Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

EDIT: AS POINTED BY SOME COMMUNITY MEMBERS, THIS HACK HAS NO EFFECT IN N900.

Hi,

I packaged the quick hack mentined in this page: http://www.webupd8.org/2010/11/alter...nel-patch.html

It's working in my N900 the same in my desktop and *should* increase perceived performance in the device but, of course, that's not guaranteed. But as kernel developers have themselves said they used this hack to test their "200 line patch", I think the gain is real.
To install, just install the attached deb AND REBOOT.

TRADITIONAL DEVEL WARNING APPLIES: USE AT YOUR OWN RISK, MAY BRICK YOUR DEVICE. NOT INTENDED TO REGULAR USERS.

Post here if you think it makes any difference.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Forgot to mention it's necessary to reboot after install.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

All the programs you run on N900 has probably the same TTY, so you will not see any improvements with this.

This patch all it does is give different priorities to the processes depending on their attached TTY. It helps when you are compiling a kernel or doing an intensive task on the command line or in an XTerm (which uses another TTY), but I hardly doubt anyone is doing this on the N900.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

In which processes we should notice a difference after installing this?

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Yes, someone said that in another thread but in my computer it ran exactly the same so I'm not sure if there's really no effect or if people are saying things they don't know about. I'd suggest first installing and then taking conclusions.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

It should smoother in general, but I'm not sure, waiting for feedback

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Take into account the X server nowadays, doesn't run on a TTY anymore. Just check with "ps waux" and you will see a '?' on the TTY field meaning they haven't any attached.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

I'm not sayinh you're wrong, Ignacius, you really seem to know what you're talking about. I'm not skilled enought to understand/discuss this, just using an empiric aproach. It take less time to install it and refute than to discuss.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Could you tell us what improvements did you experience and in which use cases?

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

I installed your package and see no immediate change in responsiveness. My guess is that it will increase responsiveness in high load situations when I have multiple apps running and the CPU is on high load.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

What can be the perceived performance when using applications that send dbus messages or execute shell commands? These should be using different TTYs, right?
And what about daemons running in background?

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Ignacius is absolutely correct. This wont make any difference in typical point and click usage for exactly the reason given. Whilst an empiric approach is cool simply because it's a good way to learn, there are times where it is indeed better to discuss than to try. Although I'm just a stranger from the internet, I've been smacking my head against this whole Unix thing for near on 2 decades, I do feel this is one of those times that it's better to understand why it wont make a difference than to try and see if it does.

: )

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Thanks for this approach, but it won't give us any boost.

If someone in here could just point me out the links to the maemo kernel compilation guide (got a shitty mouse in herer) I'm gonna compile the newest power kernel and the standard kernel with the mighty quickness patch and pubish it in here, if there's some interest.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

You have my interest. :)

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Also I don't understand this patch it is working great and when in x-terminal!

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Do it ,we r waiting.

:)

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

Yes, I tryed it for a while and it really makes no difference.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

What you describe in this thread is already used by Maemo.

Have a look at /usr/share/policy/etc/current/syspart.conf, which is a configuration file (most probably) for the ohmd daemon, which seems to do what you already want.

cgroups are mounted under /cgroups, so go there to see what's going on.

Re: Package: Alternative To The "200 Lines Kernel Patch That Does Wonders" Which You Can Use Right Away
 

why not use this ? http://talk.maemo.org/showthread.php?t=73315