InfiniteMac OSx86

InfiniteMac OSx86 (http://infinitemac.com/forum.php)
-   Leopard 10.5 (http://infinitemac.com/forumdisplay.php?f=85)
-   -   [Solved] AMD Dual Core Timing Issue AKA Mouse Bug (http://infinitemac.com/showthread.php?t=734)

Puttabong 10-11-2008 03:50 PM

Am I reading right? The 'Mouse movement bug' has finally been put to an end? Wonderful!

Thank you XNU Dev Labs. :)

R0GUE 10-11-2008 04:06 PM

Yes, you read it right. I've been testing a pre beta 2 kernel with the relevant fix and I have to say that my system has never been so stable. No mDNSResponder errors, no divide by zero, no -10810 errors -- nothing!

Good job XNU-Dev!

Taisto 10-11-2008 05:57 PM

it`s not yet released right? is it possible to get this beta ? and what is the release date cause 8th was delayed

R0GUE 10-11-2008 06:40 PM

According to mercurysquad the official beta 2 should be released before Monday, "hopefully".

naquaada 10-11-2008 06:41 PM

Does the 9.5 kernel work with 10.5.3 too? I don't have the mouse mevement bug, but sometimes dividion by zero and maybe two times the -10810 error since I own my dual-core cpu which is since May. Which packages do I need?

Taisto 10-11-2008 08:09 PM

i think it may work, at least it works with 10.5.4. anyway i wonder why some bugs appear on some cpus but do not show up on others...i`ve never seen division by zero problem or any errors. i`ve got only mouse-bug and sound stuttering

R0GUE 10-11-2008 10:24 PM

The 'mouse bug' and audio stutter are both symptoms of TSC drift which has now been resolved.

nfoav8or 10-11-2008 10:44 PM

Not to throw this into the wrong discussion but with the voodoo_beta kernel, my original TSC delta was -235690 and the new delta with CPU1 is 36.

Is there any way to narrow the gap down closer to 0?

Voyn1x 10-12-2008 12:40 PM

Quote:

Originally Posted by mercurysquad (Post 16861)
This is now fixed as of Voodoo beta 2 :)

One question though: would people prefer to have it as a kext or in-kernel as it is now?

This is fantastic news, huge thanks to the XNU team for solving this at last! :cool:

Is there any benefit to having this fix as a kext? Would it not be simpler if it stays in-kernel?

mercurysquad 10-12-2008 07:40 PM

R0GUE just identified a bug where sleep/resume resets the workaround. So we'll probably package this up as a .kext which AMD users can dump in their Extensions folder and it will take care of everything - rather than modifying the kernel's sleep/resume functions. People who use other kernels can use the kext too.

This also means beta 2 will be further delayed. Please be patient :)