Wine performance improvement

[Wolfram Sang]

Stefan Dösinger wrote:
> Am 09.12.2009 um 09:36 schrieb 임은지:
>> Thank you for your comments. I will cleanup the patch as you
>> suggested and send it.
> Did you contact the kernel maintainers regarding the kernel side of
> your changes? I think including the Wine patch doesn't make sense
> until the kernel developers accept the kernel side of it.

In the current form, this is unlikely. There are a number of formal 
issues alone:

- needs to be based on a somewhat recent version of the kernel
- too much #ifdef
- just one big patch
- too much debug output
- too much code commented out
- commenting not consistent (//)

The latter three can be easily worked out, of course. The others are a 
bit more troublesome, but are major show-stoppers nevertheless.

Also, adding syscalls is a very, very sensitive area. You really need a 
gallore of good arguments for that. IMHO I have doubts here, because the 
rest of the code looks also a bit intrusive to me (i.e. patching 
fork|sched.c).

Then again, was it really the aim to send this upstream? I didn't get 
the impression when I read the original posting, more like "We did 
something interesting. Maybe you want to try it?"

Kind regards,

    Wolfram