Re: [lc-devel] [RFC] LZO de/compression support - take 6

[Nitin G. <nit...@gm...>]

On 5/30/07, Adrian Bunk <bu...@st...> wrote:
> On Tue, May 29, 2007 at 11:10:05PM +0200, Jan Engelhardt wrote:
> >
> > On May 28 2007 19:11, Adrian Bunk wrote:
> You completely miss the point of my question.
>
> It's about the performance improvements of the modified code that were
> mentioned.
>
> What you are talking about shouldn't have any effect on the generated
> code.
>

After Daniel refined this testing program, we can see that perf gain
is < 2% which can surely be accounted to cleanups like unnecessary
castings in tight loops.

Again, all the original code has been retained _as-is_. Whatever was
changed, has been mentioned in that detailed changelog that I post
along with patch.

If someone want to review these 500 lines with this changelog in hand,
it should not take more than couple of hours. If no-one can see any
problem in code by now and considering that it's tested on x86(_32),
amd64, ppc and giving somewhat better perf. than original then I
believe it is unnecessarily hanging outside of -mm tree.

I also contacted author (Markus Oberhumer) regarding above changes but
he seems not be responding right now. But still, if it gets into -mm
and gets used by various related projects then it should be good
enough for mainline also.


Cheers,
Nitin