ISST-LTE:pVM:roselp4:ubuntu 16.04.2: vmcore cannot be analysed by crash

[SRU justification]
This fix is required to make the crash tool usable. It does also improve makedumpfile filtering of pages.

[Impact]
Kernel crashes cannot be analysed with the crash tool.
makedumpfile incorrectly filter pages.

[Fix]
Cherry-pick upstream commits fixing those issues.

[Test Case]
Running crash tool on a kernel crash file will display something like :

# crash -s usr/lib/debug/boot/vmlinux-4.8.0-34-generic
crash: read error: kernel virtual address: ffffffff81e29ff0 type: "pv_init_ops"
crash: this kernel may be configured with CONFIG_STRICT_DEVMEM, which
       renders /dev/mem unusable as a live memory source.
crash: trying /proc/kcore as an alternative to /dev/mem

crash: seek error: kernel virtual address: ffffffff81e29ff0 type: "pv_init_ops"
crash: seek error: kernel virtual address: ffffffff82166130 type: "shadow_timekeeper xtime_sec"
crash: seek error: kernel virtual address: ffffffff81e0d304 type: "init_uts_ns"
crash: usr/lib/debug/boot/vmlinux-4.8.0-34-generic and /var/crash/201701191308/dump.201701191308 do not match!

With the fix, the crash command will work as expected

Running the crash tool on a vmcore file produced by makedumpfile may return :

crash: page excluded: kernel virtual address: <> type: "fill_task_struct"

[Regression]
None expected as those modifications are part of the Zesty and upstream version.

The makedumpfile patches are in Yakkety and Zesty 1.6.0 & after

[Original description of the problem]
vmcore captured by kdump cannot be opened with crash:

% sudo crash -d1 /usr/lib/debug/boot/vmlinux-4.8.0-34-generic /var/crash/201612282137/dump.201612282137
... ...
base kernel version: 0.8.0
linux_banner:
????????
crash: /usr/lib/debug/boot/vmlinux-4 and /var/crash/201612282137/dump.201612282137 do not match!

Usage:

  crash [OPTION]... NAMELIST MEMORY-IMAGE[@ADDRESS] (dumpfile form)
  crash [OPTION]... [NAMELIST] (live system form)

Enter "crash -h" for details.

Looks like the 'linux_banner' cannot be understood by crash.

And when the vmcore was dumping, this message being showed:

[ 729.609196] kdump-tools[5192]: The kernel version is not supported.
[ 729.609447] kdump-tools[5192]: The makedumpfile operation may be incomplete.
---uname output---
Linux roselp4 4.8.0-34-generic #36~16.04.1-Ubuntu SMP Wed Dec 21 18:53:20 UTC 2016 ppc64le ppc64le ppc64le GNU/Linux

Machine Type = lpar

---Debugger---
A debugger is not configured

---Steps to Reproduce---
 1. config kdump
2. trigger kdump
3. analyse vmcore with crash

Userspace tool common name: crash/makedumpfile

The userspace tool has the following bit modes: 64-bit

Userspace rpm: makedumpfile 1.5.9-5ubuntu0.3/crash 7.1.4-1ubuntu4

Userspace tool obtained from project website: na

*Additional Instructions for Ping Tian <email address hidden>:
-Post a private note with access information to the machine that the bug is occuring on.
-Attach ltrace and strace of userspace application.

xtime timespec.tv_sec: 586481e8: Wed Dec 28 21:24:24 2016
utsname:
     sysname: Linux
    nodename: boblp1
     release: 4.8.0-32-generic
     version: #34~16.04.1-Ubuntu SMP Tue Dec 13 17:01:57 UTC 2016
     machine: ppc64le
  domainname: (none)
base kernel version: 4.8.0
verify_namelist:
dumpfile /proc/version:
Linux version 4.8.0-32-generic (buildd@bos01-ppc64el-001) (gcc version 5.4.0 20160609 (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.4) ) #34~16.04.1-Ubuntu SMP Tue Dec 13 17:01:57 UTC 2016 (Ubuntu 4.8.0-32.34~16.04.1-generic 4.8.11)
/usr/lib/debug/boot/vmlinux-4.8.0-32-generic:
Linux version 4.8.0-32-generic (buildd@bos01-ppc64el-001) (gcc version 5.4.0 20160609 (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.4) ) #34~16.04.1-Ubuntu SMP Tue Dec 13 17:01:57 UTC 2016 (Ubuntu 4.8.0-32.34~16.04.1-generic 4.8.11)

