Re: [GIT PULL v4.6] MDB Linux Kernel Debugger x86/x86_64

[Jeffrey Merkey]

On 3/25/16, Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote:
> Hi Jeff,
>
> On Fri, 25 Mar 2016 17:01:27 -0600 Jeffrey Merkey <jeffmerkey@xxxxxxxxx>
> wrote:
>>
>> I went back and checked the code and as it turns out, none of the
>> patches you nak'd are in the current branch, there are different
>> patches there now.   There are two patches you ignored that are in it,
>> but no record of a Nak for either of them.
>
> OK, so one obvious problem is that the tree you want merged has a
> single patch in it and the diffstat looks like this:
>
>  Documentation/sysrq.txt                    |    2 +-
>  MAINTAINERS                                |    6 +
>  arch/x86/include/asm/bug.h                 |    9 +-
>  arch/x86/include/uapi/asm/debugreg.h       |    1 +
>  arch/x86/kernel/Makefile                   |    1 +
>  arch/x86/kernel/apic/io_apic.c             |    2 +
>  arch/x86/kernel/debug/Makefile             |    3 +
>  arch/x86/kernel/debug/mdb/Makefile         |    6 +
>  arch/x86/kernel/debug/mdb/Makefile.local   |  106 +
>  arch/x86/kernel/debug/mdb/mdb-base.c       | 3293 +++++++++++++
>  arch/x86/kernel/debug/mdb/mdb-base.h       |  447 ++
>  arch/x86/kernel/debug/mdb/mdb-ia-apic.c    |  243 +
>  arch/x86/kernel/debug/mdb/mdb-ia-proc.h    |  819 ++++
>  arch/x86/kernel/debug/mdb/mdb-ia-support.c | 5342 +++++++++++++++++++++
>  arch/x86/kernel/debug/mdb/mdb-ia-support.h |   76 +
>  arch/x86/kernel/debug/mdb/mdb-ia.c         | 6887
> ++++++++++++++++++++++++++++
>  arch/x86/kernel/debug/mdb/mdb-ia.h         |  209 +
>  arch/x86/kernel/debug/mdb/mdb-keyboard.h   |  127 +
>  arch/x86/kernel/debug/mdb/mdb-list.c       |  534 +++
>  arch/x86/kernel/debug/mdb/mdb-list.h       |   96 +
>  arch/x86/kernel/debug/mdb/mdb-logic.c      | 2118 +++++++++
>  arch/x86/kernel/debug/mdb/mdb-main.c       |  786 ++++
>  arch/x86/kernel/debug/mdb/mdb-os.c         | 1474 ++++++
>  arch/x86/kernel/debug/mdb/mdb-os.h         |  141 +
>  arch/x86/kernel/debug/mdb/mdb-proc.h       |  179 +
>  arch/x86/kernel/debug/mdb/mdb.h            |   40 +
>  arch/x86/kernel/dumpstack_32.c             |    6 +-
>  arch/x86/kernel/dumpstack_64.c             |   18 +
>  arch/x86/kernel/hw_breakpoint.c            |   69 +-
>  arch/x86/kernel/reboot.c                   |    1 +
>  arch/x86/kernel/traps.c                    |   14 +
>  drivers/tty/vt/vt.c                        |    4 +
>  include/asm-generic/bug.h                  |    4 +
>  include/linux/console.h                    |    4 +
>  kernel/debug/kdb/kdb_debugger.c            |    2 +-
>  kernel/events/hw_breakpoint.c              |    2 +
>  kernel/extable.c                           |    1 +
>  kernel/kallsyms.c                          |   45 +
>  kernel/module.c                            |   43 +
>  kernel/rcu/tree.c                          |    1 +
>  kernel/sched/core.c                        |   13 +-
>  kernel/time/clocksource.c                  |    1 +
>  kernel/watchdog.c                          |   17 +-
>  lib/Kconfig.debug                          |   66 +
>  44 files changed, 23240 insertions(+), 18 deletions(-)
>
> Having it in that form makes it pretty hard for anyone to review ...
>
> If you have a set if incremental patches, maybe you should post those
> instead.
>
> I am afraid that if the maintainers affected by this code will not
> merge it, I cannot justify putting it in linux-next myself.
> --
> Cheers,
> Stephen Rothwell
>

I am not going to waste time arguing with the maintainers and chasing
after them.  I have a debugger, it has lots of users and a big
installed base that goes back many years.  So long as they can get it
and use it, things are cool.  Putting it in Linux is actually more of
a pain than just doing what I have been over the years and maintaining
it alongside linux.

Each merge cycle it will be submitted for the linux community, but
like I said, I am not wasting time with folks who are not my customers
and or really not that interested in it.

I already played the bait and switch game with Ingo once before and I
did a lot a work and it was ignored, so he can waste someone else's
time.  Linux is only as good as the people maintaining it.

Jeff