crash version 4.0-3.3 is available

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



-  Addressed a number of issues associated with CONFIG_SPARSEMEM
   kernels and kernels using updated manners for the linkage of
   their pglist_data structures, and pointers to their mem_map arrays.
   (anderson@xxxxxxxxxx)

-  Implemented "kmem -n" for CONFIG_SPARSEMEM kernels; in addition
   to the pgdat- and zone-related data command output, it also
   displays a list of the SPARSEMEM mem_sections.  Here is an
   example from an ia64:
 
    crash> kmem -n
    NODE    SIZE      PGLIST_DATA       BOOTMEM_DATA       NODE_ZONES
      0   2359296   e000000008c00000  a000000100749b70  e000000008c00000
                                                        e000000008c02400
                                                        e000000008c04800
                                                        e000000008c06c00
        MEM_MAP       START_PADDR  START_MAPNR
    e0000001040a3f00       0            0
 
    ZONE  NAME         SIZE       MEM_MAP      START_PADDR  START_MAPNR
      0   DMA        262144  e0000001040a3f00            0            0
      1   DMA32           0                 0            0            0
      2   Normal    2097152  e0000001048a3f00    100000000       262144
      3   HighMem         0                 0            0            0
 
    -------------------------------------------------------------------
 
    NR      SECTION        CODED_MEM_MAP        MEM_MAP       PFN
     0  e00000010409ff00  e0000001040a3f00  e0000001040a3f00  0
     1  e00000010409ff08  e0000001040a3f00  e0000001044a3f00  65536
     4  e00000010409ff20  e0000001038a3f00  e0000001048a3f00  262144
     5  e00000010409ff28  e0000001038a3f00  e000000104ca3f00  327680
     6  e00000010409ff30  e0000001038a3f00  e0000001050a3f00  393216
     7  e00000010409ff38  e0000001038a3f00  e0000001054a3f00  458752
     8  e00000010409ff40  e0000001038a3f00  e0000001058a3f00  524288
     9  e00000010409ff48  e0000001038a3f00  e000000105ca3f00  589824
    10  e00000010409ff50  e0000001038a3f00  e0000001060a3f00  655360
    11  e00000010409ff58  e0000001038a3f00  e0000001064a3f00  720896
    12  e00000010409ff60  e0000001038a3f00  e0000001068a3f00  786432
    13  e00000010409ff68  e0000001038a3f00  e000000106ca3f00  851968
    14  e00000010409ff70  e0000001038a3f00  e0000001070a3f00  917504
    15  e00000010409ff78  e0000001038a3f00  e0000001074a3f00  983040
    16  e00000010409ff80  e0000001038a3f00  e0000001078a3f00  1048576
    17  e00000010409ff88  e0000001038a3f00  e000000107ca3f00  1114112
    18  e00000010409ff90  e0000001038a3f00  e0000001080a3f00  1179648
    19  e00000010409ff98  e0000001038a3f00  e0000001084a3f00  1245184
    20  e00000010409ffa0  e0000001038a3f00  e0000001088a3f00  1310720
    21  e00000010409ffa8  e0000001038a3f00  e000000108ca3f00  1376256
    22  e00000010409ffb0  e0000001038a3f00  e0000001090a3f00  1441792
    23  e00000010409ffb8  e0000001038a3f00  e0000001094a3f00  1507328
    34  e0000001040a0010  e0000001010a3f00  e0000001098a3f00  2228224
    35  e0000001040a0018  e0000001010a3f00  e000000109ca3f00  2293760
    crash>
   (anderson@xxxxxxxxxx)

-  Fix for "kmem -i" failure in CONFIG_SPARSEMEM kernels that would
   typically fail with the error message: "kmem: invalid kernel virtual
   address: 0  type: node_zones free_pages". (anderson@xxxxxxxxxx)

-  Fix for "kmem -f" failure in CONFIG_SPARSEMEM kernels that would
   typically fail with the error message: "kmem: invalid kernel virtual
   address: ab8  type: node_zones name".  (anderson@xxxxxxxxxx)

-  Fix for "kmem -f" failure in 2.6.17 kernels (possibly earlier) that
   would fail with the error message: "kmem: invalid structure member
   offset: zone_zone_mem_map".  (anderson@xxxxxxxxxx)

-  Fix for "kmem [address]" failure in 2.6.17 kernels (possibly earlier)
   that would fail with the error message: "kmem: invalid structure
   member offset: zone_zone_mem_map".  (anderson@xxxxxxxxxx)

-  Fix for "kmem -i" that resulted in a bogus "CACHED" page count
   value.  (anderson@xxxxxxxxxx)

-  As an result of the last "kmem -i" fix, I've added a new "kmem -V"
   option that dumps the kernel's new vm_stat[] array contents by
   their enum values:

    crash> kmem -V
           NR_ANON_PAGES: 38656
          NR_FILE_MAPPED: 3116
           NR_FILE_PAGES: 141106
                 NR_SLAB: 58605
            NR_PAGETABLE: 1059
           NR_FILE_DIRTY: 7
            NR_WRITEBACK: 0
         NR_UNSTABLE_NFS: 0
               NR_BOUNCE: 0
                NUMA_HIT: 86475467
               NUMA_MISS: 0
            NUMA_FOREIGN: 0
     NUMA_INTERLEAVE_HIT: 31523
              NUMA_LOCAL: 86475467
              NUMA_OTHER: 0
    crash>

  Interally, a new dump_vm_stat() function has been added to access
  any of the items in the list.  (anderson@xxxxxxxxxx)

- Implemented support for relocatable x86_64 live kernels and kdump
  generated vmcores.  Without this patch, attempts to analyze those
  kernels would fail during initialization with the error message:
  "crash: vmlinux and vmcore do not match!"  (anderson@xxxxxxxxxx)

- Support for recognizing real-time signals in the "sig" command.
  (olivier.daudel@xxxxxxxxxxxx)

- Fix for "sys -c" display of "sys_ni_syscall" entries that showed
  different system call names that have the same (W) symbol value
  as the (T) symbol "sys_ni_syscall".  For example:

   crash> sym -l | grep ffffffff802a38b6
   ffffffff802a38b6 (W) compat_sys_ipc
   ffffffff802a38b6 (W) compat_sys_keyctl
   ffffffff802a38b6 (W) compat_sys_sysctl
   ffffffff802a38b6 (W) ppc_rtas
   ffffffff802a38b6 (T) sys_ni_syscall
   ffffffff802a38b6 (W) sys_pciconfig_iobase
   ffffffff802a38b6 (W) sys_pciconfig_read
   ffffffff802a38b6 (W) sys_pciconfig_write
   ffffffff802a38b6 (W) sys_spu_create
   ffffffff802a38b6 (W) sys_spu_run
   ffffffff802a38b6 (W) sys_vm86
   ffffffff802a38b6 (W) sys_vm86old
   crash>

- Depending upon the kernel, one of those symbols would be displayed
  instead of sys_ni_syscall.  (olivier.daudel@xxxxxxxxxxxx)

- Fix for "sig" command where in later 2.6 kernels, the queued signal
  list at the end of the display would loop back on itself, repeatedly
  displaying the same queued signal(s).  (olivier.daudel@xxxxxxxxxxxx)
 

Download from:  http://people.redhat.com/anderson
 
 

--
Crash-utility mailing list
Crash-utility@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/crash-utility

[Index of Archives]     [Fedora Development]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [KDE Users]     [Fedora Tools]

 

Powered by Linux