El Capitan 10.11.1 versus the 5960X - A Collaborative Work-Thru

ertylol · March 22, 2016

I don't know man, I'm a bit discouraged. I'm pretty unmotivated to dedicate time to it. I tried diffing some of the files, I wanted to export a comparison of them but I don't know of a difftool that still do that... Diff merge can I think but I didn't want to spend the money on the full version right now.

Anyone know of a free diff tool for ANY platform that can export a diff/comparison of the 2 files?

And by all means, if you know some expert coders that may be able to help, feel free to reach out!

Hmm what specifically do you mean as far as exporting? Like keeping colors when outputting to a file instead of tty?

To retain color formatting you could do something like

script -q diffed-DSL.export colordiff --unified --ignore-matching-lines "DarwinDumper" DarwinDumper_2.9.9.2_AMI_X64_3333_Cap_danny/ACPI\ Tables/DSL/ DarwinDumper_2.9.9.2_AMI_X64_3333_Yos_danny/ACPI\ Tables/DSL/

Script is bsdutils, runs a command and saves it to file escape sequences and all. Other important thing to remember wrt colors is to do --color=always when piping through grep.

Should net you a nice cat for later, will look {censored}ed in a text editor tho obviously. But I mean why not just regular diff, open in Sublime and set syntax to diff? Much easier.

here's the output anyways in case it helps anyone

regular without color

Anyways I know basically nothing about any of this, there's probably a reason no one has been talking about it but like, so far all the info we have seems to be from boot logs. Has anyone actually tried attaching lldb from a second box and setting some breakpoints?

Edit: considering the xnu source is out and there's symbols for all the kexts etc, shouldn't that be possible at this point?

jamiethemorris · March 22, 2016

Hmm what specifically do you mean as far as exporting? Like keeping colors when outputting to a file instead of tty?

To retain color formatting you could do something like
script -q diffed-DSL.export colordiff --unified --ignore-matching-lines "DarwinDumper" DarwinDumper_2.9.9.2_AMI_X64_3333_Cap_danny/ACPI\ Tables/DSL/ DarwinDumper_2.9.9.2_AMI_X64_3333_Yos_danny/ACPI\ Tables/DSL/
Script is bsdutils, runs a command and saves it to file escape sequences and all. Other important thing to remember wrt colors is to do --color=always when piping through grep.

Should net you a nice cat for later, will look {censored}ed in a text editor tho obviously. But I mean why not just regular diff, open in Sublime and set syntax to diff? Much easier.

here's the output anyways in case it helps anyone

regular without color

Anyways I know basically nothing about any of this, there's probably a reason no one has been talking about it but like, so far all the info we have seems to be from boot logs. Has anyone actually tried attaching lldb from a second box and setting some breakpoints?

Edit: considering the xnu source is out and there's symbols for all the kexts etc, shouldn't that be possible at this point?

I honestly didn't know you could do that with sublime, thanks for the tip.

I didn't even think of using a debugger, but I don't own one nor would I really know what to do with it.

onmybikedrunk · April 8, 2016

PSA: GoFundMe.com Donations

Hey folks,

It seems as if this won't get resolved anytime soon. I've moved on with my 6-core 5930K and I know others have as well. In the interest of transparency and honesty:

I got an email saying that the gofundme page would be halted if I didn't sign up for and make a withdraw. There is no easy way to refund money thru gofundme.com so I made a withdraw in the full amount. The total donations was $250 (I donated the first hundred) in which $228 was transferred to my bank after the gofundme fees. Somewhere in the neighborhood of 9% (yikes!). I will be paypaling refunds minus the gofundme percentage. I expect the funds to hit my account around the 15th of April, according to gofundme's calculations. HOWEVER - in order to issue a refund, I need a bit more info than what gofundme gave me.

If anyone who has made a donation wants their money back (and I'm sure you do!) please contact me directly here via PM with your paypal or google wallet email address and your full name (that's all gofundme gave me) and I will send it over the second it hits my account! This forum is under my watch list, and I get a direct email whenever there is a new post if there is any other input on this...

What a shame... this totally just fizzled out. Seems that Apple has pushed the pros to the side (once again) and their "pro" hardware is now seriously outdated. Maybe we'll see something this summer... Who knows, but for now, I'm throwing in the towel.

Thank you to everyone who made a donation...

Danny

jamiethemorris · April 10, 2016

I tried to use the USB kexts from Yosemite to see if that had anything to do with the problem. For some reason it was still complaining about IOUSBHostFamily even though I was using older versions of the kexts that don't depend on that (doesn't exist in Yosemite). However, I noticed something interesting when changing the kexts back to how they were and clearing the kernel cache...

kxld[com.apple.iokit.IOPCIFamily]: The super class vtable 'vtable for IOMapper' for vtable 'vtable for AppleVTD' is out of date. Make sure your kext has been built against the correct headers.

kxld[com.apple.iokit.IOPCIFamily]: The super class vtable 'vtable for IOMapper' for vtable 'vtable for AppleVTD' is out of date. Make sure your kext has been built against the correct headers.

kxld[com.apple.iokit.IOPCIFamily]: The super class vtable 'vtable for IOMapper' for vtable 'vtable for AppleVTD' is out of date. Make sure your kext has been built against the correct headers.

Link failed (error code 5).

Prelink failed for com.apple.iokit.IOPCIFamily; omitting from prelinked kernel.

Can't use LegacyIOPCIFamily.kext - not linked.

It didn't say anything about AppleACPIPlatform at all.

We DO have the source available for IOPCIFamily. I can probably set up my computer to build this from the source, but does anyone know what the "correct headers" would be? Are these the xnu kernel headers? Basically I think we just need to build Mavericks' IOPCIFamily against El Capitan's headers.

jamiethemorris · April 11, 2016

running into this now...

/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:28:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:30:10: error: 
      'IOKit/pci/IOPCIConfigurator.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIConfigurator.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIConfigurator.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:31:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:264:9: error: use of
      undeclared identifier 'bzero'
        bzero(reserved, sizeof(IOPCIDeviceExpansionData));
        ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:862:46: error: use of
      undeclared identifier 'strlen'
            if (aString->isEqualTo("D3cold", strlen("D3cold")))
                                             ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:864:50: error: use of
      undeclared identifier 'strlen'
            else if (aString->isEqualTo("D3Hot", strlen("D3Hot")))

Seems like it can't find the C/C++ libraries. Not sure why. I think something is set up wrong with XCode or something.

augustopaulo · April 12, 2016

running into this now...

/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:28:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:30:10: error: 
      'IOKit/pci/IOPCIConfigurator.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIConfigurator.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIConfigurator.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:31:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:264:9: error: use of
      undeclared identifier 'bzero'
        bzero(reserved, sizeof(IOPCIDeviceExpansionData));
        ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:862:46: error: use of
      undeclared identifier 'strlen'
            if (aString->isEqualTo("D3cold", strlen("D3cold")))
                                             ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:864:50: error: use of
      undeclared identifier 'strlen'
            else if (aString->isEqualTo("D3Hot", strlen("D3Hot")))

Seems like it can't find the C/C++ libraries. Not sure why. I think something is set up wrong with XCode or something.

You're slowly becoming my hero :superman:

C'mon guys, jamiethemorris may have stumble into a very important finding for loading the IOPCIFamily.kext for the 5960X CPU!!

:blowup: I'M CALLING ALL THE DEVELOPERS, EXPERIENCE & NON-EXPERIENCE CODERS OF THIS FORUM TO HELP :gathering:

It's for this kind of persevering attitude that this whole site was built for...

Boa sorte,

Paulo

jamiethemorris · April 20, 2016

Still stuck on this. :-( :-( :-(

onmybikedrunk · April 20, 2016

Gofundme.com money has hit my account. I'm processing returns. If you do not live in the U.S.A. and you've donated, our only option is unfortunately PayPal, as Google Wallet will not let me send cash outside of the country (this is news to me).

1 contributor refunded: zedn

1 contributor pending PayPal account verification: thyselius

Thanks again for your donations. I will continue to monitor this thread and PM's.

Glad to see you guys are still trying :thumbsup_anim:

ertylol · April 21, 2016

running into this now...

/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:28:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:30:10: error: 
      'IOKit/pci/IOPCIConfigurator.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIConfigurator.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIConfigurator.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:31:10: error: 
      'IOKit/pci/IOPCIPrivate.h' file not found with <angled> include; use
      "quotes" instead
#include <IOKit/pci/IOPCIPrivate.h>
         ^~~~~~~~~~~~~~~~~~~~~~~~~~
         "IOKit/pci/IOPCIPrivate.h"
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:264:9: error: use of
      undeclared identifier 'bzero'
        bzero(reserved, sizeof(IOPCIDeviceExpansionData));
        ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:862:46: error: use of
      undeclared identifier 'strlen'
            if (aString->isEqualTo("D3cold", strlen("D3cold")))
                                             ^
/usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:864:50: error: use of
      undeclared identifier 'strlen'
            else if (aString->isEqualTo("D3Hot", strlen("D3Hot")))

Seems like it can't find the C/C++ libraries. Not sure why. I think something is set up wrong with XCode or something.

Looks like it's trying to include std/framework headers that you you don't have set up as such, so either change

#include <IOKit/pci/IOPCIPrivate.h>

to

#include "IOKit/pci/IOPCIPrivate.h"

etc, or set "always search user paths" to yes unless that causes further errors.

other errors suggest string.h isn't getting included so probably related to the others. might try to look at it if I find the time, but see if you can get it going

jamiethemorris · April 21, 2016

Thanks, I'll check that out. Changing those includes fixed that problem when I tried it before (didn't know about the search user paths thing though), so the main issue is not being able to find the c++ headers or whatever. I know string.h is definitely one of them, probably others.

jamiethemorris · April 23, 2016

I don't know if it's relevant, but I'm building as root in /usr/local/src.

jamiethemorris · April 24, 2016

I just checked and "always search user paths" is set to YES. I didn't have anything in there anyway though. I did discover that I had to add /usr/include to my header search paths and that fixed some of the errors for not being able to find the c headers. Now I'm running into a bunch of these:

In file included from /usr/include/c++/4.2.1/tr1/stdarg.h:37:
In file included from /usr/include/c++/4.2.1/tr1/cstdarg:37:
/usr/include/c++/4.2.1/cstdarg:60:11: error: no member named 'va_list' in the
      global namespace
  using ::va_list;
        ~~^
In file included from /usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:25:
In file included from /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/system.h:57:
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:172:41: error: 
      unknown type name 'va_list'
int vsscanf(const char *, char const *, va_list);
                                        ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:174:34: error: 
      unknown type name 'va_list'
extern int      vprintf(const char *, va_list);
                                      ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:175:52: error: 
      unknown type name 'va_list'
extern int      vsnprintf(char *, size_t, const char *, va_list);
                                                        ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:178:47: error: 
      unknown type name 'va_list'
extern int      vsprintf(char *bufp, const char *, va_list) __deprecated;
                                                   ^
In file included from /usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:25:
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/system.h:61:42: error: 
      unknown type name 'va_list'
extern void     _doprnt( const char *format, va_list *arg,

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:130:64: error: 
      expected function body after function declarator
int     IOLockSleep( IOLock * lock, void *event, UInt32 interType) __DARWIN14_AL...
                                                                   ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:133:49: error: 
      expected function body after function declarator
  ...AbsoluteTime deadline, UInt32 interType) __DARWIN14_ALIAS(IOLockSleepDea...
                                              ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:135:63: error: 
      expected function body after function declarator
void    IOLockWakeup(IOLock * lock, void *event, bool oneThread) __DARWIN14_ALIA...
                                                                 ^

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOMemoryDescriptor.h:155:11: error: 
      unknown type name 'upl_page_info_t'; did you mean 'vm_page_info_t'?
    const upl_page_info_t * pageList;
          ^~~~~~~~~~~~~~~
          vm_page_info_t
In file included from /usr/local/src/IOPCIFamily-257.20.1/IOPCIDevice.cpp:25:
In file included from /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/system.h:39:
In file included from /usr/include/mach/mach_types.h:108:
/usr/include/mach/vm_region.h:324:14: note: 'vm_page_info_t' declared here
typedef int *vm_page_info_t;
             ^

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLib.h:309:33: error: 
      unknown type name 'va_list'
void IOLogv(const char *format, va_list ap);
                                ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLib.h:339:6: error: 
      conflicting types for 'Debugger'
void Debugger(const char * reason);
     ^

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLib.h:309:33: error: 
      unknown type name 'va_list'
void IOLogv(const char *format, va_list ap);
                                ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLib.h:339:6: error: 
      conflicting types for 'Debugger'
void Debugger(const char * reason);
     ^

In file included from /usr/include/c++/4.2.1/tr1/cstdarg:37:
/usr/include/c++/4.2.1/cstdarg:60:11: error: no member named 'va_list' in the
      global namespace
  using ::va_list;
        ~~^

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:172:41: error: 
      unknown type name 'va_list'
int vsscanf(const char *, char const *, va_list);
                                        ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:174:34: error: 
      unknown type name 'va_list'
extern int      vprintf(const char *, va_list);
                                      ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:175:52: error: 
      unknown type name 'va_list'
extern int      vsnprintf(char *, size_t, const char *, va_list);
                                                        ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/libkern/libkern.h:178:47: error: 
      unknown type name 'va_list'
extern int      vsprintf(char *bufp, const char *, va_list) __deprecated;
                                                   ^
In file included from /usr/local/src/IOPCIFamily-257.20.1/IOPCIBridge.cpp:23:
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/system.h:61:42: error: 
      unknown type name 'va_list'
extern void     _doprnt( const char *format, va_list *arg,
                                             ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:130:64: error: 
      expected function body after function declarator
int     IOLockSleep( IOLock * lock, void *event, UInt32 interType) __DARWIN14_AL...
                                                                   ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:133:49: error: 
      expected function body after function declarator
  ...AbsoluteTime deadline, UInt32 interType) __DARWIN14_ALIAS(IOLockSleepDea...
                                              ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOLocks.h:135:63: error: 
      expected function body after function declarator
void    IOLockWakeup(IOLock * lock, void *event, bool oneThread) __DARWIN14_ALIA...
                                                                 ^

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/Kernel.framework/Headers/IOKit/IOMemoryDescriptor.h:155:11: error: 
      unknown type name 'upl_page_info_t'; did you mean 'vm_page_info_t'?
    const upl_page_info_t * pageList;
          ^~~~~~~~~~~~~~~
          vm_page_info_t

Perhaps I need to change my build flags or something?

jamiethemorris · April 25, 2016

Apparently I'm missing AspenSDK.xcconfig? Not sure if I actually need this or if it's an Apple internal thing.

jonisign · April 25, 2016

Hmm, maybe Apple will supply it if you ask nicely? How are you supposed to compile this source if it is dependent on things that Apple doesn't release?

jamiethemorris · April 25, 2016

Hmm, maybe Apple will supply it if you ask nicely? How are you supposed to compile this source if it is dependent on things that Apple doesn't release?

I guess you have to figure out what the build configuration is yourself. I'm sure to someone with more knowledge than me it's probably obvious what settings etc. you need to use to build a kext

jonisign · April 26, 2016

I guess you have to figure out what the build configuration is yourself. I'm sure to someone with more knowledge than me it's probably obvious what settings etc. you need to use to build a kext

Have you tried building the kext against mavericks sources? Does the same error occur?

jamiethemorris · April 26, 2016

Actually I haven't even tried building the mavericks kext yet. Right now I'm just trying to build the el Capitan one to make sure I have everything set up right, which clearly I don't.

maleorderbride · May 4, 2016

Not sure what to make of this, but I took an 10.11.4 drive from another X99 build, then connected it to my 14-core (which has the same problem as the 5960x) and I was able to boot up to desktop.

I was able to do that four times, then I am back to halt and restart on PCI Configuration Begins.

jonisign · May 4, 2016

Not sure what to make of this, but I took an 10.11.4 drive from another X99 build, then connected it to my 14-core (which has the same problem as the 5960x) and I was able to boot up to desktop.

I was able to do that four times, then I am back to halt and restart on PCI Configuration Begins.

Interesting, that's really strange––

So you're saying you got a halt at PCI Configuration Begin when installing from scratch, but taking a system that already has 10.11.4 installed booted right to the desktop with an E5-2697?

Any idea what changed when booting the fourth time? If you continue to try to boot can you get passed PCI Configuration Begin?

maleorderbride · May 4, 2016

Interesting, that's really strange––

So you're saying you got a halt at PCI Configuration Begin when installing from scratch, but taking a system that already has 10.11.4 installed booted right to the desktop with an E5-2697?

Any idea what changed when booting the fourth time? If you continue to try to boot can you get passed PCI Configuration Begin?

I took the change to mean something around caches, but I was also screwing with Clover on each boot, plus moving the PCI-e devices around.

When it was working, I was using UEFI clover and my EFI folder, but when it stopped I was trying to do a non-UEFI install as I have enough PCI-e devices in that build that it would hit a memmap issue 50% of the time.

I will test more today to see if I can get either a real answer, or at least a reproducible method. However, whether this was caches, something in the EFI, or some other variable I have overlooked, this still bodes rather well.

shilohhh · May 4, 2016

I took the change to mean something around caches, but I was also screwing with Clover on each boot, plus moving the PCI-e devices around.

When it was working, I was using UEFI clover and my EFI folder, but when it stopped I was trying to do a non-UEFI install as I have enough PCI-e devices in that build that it would hit a memmap issue 50% of the time.

I will test more today to see if I can get either a real answer, or at least a reproducible method. However, whether this was caches, something in the EFI, or some other variable I have overlooked, this still bodes rather well.

I've had a look at the X99 ACPI tables and they are huge! I don't own any X99 hardware so I can't really do any testing. It would be awesome if you could post a .ioreg file from ioregistryexplorer v2.1 and a Darwin dump of your system while running el cap (functional).

maleorderbride · May 4, 2016

I will post here if I get it back to desktop, take an IOreg, DD, anything else. Hell, remote you in!

shilohhh · May 4, 2016

I will post here if I get it back to desktop, take an IOreg, DD, anything else. Hell, remote you in!

If you don't mind, your entire clover folder from the EFI partition. I've already downloaded all the posted files from this thread to compare with. My thinking thus far is that the structure of the PCI devices in the ACPI tables may be conflicting with OS X.

jamiethemorris · May 4, 2016

I took the change to mean something around caches, but I was also screwing with Clover on each boot, plus moving the PCI-e devices around.

When it was working, I was using UEFI clover and my EFI folder, but when it stopped I was trying to do a non-UEFI install as I have enough PCI-e devices in that build that it would hit a memmap issue 50% of the time.

I will test more today to see if I can get either a real answer, or at least a reproducible method. However, whether this was caches, something in the EFI, or some other variable I have overlooked, this still bodes rather well.

interesting... Keep us posted on what else you find!

maleorderbride · May 4, 2016

Here is the EFI, It think that is back in the exact state it was when it first worked: https://dl.dropboxusercontent.com/u/6316213/Resources/Previously%20working%20EFI%20for%20X99-E%20WS%20%26%20E5-2697%20V3.zip

When it worked, I just assumed something had changed in 10.11.4 and I was out of the loop (haven't read this thread in awhile). I only came here after it stopped working =/

El Capitan 10.11.1 versus the 5960X - A Collaborative Work-Thru

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites