Jump to content
6414 posts in this topic

Recommended Posts

I've revised the kernel some more.

Thanks for the tip Bronzovka, it seems it was assertion related (assert which translates into Assert function).

This made me thinking...

It can be that some things panic due to some assertion (code signature for example).

I've disabled some panics...

I also added rtc_clock_stepping and rtc_clock_stepped algorithms (for use in some PM kexts).

To everybody, please test this kernel and let me know if it still panics at things like iPhoto '08 or Preview...

EDIT: Oh yeah, I also disabled restart on panic... thought that would be useful...

 

It read all kexts and then screen gone blank, halted (With Chameleon)

 

i tried a few times and just couldnt install Clover......i would like to help if someone could advise how to install

didnt panic

 

Preview just froze for 30 seconds  and opened up after normal 

 

http://cl.ly/153n392y0g3M

 

and the Red Bug in Applauncher isnt there  this time 

 

and the Output from that GPU Check Log i guess u implemented

 

http://pastebin.com/50bdA5yu

  • Like 1

I've revised the kernel some more.

Thanks for the tip Bronzovka, it seems it was assertion related (assert which translates into Assert function).

This made me thinking...

It can be that some things panic due to some assertion (code signature for example).

I've disabled some panics...

I also added rtc_clock_stepping and rtc_clock_stepped algorithms (for use in some PM kexts).

To everybody, please test this kernel and let me know if it still panics at things like iPhoto '08 or Preview...

EDIT: Oh yeah, I also disabled restart on panic... thought that would be useful...

works on Phenom 2 x6 but this to my asus ati 5770 creates a bit 'of artifacts
but also serves as a good all iLife
 
with Test5 it works for me and all my asus ati 5770 does not create artifacts

Schermata 2013-11-01 alle 16.39.03.png

works on FX6100 too

the booting is more faster

the artefact on login screen is undeniable :(

there's an artefact on preview

 

znn0.th.png

 

==============

 

why i cant download anything from appstore

everytime inserting the apple id, ended with "unknown error ocurred"

is someone know how to fixed it ?? :( :( :(

works on FX6100 too

the booting is more faster

the artefact on login screen is undeniable :(

there's an artefact on preview

 

znn0.th.png

 

==============

 

why i cant download anything from appstore

everytime inserting the apple id, ended with "unknown error ocurred"

is someone know how to fixed it ?? :( :( :(

Wow... that's horrible...

Has anybody else got this issue?

Andy , panic at "rtclock_init" in commpage.c line 441 )))).

Oh that...

It's nothing to worry about...

kTscPanicOn needs to be set in order for this panic to occur...

Oh... I see... problem is tsc.c

One sec, I'll mod it...

Wow... that's horrible...

Has anybody else got this issue?

Oh that...

It's nothing to worry about...

kTscPanicOn needs to be set in order for this panic to occur...

Oh... I see... problem is tsc.c

One sec, I'll mod it...

You see "if (kTscPanicOn)

         panic("rtclock_init panic");"

 

It is maybe for 32 bit . No need it is =))) . i think )

Ok, building new kernel now...

It now by default should also get the frequency from EFI (boot loader).

If no FSB is found, default new frequency is 200 Mhz (as needed on FX and so)...

Hold on, post soon ;)

Ok, building new kernel now...

It now by default should also get the frequency from EFI (boot loader).

If no FSB is found, default new frequency is 200 Mhz (as needed on FX and so)...

Hold on, post soon ;)

Yes right ! ;)

Please try new kernel for 10.9 on old cpus and FX ! =))

DELETED , i continue fix kernel for reboot ...

If no FSB is found, default new frequency is 200 Mhz 

 

 

wasnt that corrected long time ago in chameleon so it wouldnt have to be done kernelside anymore

 

http://forge.voodooprojects.org/p/chameleon/issues/92/#ic826

 

Comment 15 by Evan Lojewski, Jul 20, 2011

Andy, that's because the legacy kernel multiplies the FSB by two for

K10s. It was a hack added to work around the bug in chameleon. Now

that it's fixed, it should be removed from xnu.