hypervisor: (undetermined)
crash: per_cpu_symbol_search(per_cpu__tvec_bases): NULL
ppc64_vmemmap_init: vmemmap base: f000000000000000

crash: PPC64: cannot find 'cpu_possible_map', 'cpu_present_map', 'cpu_online_map' or 'cpu_active_map' symbols

root@boblp1:/usr/lib/debug/boot# uname -a
Linux boblp1 4.8.0-32-generic #34~16.04.1-Ubuntu SMP Tue Dec 13 17:01:57 UTC 2016 ppc64le ppc64le ppc64le GNU/Linux
root@boblp1:/usr/lib/debug/boot#

1. Missing v4.8 support related patches in crash tool

   commit 098cdab16dfa6a85e9dad2cad604dee14ee15f66
   Author: Dave Anderson <email address hidden>
   Date: Fri Feb 12 14:32:53 2016 -0500

    Fix for the changes made to the kernel module structure introduced by
    this kernel commit for Linux 4.5 and later kernels:

      commit 8244062ef1e54502ef55f54cced659913f244c3e
      modules: fix longstanding /proc/kallsyms vs module insertion race.

    Without the patch, the crash session fails during initialization
    with the error message: "crash: invalid structure member offset:
    module_num_symtab".
    (<email address hidden>)

   commit 6f1f78e33474d00d5f261d7ed9d835c558b34d61
   Author: Dave Anderson <email address hidden>
   Date: Wed Jan 20 09:56:36 2016 -0500

    Fix for the changes made to the kernel module structure introduced by
    this kernel commit for Linux 4.5 and later kernels:

      commit 7523e4dc5057e157212b4741abd6256e03404cf1
      module: use a structure to encapsulate layout.

    Without the patch, the crash session fails during initialization
    with the error message: "crash: invalid structure member offset:
    module_init_text_size".
    (<email address hidden>)

   commit 1e92f9fad3a7e3042b16996306cb2335760ef8c8
   Author: Dave Anderson <email address hidden>
   Date: Mon Feb 1 16:10:49 2016 -0500

    Fix for the replacements made to the kernel's cpu_possible_mask,
    cpu_online_mask, cpu_present_mask and cpu_active_mask symbols in
    this kernel commit for Linux 4.5 and later kernels:

      commit 5aec01b834fd6f8ca49d1aeede665b950d0c148e
      kernel/cpu.c: eliminate cpu_*_mask

    Without the patch, behavior is architecture-specific, dependent upon
    whether the cpu mask values are used to calculate the number of cpus.
    For example, ARM64 crash sessions fail during session initialization
    with the error message "crash: zero-size memory allocation! (called
    from <address>)", whereas X86_64 sessions come up normally, but
    cpu mask values of zero are stored internally.
    (<email address hidden>)

   commit 182914debbb9a2671ef644027fedd339aa9c80e0
   Author: Dave Anderson <email address hidden>
   Date: Fri Sep 23 09:09:15 2016 -0400

    With the introduction of radix MMU in Power ISA 3.0, there are
    changes in kernel page table management accommodating it. This patch
    series makes appropriate changes here to work for such kernels.
    Also, this series fixes a few bugs along the way:

      ppc64: fix vtop page translation for 4K pages
      ppc64: Use kernel terminology for each level in 4-level page table
      ppc64/book3s: address changes in kernel v4.5
      ppc64/book3s: address change in page flags for PowerISA v3.0
      ppc64: use physical addresses and unfold pud for 64K page size
      ppc64/book3s: support big endian Linux page tables

    The patches are needed for Linux v4.5 and later kernels on all
    ppc64 hardware.
    commit 8ceb1ac628bf6a0a7f0bbfff030ec93081bca4cd
   Author: Dave Anderson <email address hidden>
   Date: Mon May 23 11:23:01 2016 -0400

    Fix for Linux commit 0139aa7b7fa12ceef095d99dc36606a5b10ab83a, which
    renamed the page._count member to page._refcount. Without the patch,
    certain "kmem" commands fail with the "kmem: invalid structure member
    offset: page_count".
    (<email address hidden>)

   commit 7136bf8495948cb059e5595b8503f8ae37019fa1
   Author: Dave Anderson <email address hidden>
   Date: Thu May 19 14:01:19 2016 -0400

    Fix for Linux commit edf14cdbf9a0e5ab52698ca66d07a76ade0d5c46, which
    has appended a NULL entry as the final member of the pageflag_names[]
    array. Without the patch, a message that indicates "crash: failed to
    read pageflag_names entry" is displayed during session initialization
    in Linux 4.6 kernels.
    (<email address hidden>)

2. The following makedumpfile commits are needed:

   commit 5bc1f520cc7ab6e18abdd5af21c80ecda6339eb5
   Author: Atsushi Kumagai <email address hidden>
   Date: Tue Jan 26 10:11:33 2016 +0900

    [PATCH] Looking for page.compound_order/compound_dtor to exclude hugepages

    * Required for kernel 4.4

    Due to some changes in struct page, hugepages wouldn't be removed on
    linux 4.4. makedumpfile reads page.lru.prev to get "order" (number of hugepages)
    and page.lru.next to get "dtor" (destructor for hugepages) to detect hugepages,
    but the offsets of the two was changed in linux 4.4.

      kernel version | where is order | where is dtor
     ----------------+---------------------------+---------------------------
            - v3.19 | lru.prev | lru.next
       v4.0 - v4.3 | compound_order(=lru.prev) | compound_dtor(=lru.next)
       v4.4 - | compound_order | compound_dtor

    As above, OFFSET(page.compound_order) and OFFSET(page.compound_dtor) are
    definitely necessary in VMCOREINFO on linux 4.4 and later.

    Further, the content of page.compound_dtor was changed from direct address
    of dtor to the ID of it in linux 4.4.

    Signed-off-by: Atsushi Kumagai <email address hidden>

   commit 13b4233e91a9d5aa14c4b0643af36cbc29b9fa7a
   Author: Atsushi Kumagai <email address hidden>
   Date: Wed Feb 24 17:09:44 2016 +0900

    [PATCH] Skip examining compound tail pages

    * Required for kernel 4.5

    For filtering user pages, we check whether each page's
    page->mapping have PAGE_MAPPING_ANON bit.
    However, unexcludable compound tail pages can have
    PAGE_MAPPING_ANON since kernel 4.5, they can be excluded
    as user page wrong.

    Now, we don't need to check compound tail pages because
    excludable compound pages must be excluded at a time by
    exclude_range() when the corresponding head page is checked.
    So just skipping tail pages can avoid wrong filtering.

    Signed-off-by: Atsushi Kumagai <email address hidden>

3. The linux-image dbgsym version installed must be pulled from a different repo
   instead of the one meant for 16.04.2 because the gcc version of kernel
   image (/boot/vmlinux-4.8.0-34-generic) and the vmlinux with debug
   symbols(usr/lib/debug/boot/vmlinux-4.8.0-34-generic) don't match.

   Please use the following repos

     sudo tee /etc/apt/sources.list.d/ddebs.list << EOF
     deb http://ddebs.ubuntu.com/ $(lsb_release -cs) main restricted universe multiverse
     deb http://ddebs.ubuntu.com/ $(lsb_release -cs)-security main restricted universe multiverse
     deb http://ddebs.ubuntu.com/ $(lsb_release -cs)-updates main restricted universe multiverse
     deb http://ddebs.ubuntu.com/ $(lsb_release -cs)-proposed main restricted universe multiverse
     EOF

   to install linux-image-4.8.0-34-generic-dbgsym package.

Thanks

[snip]

>
> 3. The linux-image dbgsym version installed must be pulled from a different
> repo

s/must be pulled/must have been pulled/

Applied crash utility's missing patches on top of
crash-7.1.4-1ubuntu4 and makedumpfile tool's missing patches on top of
makedumpfile-1.5.9-5ubuntu0.3. Did some sanity testing of the
patched binaries. The binaries were working as expected.