 (st20):
      <HP Ultrium LTO>
      kernel memory allocator:
      redzone violation: write past end of buffer
      buffer=e4491a40 bufctl=e4589488 cache: kmem_alloc_768
      previous transaction on buffer e4491a40:
      thread=dc135da0 time=T-9.220405861 slab=e4460450 cache: kmem_alloc_768
      kmem_cache_alloc_debug+252
      NOTICE: e1000g1 unregistered
      kmem_cache_alloc+205
      kmem_zalloc+57
      scsi_hba_pkt_alloc+ae
      mega_scsi_init_pkt+46
      scsi_init_pkt+87
      st_doattach+d9
      st_attach+68
      devi_attach+a5
      attach_node+9a
      i_ndi_config_node+c1
      i_ddi_attachchild+3d
      devi_attach_node+bb
      config_immediate_children+e6
      ndi_busop_bus_config+74

      panic[cpu3]/thread=de2e8da0:
      kernel heap corruption detected


      de2e8a38 genunix:kmem_error+4d5 (1, cb829c88, e4491a)
      de2e8a78 genunix:kmem_free+1f2 (e4491a40, 290, e425)
      de2e8a98 scsi:scsi_hba_pkt_free+12 (e4491a44, e4491a40,)
      de2e8ab8 lsimega:mega_scsi_destroy_pkt+46 (e4491a44, e4491a40,)
      de2e8ad8 scsi:scsi_destroy_pkt+19 (e4491a40)
      de2e8b08 st:st_detach+4d8 (d7f87418)
      de2e8b48 genunix:devi_detach+7a (d7f87418, 0, de2e8b)
      de2e8b78 genunix:detach_node+5b (d7f87418, 2000, de2)
      de2e8bb8 genunix:i_ndi_unconfig_node+d0 (d7f87418, 4, 2000, )
      de2e8bd8 genunix:i_ddi_detachchild+1f (d7f87418, 2000, de2)
      de2e8c18 genunix:devi_detach_node+b8 (d7f87418, 2000, de2)
      de2e8c68 genunix:unconfig_immediate_children+196 (ce529860, 0, 2000, )
      de2e8c98 genunix:ndi_busop_bus_unconfig+5b (ce529860, 2000, 7, )
      de2e8cc8 scsi:scsi_hba_bus_unconfig_spi+c8 (ce529860, 2000, 7, )
      de2e8cf8 scsi:scsi_hba_bus_unconfig+e9 (ce529860, 2000, 7, )
      de2e8d38 genunix:devi_unconfig_common+dc (ce529860, 0, 2000, )
      de2e8d88 genunix:mt_config_thread+89 (d7f79920, 0)
      de2e8d98 unix:thread_start+8 ()

      syncing file systems...
      done
      dumping to /dev/zvol/dsk/rpool/dump, offset 65536, content: kernel

      The server is a Dell PowerEdge 2650, and the tape machine is a PowerVault 122T, but this looks like a software problem.
      If it is a memory leak then that might explain why the system suffers from soft hangs.

      # uname -a
      SunOS backupsrv02 5.11 snv_151a i86pc i386 i86pc Solaris
      # modinfo | grep lsimega
      192 f63ac000 78f0 231 1 lsimega (LSI MegaRAID 2.05.06)