 

just throwing my thoughts in 

 

Yes right !  ;)

 

Please try new kernel for 10.9 on old cpus and FX ! =))

 

Paniced ! here http://cl.ly/image/3D1F431b321H

Ok, built the new kernel...

I also corrected some old stuff in tsc.c...

Removed the panic in commpage.c as shown by Bronzovka...

FX should now work too (with Chameleon, but this needs testing)...

Test please and report back ;)

mach_kernel.zip

SinetekBronzovkAnVoodoo_RC3.diff.zip

@andy

 

for code signing its much easier than that, only need to do an "if (0 &&" in ubc_subr and the SHA256's will always equate :-)

 

 

for the graphics corruption problem, i'm pretty sure the TSC is not to blame now.

issues occur on ATI too. my guess is that the nvidia driver shuts everything down as soon as it detects some sort of corruption (visibly only memory seems affected in some sort, the gpu or driver never seems to hard-crash) whereas the ATI driver for some reason does not.

 

perhaps there is a way to disable this failure mechanism and go on. However, things seem  horribly messed up graphically.

 

another lead is the npci=~~ flag, maybe this causes the PCI bus to become confused and update the wrong portions of memory or something.

i've tried an older IOPCIPlatform kext, but no dice.

 

- Sinetek

  • Like 3

wasnt that corrected long time ago in chameleon so it wouldnt have to be done kernelside anymore

 

http://forge.voodooprojects.org/p/chameleon/issues/92/#ic826

just throwing my thoughts in 

 

Paniced !

where picture ?

@andy

 

for code signing its much easier than that, only need to do an "if (0 &&" in ubc_subr and the SHA256's will always equate :-)

Where precisely?

Could you show what portion of ubc_subr.c you mean?

Thanks for the info :)

Athlon 64 x2 5000+ - K9N Ultra

 

Boot with kernel of the post #3931 but more graphic corruption than before. I go to test the latest Andy's kernel now...

 

Instant reboot after kext listing with the kernel of the post #3942.

 

Can the output of an iOS App using OpenGL framework crash in simulator can help in this research ?

line 3310

 

 

 

anyone else noticed mavericks is slow as molasses during shutdown?

I already did something similar in the past...

Unfortunetely it won't fully beat the CS enforcement...

If you load an app with something modified, it will still crash the kernel...

The only way to do app modding is by stripping the code signature with my tool (stripcodesig).

unsigned long cs_validate_page_no_hash = 0;
unsigned long cs_validate_page_bad_hash = 0;
boolean_t
cs_validate_page(
	__unused void			*_blobs,
	__unused memory_object_t		pager,
	__unused memory_object_offset_t	page_offset,
	__unused const void		*data,
	boolean_t		*tainted)
{
	*tainted = FALSE;
	
	return TRUE;
}

Ok, built the new kernel...

I also corrected some old stuff in tsc.c...

Removed the panic in commpage.c as shown by Bronzovka...

FX should now work too (with Chameleon, but this needs testing)...

Test please and report back ;)

with chameleon blackscreen after loading kext :(

with clover success boot to desktop

and fsb now detected 200Mhz with CPU-X

Kernel #3944 boots to the desktop. Same graphics glitches (ATI 5450).

 

Athlong 64 x2 5000+ on K9N Ultra

 

For information, for my latest issue with clock going backward 40min each second, Balking N300 dongle was involved... unplugged it and it solved the issue.

i might be totally wrong here, but maybe its worth a shot for the Graphics Corruption,

 

when booting in verbose mode it logs to /dev/console now if i try a symlink to /dev/console.txt couldnt we see what exactly happens when we dont see anything anymore on the screen with nvidia after booting with a ATI card? or with -x ?

maybe wrong sysenter in opemu )))

Should work.

32-bit apps load fine...

It has both 32 and 64 bit sysenter and detects this just fine based on the arch of the binary it is being fed...

Should work.

32-bit apps load fine...

It has both 32 and 64 bit sysenter and detects this just fine based on the arch of the binary it is being fed...

Andy , yes work fine , i tested !!! 

But why other programm is code signing in crash ? I see in code signing - maybe kevent64 - it is syscall is no compatible with sysenter ? )))

×
×
  • Create New...