Re: [Qemu-devel] [PATCH] m25p80.c Added support for N25Q256 and N25Q512

[Peter Crosthwaite]

On Mon, Nov 30, 2015 at 1:33 PM, Krzeminski, Marcin (Nokia -
PL/Wroclaw) <address@hidden> wrote:
>
>
>> -----Original Message-----
>> From: EXT Peter Crosthwaite [mailto:address@hidden
>> Sent: Monday, November 30, 2015 9:55 PM
>> To: Krzeminski, Marcin (Nokia - PL/Wroclaw)
>> Cc: address@hidden; address@hidden; Sai Pavan Boddu
>> Subject: Re: [Qemu-devel] [PATCH] m25p80.c Added support for N25Q256
>> and N25Q512
>>
>> On Mon, Nov 30, 2015 at 11:51 AM, Krzeminski, Marcin (Nokia -
>> PL/Wroclaw) <address@hidden> wrote:
>> >
>> >
>> >> -----Original Message-----
>> >> From: EXT Peter Crosthwaite [mailto:address@hidden
>> >> Sent: Sunday, November 29, 2015 8:19 PM
>> >> To: Krzeminski, Marcin (Nokia - PL/Wroclaw)
>> >> Cc: address@hidden; address@hidden; Sai Pavan Boddu
>> >> Subject: Re: [Qemu-devel] [PATCH] m25p80.c Added support for
>> N25Q256
>> >> and N25Q512
>> >>
>> >> On Sun, Nov 29, 2015 at 5:12 AM, Krzeminski, Marcin (Nokia -
>> >> PL/Wroclaw) <address@hidden> wrote:
>> >> >
>> >> >
>> >> >> -----Original Message-----
>> >> >> From: EXT Peter Crosthwaite [mailto:address@hidden
>> >> >> Sent: Saturday, November 28, 2015 7:50 PM
>> >> >> To: Krzeminski, Marcin (Nokia - PL/Wroclaw)
>> >> >> Cc: address@hidden; address@hidden; Sai Pavan Boddu
>> >> >> Subject: Re: [Qemu-devel] [PATCH] m25p80.c Added support for
>> >> N25Q256
>> >> >> and N25Q512
>> >> >>
>> >> >> These features are also available in Xilinx QEMU if you want to
>> >> >> compare
>> >> >> implementation:
>> >> >>
>> >> >>
>> >>
>> https://github.com/Xilinx/qemu/blob/pub/2015.2.plnx/hw/block/m25p80.c
>> >> >>
>> >> >> That work also handles the larger and newer Spansion flash parts,
>> >> >> as well as the quad and dual mode commands for QSPI (also features
>> >> >> of
>> >> n25qXXX).
>> >> >>
>> >> > Too bad I did not checked xilix fork, so V2 of this patch does not
>> >> > make
>> >> sense since all is already implemented.
>> >>
>> >> V2 still makes sense. My V1 comments are pretty minor and that Xilinx
>> >> work will need cleanup before upstreaming. It is not a competing
>> >> implementaiton and the diff there will go down when it is rebased on
>> yours.
>> >>
>> > Since this is start with open source, making this feature at least
>> > ready to pull seem to be worth to try. I'll prepare v2 when I got some free
>> time.
>>
>> There is a learning curve for some people on getting into patch respinning.
>> You need some features of git that let you rewrite history. Specifically you
>> want to look at:
>>
>> git rebase -i
>> git add -p
>> git commit --amend
>>
>> If you are unfamiliar.
>>
> For me it is much more complicated. I added support for those flash devices 
> in one commit,
> then made one or two fixes and that's it. It is really hard to get patch set 
> from that..
> The way of working is the main issue here...
>

git reset HEAD^ then git add -p. It lets you interactively select
hunks to commit. Then when you git commit only the ones you selected
are committed. Use "e" mode if you have a single hunk with multiple
changes. Keep adding and commiting until you have no diff.

Regards,
Peter

>> >> > Why didn't xilinks merge it with mainline?
>> >>
>> >> There's a lot in that tree and Xilinx hasn't gotten around to it yet.
>> >>
>> > Yes, I noticed one interesting feature.
>>
>> Are you able to share more? :) If there is known community interest in
>> specific features it helps upstreaming.
>>
> Yes, most interesting thing is the I2C eeproms (m24cxx.c file). There is 
> smbus eeprom,
> but this is not exactly I2C. At least from my point of view it is interesting 
> to have it in upstream :)
>
> Regards,
> Marcin
>
>> Regards,
>> Peter