Jump to content

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


Duran Keeley
 Share

1,948 posts in this topic

Recommended Posts

and… palignr
 
in opemu
 
void palignr128(ssp_m128 *src, ssp_m128 *dst, ssp_m128 *res, uint8_t IMM)
{
    
    char buf [32];
    int i;
    
    memcpy (&buf[0], src->s32, 16);
    memcpy (&buf[16], dst->s32, 16);
    
    for (i = 0; i < 16; i++)
        if (IMM >= 32 || IMM + i >= 32)
            res->s8[i] = 0;
        else
            res->s8[i] = buf[IMM + i];
}
this is from gcc src
 
 
orig src
 
compute_correct_result_128 (int *i1, int *i2, unsigned int imm, int *r)
{
  char buf [32];
  char *bout = (char *) r;
  int i;


  memcpy (&buf[0], i2, 16);
  memcpy (&buf[16], i1, 16);


  for (i = 0; i < 16; i++)
    if (imm >= 32 || imm + i >= 32)
      bout[i] = 0;
    else
      bout[i] = buf[imm + i];
}
This function is compute imm num
 
Why the result be res->s8 = buf[iMM + i];
 
in ssse3-palignr.c
Return value is   *(__m128i *) r = _mm_alignr_epi8 (t1, t2, imm(0-31));
 
 
in MSDN data
 
    a.m128i_u32[3] = 0x01234567;
    a.m128i_u32[2] = 0x89ABDCEF;
    a.m128i_u32[1] = 0x01234567;
    a.m128i_u32[0] = 0x89ABCDEF;
    b.m128i_u32[3] = 0xFFFFEEEE;
    b.m128i_u32[2] = 0xDDDDCCCC;
    b.m128i_u32[1] = 0xBBBBAAAA;
    b.m128i_u32[0] = 0x99998888;


    // A right align value of four should remove the lowest 4 bytes of "b"
    __m128i res = _mm_alignr_epi8( a, b, 4 );


    printf_s("Original a: 0x%016I64x%016I64x\nOriginal b: 0x%016I64x%016I64x\n",
        a.m128i_u64[1], a.m128i_u64[0],
        b.m128i_u64[1], b.m128i_u64[0]);


    printf_s("Result res: 0x%016I64x%016I64x\n",
        res.m128i_u64[1], res.m128i_u64[0]);


Return value


ralign=IMM


r := (CONCAT(a,  >> (ralign * 8)) & 0xffffffffffffffff
and Should be res u64 dst & src u32
 
 
 
this is PALIGNR Comment
 
 
PALIGNR combines two 16-byte vectors into a single, 32-byte, and if it chooses from a range of 16 bytes and stores in the targeting vector. Start of the range is indicated by the third argument to the command, 8-bit immediate constant imm8.
 
PALIGNR commandment XMM1, XMM2, imm8 (128-bit arguments, ie. 16-element vectors) corresponds
 {Link vectors} 
 {Temp - 32-element vector}
 for i := 0 to 15 do begin
   temp[i]    := xmm1[i];
   temp[i+16] := xmm2[i];
 end;


 {Selects a portion}
 for i := 0 to 15 do
   xmm1[i] := temp[i + imm8];
Eg.
 
        0  1  2  3  4  5  6  7  8  9  10 11 12 13 14 15
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
xmm1 = |W |i |k |i |p |e |d |i |a |, |  |W |o |l |n |a |  
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
       
        16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
xmm2 = |  |E |n |c |y |k |l |o |p |e |d |i |a |  |  |  |
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
The result of PALIGNR XMM1, XMM2, 11:
 
        0  1  2  3  4  5  6  7  8  9  10 11 12 13 14 15
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
xmm1 = |W |o |l |n |a |  |E |n |c |y |k |l |o |p |e |d |  
       +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+

this Function PALIGNR Is simply wrong

 

I dont understand Why the test is successful

Link to comment
Share on other sites

New program for VMWare bit mask calculation:

#include <stdint.h>
#include <stdio.h>

#define _Bit(n)			(1U << n)
#define CPUID_FEATURE_SSSE3     _Bit(9)  /* Supplemental SSE3 instructions */

static inline void
do_cpuid(uint32_t selector, uint32_t *data)
{
	__asm__ volatile ("cpuid"
		: "=a" (data[0]),
		  "=b" (data[1]),
		  "=c" (data[2]),
		  "=d" (data[3])
		: "a"(selector),
		  "b" (0),
		  "c" (0),
		  "d" (0));
}

static inline void
print_vmware_flag(uint32_t data)
{
    int i;
    printf("cpuid.1.edx = \"");
    for (i = 31; i > 27; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 27; i > 23; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 23; i > 19; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 19; i > 15; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 15; i > 11; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 11; i > 7; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 7; i > 3; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf(":");
    for (i = 3; i > -1; i--)
    {
        if (data & _Bit(i))
        {
            printf("1");
        } else {
            printf("0");
        }
    }
    printf("\"\n");
}

int main(void)
{
	uint32_t reg[4];
	do_cpuid(1, reg);
    printf("With SSSE3:\n");
	printf("ECX: 0x%X\n", reg[2]);
	printf("EDX: 0x%X\n", reg[3]);
    print_vmware_flag(reg[3]);
    printf("\n");

    printf("Without SSSE3:\n");
    printf("ECX: 0x%X\n", reg[2]);
    printf("EDX: 0x%X\n", (reg[3] & ~CPUID_FEATURE_SSSE3));
    print_vmware_flag(reg[3] & ~CPUID_FEATURE_SSSE3);
	return 0;
}

Hard work made easy

Link to comment
Share on other sites

New kernel to test.

I again modded it to software data types.

Hopefully with more success now.

BSA_YOS_R7D.zip

  • Like 5
Link to comment
Share on other sites

With BSA_YOS_R7D and system.kext replaced:

 

IMG_0764_www.kepfeltoltes.hu_.jpg

Which IOPCIFamily and AppleACPIPlatform kexts do you use?

It seems like an incompatibility.

This is not kernel related...

Link to comment
Share on other sites

My first results:

 

- boot successful

- able to make screenshot, but wrong colors:

 

post-302440-0-31570400-1422200906_thumb.png

post-302440-0-34873500-1422200947_thumb.png

 

- selection of file only with name

- Safari works, App Store does not work with Webkit patch. Will try with original frameworks files.

 

Update:

Errors in log:

Jan 25 17:49:01 Mac-Pro-Andrey.local com.apple.WebKit.WebContent[423]: Error returned from iconservicesagent: Error Domain=NSCocoaErrorDomain Code=4099 "Не удалось установить связь с программой-помощником." (The connection to service named com.apple.iconservices was invalidated.) UserInfo=0x7f8ad96277c0 {NSDebugDescription=The connection to service named com.apple.iconservices was invalidated.}
Jan 25 17:49:01 Mac-Pro-Andrey.local com.apple.WebKit.WebContent[423]: Error returned from iconservicesagent: Error Domain=NSCocoaErrorDomain Code=4099 "Не удалось установить связь с программой-помощником." (The connection to service named com.apple.iconservices was invalidated.) UserInfo=0x7f8ad94848f0 {NSDebugDescription=The connection to service named com.apple.iconservices was invalidated.}
Jan 25 17:49:01 Mac-Pro-Andrey.local com.apple.WebKit.WebContent[423]: Error returned from iconservicesagent: Error Domain=NSCocoaErrorDomain Code=4099 "Не удалось установить связь с программой-помощником." (The connection to service named com.apple.iconservices was invalidated.) UserInfo=0x7f8adb511a10 {NSDebugDescription=The connection to service named com.apple.iconservices was invalidated.}
Jan 25 17:49:01 Mac-Pro-Andrey.local com.apple.WebKit.WebContent[423]: Error returned from iconservicesagent: Error Domain=NSCocoaErrorDomain Code=4099 "Не удалось установить связь с программой-помощником." (The connection to service named com.apple.iconservices was invalidated.) UserInfo=0x7f8ad96778f0 {NSDebugDescription=The connection to service named com.apple.iconservices was invalidated.} 

Testing continues... 

  • Like 1
Link to comment
Share on other sites

I am using the original ones, so that is the issue? Can you give me the link to the proper ones? 

Try with original System.kext.

It would seem that is the reason the link fails.

It's either that or your AppleACPIPlatform.kext or IOPCIFamily.kext files have gone bad.

Link to comment
Share on other sites

Try with original System.kext.

It would seem that is the reason the link fails.

It's either that or your AppleACPIPlatform.kext or IOPCIFamily.kext files have gone bad.

 I've changed them to the ones you updated, and now the issue is the boot process stucks at hfs: mounted...

Link to comment
Share on other sites

It seems for fx the same i manage to fix USB 3.0 port that wasnt at all working and faster booting for my not taken place i thing the same or slower 

Not working dvd/cd i need to fix it

Install lastes AppleNforce but still not working dvd

Link to comment
Share on other sites

New kernel to test.

I again modded it to software data types.

Hopefully with more success now.

Not goes white screen with rainbow spinning
IMG_0033.JPG
 

 

is fine if I install my kext AppleAcpi ... / IOPCIFAmily ......
file:///Users/carlo/Desktop/sysctl%20-a%20
Link to comment
Share on other sites

My result with r7d:

Boot ok

Safari and Appstore work (testet with webkit fix)

Like any other Kernel before:

sometimes wrong colors in covers in itunes, in icons for jpg in finder and pictures in preview an iwork

iphoto gives me an error at importing pictures: not enough space on harddisk

Link to comment
Share on other sites

Hi AnV

Your R7D I try Fix

And delete the old code

Please check my changes

 

in opemu ssse3 opcode

m128 palignr
                case 0x0F:
                    xmmres.i = ssp_alignr_epi8_SSSE3 (xmmdst.i, xmmsrc.i, (const int)operand);
                    ins_size++;
                    break;

m64 palignr
                case 0x0F:
                    mmres.m64 = ssp_alignr_pi8_SSSE3 (mmdst.m64, mmsrc.m64, (const int)operand);
                    ins_size++;
                    break;

/EXTERNAL_HEADERS/SSEPlus/native/SSEPlus_native_SSSE3.h

/** \SSSE3{Native,_mm_alignr_epi8} */
SSP_FORCEINLINE __m128i ssp_alignr_epi8_SSSE3 (__m128i a, __m128i b, int n)
{
    n = (n>=32) ?  32 : n;
    switch( n )
    {
        CASE_32( ssp_alignr_epi8_REF, a, b );
    }
}
/** \SSSE3{Native,_mm_alignr_pi8} */
SSP_FORCEINLINE __m64 ssp_alignr_pi8_SSSE3 (__m64 a, __m64 b, int n)
{
    n = (n>=16) ?  16 : n;
    switch( n )
    {
        CASE_16( ssp_alignr_pi8_REF, a, b );
    }
}

/EXTERNAL_HEADERS/SSEPlus/emulation/SSEPlus_emulation_REF.h

SSP_FORCEINLINE __m128i ssp_alignr_epi8_REF (__m128i a, __m128i b, const int ralign)
{
    ssp_m128 C[3];
    ssp_s8 * tmp;
    int i, j;

    if (ralign <0) return b; //only shift to right, no negative
    C[2].i = _mm_setzero_si128();
    if (ralign > 32) return C[2].i;
    C[1].i = a;
    C[0].i = b;
    tmp = & (C[0].s8[0]);

    for (i=ralign+15, j=15; i >=ralign; i--, j--) {
        C[2].s8[j] = tmp[i];
    }

    return C[2].i;
}

/**  \SSSE3{Reference,_mm_alignr_pi8}
 \n \b NOTE: The user must call _mm_empty() after a call to this function.
 */
SSP_FORCEINLINE __m64 ssp_alignr_pi8_REF (__m64 a, __m64 b, const int ralign)
{
    ssp_m64 C[3];
    ssp_s8 * tmp;
    int i, j;
    
    if (ralign <0) return b; //only shift to right, no negative
    C[2].u32[0] = 0;
    C[2].u32[1] = 0;
    if (ralign > 16) return C[2].m64;
    C[1].m64 = a;
    C[0].m64 = b;
    tmp = & (C[0].s8[0]);
    
    for (i=ralign+7, j=7; i >=ralign; i--, j--) {
        C[2].s8[j] = tmp[i];
    }
    
    return C[2].m64;
}

SSEPlus code seems to be some difference with MSDN or gcc

 

Returns the value of different

 

Now Screenshots can works

But the color is not correct
 

1010-anv-r7-kernel.zip

1010-anv-r7-fix-patch.diff.zip

  • Like 5
Link to comment
Share on other sites

With BSA_YOS_R7D system overall performance is very good (Geekbench score is much lower than on Mavericks), but wrong colors in Launchpad, Safari and login screen! Appstore works!

 

With 1010-anv-r7-kernel.zip almost the same experience, but also wrong colors in Desktop.

 

Is there any fix for that?

Not untill we know what causes this. We'll need to check all instructions.
  • Like 1
Link to comment
Share on other sites

New kernel to test.

I again modded it to software data types.

Hopefully with more success now.

BSA_YOS_R7D

 

Hi Andy,

 

i test the version X86_64_R7D with webkit.

up to now very stable, app store works, safari once (within one hour) an error message brought.

 

the color are clear, youtube runs well. 

I installed additionally the safari picture preview kit.

Without the two kit unfortunately safari crashes, App Store can not correctly be opened, Preview Picture are distorted / in Safari are not correctly represented.

 

thanks

Link to comment
Share on other sites

kern_return_t glrPopulateDevice(io_connect_t connect, display_info_t* dinfo)
{
	/*
	 * For Intel, dinfo->config[0] determines type
	 *
	 *   0x100000U - "HD Graphics"
	 *   0x080000U - "GMA X3100"
	 *   0x040000U - "GMA 950"
	 *   else      - "GMA 900"
	 *
	 * Strings are
	 *   "Intel %s OpenGL Engine" (%s == model)
	 *   "Intel %s Compute Engine" (%s == model)
	 */
	strcpy(dinfo->engine1, "VMware SVGA II OpenGL Engine");
	strcpy(dinfo->engine2, "VMware SVGA II Compute Engine");
	return ERR_SUCCESS;
}
_Bool gldGetVersion(int* arg0, int* arg1, int* arg2, int* arg3)
{
	if (!glr_io_data.dinfo)
		return 0;
	/*
	 * Note these numbers are from OS 10.6.5
	 *
	 * arg3 is the renderer ID, which is
	 *   0x4000 for Intel GMA 950
	 *   0x0400 for GLRendererFloat
	 *   (see CGLRenderers.h)
	 */
	*arg0 = 3;
	*arg1 = 1;
	*arg2 = 0;
	*arg3 = kCGLRendererIntel900ID & 0xFFFFU;
	GLDLog(2, "%s(%d, %d, %d, %#x)\n", __FUNCTION__, *arg0, *arg1, *arg2, (uint32_t) *arg3);
	return 1;
}
HIDDEN
IOReturn CLASS::get_config(uint32_t* c1, uint32_t* c2, uint32_t* c3, uint32_t* c4, uint32_t* c5)
{
	uint32_t const vram_size = m_provider->getVRAMSize();

	/*
	 * c1 used by GLD to discern Intel 915/965/Ironlake(HD)
	 *   0x100000U - "HD Graphics"
	 *   0x080000U - "GMA X3100"
	 *   0x040000U - "GMA 950"
	 *   else      - "GMA 900" [Unsupported in OS 10.7]
	 */
	*c1 = version_major >= 11 ? 0x40000U : 0U;
#ifdef USE_OWN_GLD
#if 0
	*c2 = static_cast<uint32_t>(m_provider->getLogLevelGLD()) & 7U;
#else
	*c2 = 1U;	// set GLD logging to error level
#endif
#else
	*c2 = 0U;
#endif
	*c3 = vram_size;	// total memory available for textures (no accounting by VMsvga2)
	*c4 = vram_size;	// total VRAM size
#if __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ >= 1060
	*c5 = m_provider->getSurfaceRootUUID();
#else
	*c5 = 0U;
#endif
	DVLog(2, "%s(*%u, *%u, *%u, *%u, *%#x)\n", __FUNCTION__, *c1, *c2, *c3, *c4, *c5);
	return kIOReturnSuccess;
}
.encoding = "windows-1252"
config.version = "8"
virtualHW.version = "10"
scsi0.present = "TRUE"
scsi0.virtualDev = "lsilogic"
sata0.present = "TRUE"
memsize = "2048"
sata0:0.present = "TRUE"
sata0:0.fileName = "Yosemite.vmdk"
sata0:1.present = "TRUE"
sata0:1.autodetect = "TRUE"
sata0:1.deviceType = "disk"
ethernet0.present = "TRUE"
ethernet0.connectionType = "nat"
ethernet0.virtualDev = "e1000e"
ethernet0.wakeOnPcktRcv = "FALSE"
ethernet0.addressType = "generated"
usb.present = "TRUE"
ehci.present = "TRUE"
ehci.pciSlotNumber = "34"
sound.present = "TRUE"
sound.virtualDev = "hdaudio"
sound.fileName = "-1"
sound.autodetect = "TRUE"
pciBridge0.present = "TRUE"
pciBridge4.present = "TRUE"
pciBridge4.virtualDev = "pcieRootPort"
pciBridge4.functions = "8"
pciBridge5.present = "TRUE"
pciBridge5.virtualDev = "pcieRootPort"
pciBridge5.functions = "8"
pciBridge6.present = "TRUE"
pciBridge6.virtualDev = "pcieRootPort"
pciBridge6.functions = "8"
pciBridge7.present = "TRUE"
pciBridge7.virtualDev = "pcieRootPort"
pciBridge7.functions = "8"
vmci0.present = "TRUE"
smc.present = "TRUE"
hpet0.present = "TRUE"
ich7m.present = "TRUE"
usb.vbluetooth.startConnected = "TRUE"
board-id.reflectHost = "TRUE"
firmware = "bios"
displayName = "Mac OS X 10.9"
guestOS = "other-64"
nvram = "Mac OS X 10.9.nvram"
virtualHW.productCompatibility = "hosted"
powerType.powerOff = "soft"
powerType.powerOn = "soft"
powerType.suspend = "soft"
powerType.reset = "soft"
extendedConfigFile = "Mac OS X 10.9.vmxf"
uuid.bios = "56 4d 60 94 7b 5f f0 07-06 3e 57 83 a3 9b ca 83"
uuid.location = "56 4d 60 94 7b 5f f0 07-06 3e 57 83 a3 9b ca 83"
replay.supported = "FALSE"
replay.filename = ""
sata0:0.redo = ""
pciBridge0.pciSlotNumber = "17"
pciBridge4.pciSlotNumber = "21"
pciBridge5.pciSlotNumber = "22"
pciBridge6.pciSlotNumber = "23"
pciBridge7.pciSlotNumber = "24"
scsi0.pciSlotNumber = "16"
usb.pciSlotNumber = "32"
ethernet0.pciSlotNumber = "160"
sound.pciSlotNumber = "33"
vmci0.pciSlotNumber = "35"
sata0.pciSlotNumber = "36"
ethernet0.generatedAddress = "00:0c:29:9b:ca:83"
ethernet0.generatedAddressOffset = "0"
vmci0.id = "-1550071165"
monitor.phys_bits_used = "40"
vmotion.checkpointFBSize = "33554432"
cleanShutdown = "FALSE"
softPowerOff = "FALSE"
usb:1.speed = "2"
usb:1.present = "TRUE"
usb:1.deviceType = "hub"
usb:1.port = "1"
usb:1.parent = "-1"
toolsInstallManager.updateCounter = "6"
checkpoint.vmState = ""
sata0:1.fileName = "C:\Users\Vivianne\Documents\Virtual Machines\OS X 10.10\OS X 10.10.vmdk"
sata0:2.present = "TRUE"
sata0:2.autodetect = "TRUE"
sata0:2.deviceType = "cdrom-raw"
sata0:1.redo = ""
workingDir = "."
tools.syncTime = "FALSE"
sata0:2.startConnected = "TRUE"
floppy0.present = "FALSE"
mouse.vusb.enable = "TRUE"
keyboard.vusb.enable = "TRUE"
usb:0.present = "TRUE"
usb:0.deviceType = "hid"
usb:0.port = "0"
usb:0.parent = "-1"

I now have access to a real AMD Athlon II X2.

It's my aunts PC which I am allowed to use.

I just converted a Chameleon based image to a vmdk (dd2vmdk) and I set the following configuration in the VMX file to enable full support.

I can now thus test out some specifics...

A tad slow (boot takes very long) but at least it works...

I will install a VMWare virtual machine and use my latest AMD kernel in order to make it work and do tests.

I also noticed on boot some things were misplaced (Black blocks on the installer which seemed to disappear after a while).

NOTE: I used Chameleon on a USB stick (and Yosemite installer), converted this to a raw image and then I made the vmdk config.

Perfect thus for hackintosh testing on AMD in VMWare Virtual Machine.

This works without any unlocker in VMWare Workstation versions even up to 11!

Oh yeah, before I forget, the download for dd2vmdk:

http://dd2vmdk.sf.net

The irony wants... so far I see no pixel bug further.

Seems to be video driver related... we should examine the video drivers in more detail...

Maybe framebuffer issues???

Some part of the code must cause this and it's either the IOFramebuffer or the IOAccelerator that causes it...

If it's IOFramebuffer related - we can make our own... the IOFramebuffer class is fully documented by Apple and fully open for anyone to use. VMWare has an open source graphics driver too...

http://sourceforge.net/projects/VMsvga2

A example of an Apple made IOFramebuffer is NVDAResman.kext.

Attached is an open source snapshot of the drivers inside VMWare.

If it's IOAccelerator related - we can only patch... in order to make an IOAccelerator we would need to implement every OpenGL instruction till OpenGL 4.

Since they're not openly coded I don't know how they work and can't thus make an IOAccelerator.

Other than that it contains only start / stop / probe (standard IOKit functions). an example of an IOAccelerator is NVDANV50Hal.kext.

EDIT: It would seem the VMWare project contains a very basic IOAccelerator.

It's only capable of accelerated block transfer (blit) functions but it's a start...

We would need to add Apple's OpenGL functions though if we wanted to make our own... this plainly will not do for HW acceleration in any way.

Only accelerated drawing is the result of this driver... not more...

The most up to date version is the 2D GL branch.

This would need still the 3D enhancements but it's nice to see these are openly coded....

Obviously some Yosemite tweaking would be required too!

Interesting chunk of code above.

Selecting the graphics device.

This would need tweaking to Intel HD in order to make it work with Yosemite.

Some functions may also need changing...

Judging by code above: it seems to emulate a GMA950...

It's still cool to know how that works thus...

Next cool piece of code - Same thing the code defining the CGL renderer...

Would also need Intel HD..

And yet another reference to the renderer id...

Would seem it really needs updating...

It also seems to implement the OpenCL engine (not bad!)

gldGetRendererInfo -> needs mods too!

Also sets a renderer id... I'm starting to think most older no more supported cards could be supported these days using a properly coded IOAccelerator.

With all of that info it should actually be possible (and with some extra reverse engineering) to make our own!

OpenCL is implemented through the gldCreateComputeContext and gldCreateComputeContext functions (that and some extra functions).

gldInitDispatch defines apparently the supported OpenGL functions...

Interesting...

Wow! The desktop surely is colorful :-/

The irony... it would seem like??? palignr still??

Color is not quite present in the screen capture.

First press of command 3 also had as result that image data could not be saved...

I saw clearly that it mostly "misses color".

Just finished installing the VMWare tools... we'll see now if things are better...

Installing VMWare tools with custom parts did not help.

It would seem the problem is in the opemu (palignr or some other instruction responsable for doing some essential work!)

Ethernet in the VM required a small fix.

Add to Info.plist 0x10d38086 of Intel82574L.kext and remove IOPCISecondaryMatch.

Then ethernet works.

I put the desktop on auto-switch between images. image is now fine... (weird but hell)

post-100023-0-94019500-1422313872_thumb.png

vmsvga2-code-166.zip

post-100023-0-22248500-1422319465_thumb.png

  • Like 7
Link to comment
Share on other sites

I also now see the Safari bug mentioned.

It seems indeed like a misalignment of pixels.

I tried capturing a screen (took 5 attempts but hell).

Final result lacked color...


Installed WebKit patch, Safari seems stable.

Gonna test App Store now as final thing for today (need Xcode...)


App Store seems to work but very slowly...

Djeez

Then again... no packages are showing up....

Interesting case...

It does however show an update (which I'm applying through the Terminal)


Nope... no luck... no App Store...

Would seem the opemu needs further fixing...

I'm betting pallignr gives incorrect results..


I'm installing all and now App Store works!

It would seem I needed to login quickly this way.

Type gcc and select download Xcode.

Force it to download it and enter apple id.

Store works...


I noticed:

When I switched the VM back to windowed... desktop distorted again.

The problem is definitely somewhere in the blitting algorithms...

If we could pinpoint that function... bingo... we know what is wrong...

I'm betting there is an SSSE3 function in there and that's the one we need to examine...

post-100023-0-83372800-1422321542_thumb.png

  • Like 3
Link to comment
Share on other sites

i am noticing lot activity on AMD side kernel lately, i think you all getting close of fixing yosemite kernel of making run faster on AMD FX  but the processor is running fine on this kernel but the problem is due to GPU drivers  are not running fast as should be like maverick last kernel was almost perfect speed in video and cpu, i understand this is very hard to code this correctly, everytime apple unleashed next os verison it gets worst for all of you, because without you, this can never work on AMD and correct specification Custom parts unlike LINUX and WIndows does very well..and still Windows 10 is something that apple couldn't counter this time around, this might be the best Windows version to reign in again...and its up to apple decision to counter that with more changes or yosemite might bust when that happen..

Link to comment
Share on other sites

I also now see the Safari bug mentioned.

It seems indeed like a misalignment of pixels.

I tried capturing a screen (took 5 attempts but hell).

Final result lacked color...

Installed WebKit patch, Safari seems stable.

Gonna test App Store now as final thing for today (need Xcode...)

App Store seems to work but very slowly...

Djeez

Then again... no packages are showing up....

Interesting case...

It does however show an update (which I'm applying through the Terminal)

Nope... no luck... no App Store...

Would seem the opemu needs further fixing...

I'm betting pallignr gives incorrect results..

I'm installing all and now App Store works!

It would seem I needed to login quickly this way.

Type gcc and select download Xcode.

Force it to download it and enter apple id.

Store works...

I noticed:

When I switched the VM back to windowed... desktop distorted again.

The problem is definitely somewhere in the blitting algorithms...

If we could pinpoint that function... bingo... we know what is wrong...

I'm betting there is an SSSE3 function in there and that's the one we need to examine...

 

Hi AnV

 

TEST this

 

Replace 10.9.5 ImageIO no pixel bug

 

But Can not replace the file on 10.10

replace the file Must operate on other OS

 

path: /System/Library/Frameworks/ImageIO.framework

ImageIO-1095.zip

  • Like 2
Link to comment
Share on other sites

 Share

×
×
  • Create New...