Jump to content

[AMD] Yosemite Kernel Testing (for help use the Help Topic)


Duran Keeley
 Share

1,898 posts in this topic

Recommended Posts

 

hello spakk, 

 

le pentium64 n'est pas 3dnow et il rencontre les mêmes problèmes qu'Athlon64 et Phenom ;)

 

le problème est opemu ssse3 ;)

 

http://www.cpu-world.com/CPUs/Pentium_D/Intel-Pentium%20D%20915%20HH80553PG0724MN%20%28BX80553915%20-%20BX80553915R%29.html

 

hello spakk,
 
the pentium64 not 3dnow and meets the same qu'Athlon64 and Phenom problems)
 
the problem is opemu SSSE3 ;)

 

 

Hi Gils,

 

look here: http://en.wikipedia.org/wiki/Phenom_II

this information accordingly to: Thuban ( Callisto, Heka, Deneb and Zosma)
  • Like 1
Link to comment
Share on other sites

 

Hi Gils,

 

look here: http://en.wikipedia.org/wiki/Phenom_II

this information accordingly to: Thuban ( Callisto, Heka, Deneb and Zosma)

 

 

the pentium is not 3dnow 

 

the problem is not 3dnow ;) 

Link to comment
Share on other sites

New version with optimised OpenSSL decrypter and Bronzovka cpuid fix.

Test please, should work now fine :D

For use of core crypto decrypter use -nodecryptor in  combination with AppleDecrypt-3.1.0.

If used within kernel itself, always panic so only option is that way ;)

BSA_Yos_R2.zip

  • Like 7
Link to comment
Share on other sites

Hi AnV ! 
I tested your kernel R2 , get glitches on desktop . You should delete pseudo code in commpage.c :

case 4:
  bits |= kHasSSE2;
+
+ bits |= kHasSSE3; <-- need delete 
+ bits |= kHasSupplementalSSE3; <-- need delete 
+
+ cpuid_info()->cpuid_features |= CPUID_FEATURE_SSSE3; <-- need delete 
+ cpuid_info()->cpuid_features |= CPUID_FEATURE_SSE3; <-- need delete 
 
2) I did new rtclock.c and tsc.c and working fine on all amd  ., because i see in your diff and speak to you that this tsc.c - is old . 
3) i did new mtrr.c and fine :)
4) i did and modified instructions for ssse3 in opemu . no crash in 10.7.5 ;)
I need free time for create diff ... 
  • Like 6
Link to comment
Share on other sites

New version with optimised OpenSSL decrypter and Bronzovka cpuid fix.

Test please, should work now fine 

 

BSA_Yos_R2

kernel starts, Safari crashes, app store no function, blurred images,no improvement

 

Andy Bronya

 

 

Question:

 

 

I did test now some Kernel, up to improvement of App Store, safari.

With the App Store the images are not shown,safari crashes quite often.

Also with the employment of the WebKitFramework (no patch copy direct) did not bringsuccess.

 

 

Youtube films are not played.

 

Only a Kernel brought an improvement with me

I will upload just a Kernel now with that the App Store and will function safari with my hardware.

Unfortunately I cannot any longer remind who this Kernel released myself. Possibly Andy Bronya can use this Kernel as basis for further optimizations.

 

 

Approach:

 

 

Kernel copy, WebKit Framework --> all files in folder Frameworks copy, with Kext wizard update, again start, App Store start.

 

 

Edit:

 

Darwin Kernel Version 14.0.0

durankeeley:xnu-2782.1.97-opemu1/BUILD/obj/RELEASE_X86_64 x86_64

 

kernel.zip

post-1439641-0-22858300-1418022209_thumb.jpg

  • Like 1
Link to comment
Share on other sites

open terminal and give : uname -a 

then give us the names of kernel


the kernel "durankeeley:xnu-2782.1.97-opemu1/BUILD/obj/RELEASE_X86_64 x86_64" has been built based on "BSA_10.10_rc1". I'm just create a new on, but would only try it before I upload here.

  • Like 2
Link to comment
Share on other sites

 

You can do separately FX & Others AMD CPU ???

 
AMD FX processors only have SSSE3. Others are SSE3.
 
Because Athlon + SSSE3 = glitches
 
e.g. :
 
if CPU=AMD FX
{
   cpuid_info()->cpuid_features |= CPUID_FEATURE_SSSE3;
}
else
{
   cpuid_info()->cpuid_features |= CPUID_FEATURE_SSE3;
}
 
 
Thank you

 

No !

Opemu  wrong !

 

 

 

no works on Athlon 5350 

yes , this old tsc.c . I need give to AnV new files ...

omg , i tested in opemu function fetchoperands() :

if(mod == 2) {...} - this no effect and don't working ! Very odd from Sinetek ! :\

  • Like 7
Link to comment
Share on other sites

I can not test it at the moment,
I have the fuc_king Yosemite adjusted before building kernel and now my amd computer does not boot from yosemite ... CPU Halted, this has nothing to do with kernel test. I do not want upload untested kernel. sory

Link to comment
Share on other sites

hello bronya , 

 

il n'y a que ton kernel qui fonctionne sur mon kabini  :rolleyes:

tu pourrais regarder le problème IOPCI s'il te plait ?   :yes: 

 

there are only your kernel running on my Kabini  :rolleyes:
you could watch the IOPCI problem please?   :yes:
Link to comment
Share on other sites

Just for the sake or archiving info here:

 

The OPEMU used in BSA_rc2 that I uploaded is attached, it's an old gen 2 bronya optimised.

 

KNOWN PROBLEMS:

WebKit crash on vanilla Yosemite frame work, works on the new compiled WebKit (used in the WebKit Patch)

Launchpad - White icons

Screenshot - Full White

Web kit problem is with that copies the WebKit Framework files (not the patch) no problem. App Store functions and to safari does not crashes (see the Picture)

 However Bronja / Andys is more again faster kernel substantially

Link to comment
Share on other sites

kernel starts, Safari crashes, app store no function, blurred images,no improvement

 

Andy / Bronya

 

 

Question:

 

 

I did test now some Kernel, up to improvement of App Store, safari.

With the App Store the images are not shown,safari crashes quite often.

Also with the employment of the WebKitFramework (no patch copy direct) did not bringsuccess.

 

 

Youtube films are not played.

 

Only a Kernel brought an improvement with me

I will upload just a Kernel now with that the App Store and will function safari with my hardware.

Unfortunately I cannot any longer remind who this Kernel released myself. Possibly Andy / Bronya can use this Kernel as basis for further optimizations.

 

 

Approach:

 

 

Kernel copy, WebKit Framework --> all files in folder Frameworks copy, with Kext wizard update, again start, App Store start.

 

 

Edit:

 

Darwin Kernel Version 14.0.0

durankeeley:xnu-2782.1.97-opemu1/BUILD/obj/RELEASE_X86_64 x86_64

 

no works on Athlon 5350 

Link to comment
Share on other sites

boot flags : kext-dev-mode=1 npci=0x2000 -v - f

Enoch-rev.2468

 

did you start with this parameter?

 

no , with chameleon 2510 , drapeau : npci=0x2000 UsekernelCache=No (kext 13A603) , kabini is ssse3 but fsb=100 Mhz and frequency=2049 Mhz . 

Link to comment
Share on other sites

no , with chameleon 2510 , drapeau : npci=0x2000 UsekernelCache=No (kext 13A603) , kabini is ssse3 but fsb=100 Mhz and frequency=2049 Mhz .

you need "kext-dev-mode=1 and -f"

@yakei

 

UseKernelCache=No is the same as "-f", they are both flags to ignore kernel cache.

  • Like 1
Link to comment
Share on other sites

1010_BSA_RC2 without opemu

 

Yosemite Version: fresh install of 10.10.1

boot time / resultboot time is good

Launchpad icons: clean, no error found

Desktop Icons: clean, no error found

Graphics Desktop:sometimes problems, raphic dissolution

AppStore: working

Safari / WebKit: Works with WebKit

Graphics driver error: not found

Video/ Youtube streams tested with Safari: video is very good, no errors found

 

blurred images see post 323

 

:thumbsup_anim:  :D

  • Like 1
Link to comment
Share on other sites

Unfortunately edit graphics with preview is still sh_it :(  only with graphics tool that are not from Apple, is the process of graphics possible.

Link to comment
Share on other sites

Just for the sake of archiving info here:

 

The OPEMU used in BSA_rc2 that I uploaded is attached, it's an old gen 2 bronya optimised.

 

 

Yes , but this don't old opemu ! This new opemu optimized !

1010_BSA_RC1 with opemu

1010_BSA_RC1_v2 with opemu

1010_BSA_RC2 without opemu

1010_BSA_RC2_v2 without opemu

 

 

What is it ? Why did you write this ?

What does it mean "without opemu" ? I checked in the kernel , there is opemu !

Please stop if you're not a programmer .

Thank you for understanding !

  • Like 3
Link to comment
Share on other sites

Yes , but this don't old opemu ! This new opemu optimized !

Hi Bronya

 

Sorry for the miss understanding, I thought it was a older OPEMU because when Sinetek released OPEMU III he had the the commented header like this

 

 

                                                         8888888 8888888 8888888 
                                                            888     888     888   
                                                            888     888     888   
   .d88b.  88888b.   .d88b.  88888b.d88b.  888  888         888     888     888   
  d88""88b 888 "88b d8P  Y8b 888 "888 "88b 888  888         888     888     888   
  888  888 888  888 88888888 888  888  888 888  888         888     888     888   
  Y88..88P 888 d88P Y8b.     888  888  888 Y88b 888         888     888     888   
   "Y88P"  88888P"   "Y8888  888  888  888  "Y88888       8888888 8888888 8888888 
           888                                                                    
           888                                                                    
           888                                                                    

 

 

Where as OPEMU II had a header like so

 

 

/*   ** SINETEK **
 *   This is called the Opcode Emulator: it traps invalid opcode exceptions
 *   and modifies the state of the running thread accordingly.
 *   There are two entry points, one for user space exceptions, and another for
 *   exceptions coming from kernel space.
 *
 * STATUS
 *  . SSSE3 is implemented.
 *  . SYSENTER is implemented.
 *  . RDMSR is implemented.
 *
 */

 

 

However I now see that the following is in your OPEMU

 

 

* STATUS
 *  . SSE3 is implemented.
 *  . SSSE3 is implemented.
 *  . SYSENTER is implemented.
 *  . SYSEXIT is implemented.
 *  . RDMSR is implemented.

 

  • Like 1
Link to comment
Share on other sites

Hi Bronya

 

Sorry for the miss understanding, I thought it was a older OPEMU because when Sinetek released OPEMU III he had the the commented header like this

 

 

                                                         8888888 8888888 8888888 
                                                            888     888     888   
                                                            888     888     888   
   .d88b.  88888b.   .d88b.  88888b.d88b.  888  888         888     888     888   
  d88""88b 888 "88b d8P  Y8b 888 "888 "88b 888  888         888     888     888   
  888  888 888  888 88888888 888  888  888 888  888         888     888     888   
  Y88..88P 888 d88P Y8b.     888  888  888 Y88b 888         888     888     888   
   "Y88P"  88888P"   "Y8888  888  888  888  "Y88888       8888888 8888888 8888888 
           888                                                                    
           888                                                                    
           888                                                                    

 

 

Where as OPEMU II had a header like so

 

 

/*   ** SINETEK **
 *   This is called the Opcode Emulator: it traps invalid opcode exceptions
 *   and modifies the state of the running thread accordingly.
 *   There are two entry points, one for user space exceptions, and another for
 *   exceptions coming from kernel space.
 *
 * STATUS
 *  . SSSE3 is implemented.
 *  . SYSENTER is implemented.
 *  . RDMSR is implemented.
 *
 */

 

 

However I now see that the following is in your OPEMU

 

 

* STATUS
 *  . SSE3 is implemented.
 *  . SSSE3 is implemented.
 *  . SYSENTER is implemented.
 *  . SYSEXIT is implemented.
 *  . RDMSR is implemented.

 

Which is opemu 2 with enhancements from me and Bronya.

I implemented SSE3 parts and SYSEXIT.

Bronzovka did other improvements.

  • Like 7
Link to comment
Share on other sites

:)

 

Merci de laisser faire à ceux qui savent faire et qui connaissent le codage , cela évitera de voir fleurir des absurdités de pseudos kernel .

je pense que seul maitrise revient à Andy , bronya et Sinetek...........?? et le nouveau membre takiguru

 

je n'ai aucune compétence dans ce domaine , donc je m'abstiens ............ 

 

Thank you to let those who know how and who know the code, it will avoid seeing bloom kernel nicks nonsense.
I think that only mastery returns to Andy, and Bronya Sinetek ........... ?? and the new member takiguru
 
I have no expertise in this area, so I refrain ............
 
with the kernel Bronya and of takiguru also works perfectly for the moment on my wayward Kabini :)
 
Merci à eux :)
Edited by gils83
  • Like 3
Link to comment
Share on other sites

 Share

×
×
  • Create New...