Re: [PATCH v2 00/29] first version of mcdstub

[Alex Bennée]

Nicolas Eder <nicolas.eder@lauterbach.com> writes:

<snip>
> Signed-off-by: Nicolas Eder <nicolas.eder@lauterbach.com>
>
> neder (29):

I think you need to fix your author attribution here.

>   mcdstub initial commit, mcdstub file structure added
>   TCP chardev added, handshake with TRACE32 working
>   TCP packet handling added
>   queries for resets and triggers added
>   queries for memory spaces and register groups added
>   query for registers added
>   query data preparation improved
>   shared header file added, used for TCP packet data
>   memory and register query data now stored per core
>   handler for resets added
>   query for the VM state added
>   handler for reading registers added
>   handler for reading memory added
>   handler for single step added
>   adapting to the qemu coding style
>   deleting the mcdd startup option
>   handler for breakpoints and watchpoints added
>   making step and go handlers core-specific
>   adding trigger ID handling for TRACE32
>   cp register read/write added
>   switching between secure and non-secure memory added
>   transitioning to unsinged integers in TCP packets and removing
>     MCD-API-specific terms
>   moved all ARM code to the ARM mcdstub and added now commom header file
>   step and go handlers now propperly perform global operations
>   Doxygen documentation added
>   moved all mcd related header files into include/mcdstub
>   MCD stub entry added to maintainers file
>   added description to out-commented gdb function
>   introducing the DebugClass. It is used to abstract the gdb/mcd
>     set_stop_cpu function.

As you need to re-base anyway for this to apply cleanly I'm going to
wait until v3 for another pass. However I have noticed these patches are
quite noisy with a number of issues:

  - commented out code
  - code introduced then deleted
  - code motion after introduction
  - random white space changes

All of which makes it hard to review. A lot of this stems from the c&p
scaffolding from gdbstub which I understand as an approach to write the
initial version. However this should be squashed and merged away in the
final patches presented for review. Also please make sure:

  - commit messages match changes
  - each patch compiles cleanly on its own
  - you run through checkpatch.pl

Thanks,

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro