Jump to content

Zenith432

Zenith432

Member Since 21 Jun 2009
Offline Last Active Jun 17 2016 10:39 AM
*****

#2239829 Clover General discussion

Posted by Zenith432 on 23 May 2016 - 09:57 AM

Is there a possibility for packing theme files into archives like tar or dmg or iso, and having Clover read them from packed instead of having them on filesystem as individual files?

#2239764 Clover General discussion

Posted by Zenith432 on 22 May 2016 - 08:30 PM

These are times I get... Grub ExfatSATA disk (MBR) 1:074USB disk (MBR) 1:417 Grub NTFSSATA disk (MBR)0:9610:9430:944SATA disk (GPT)0:9440:944 EDK2 Fat DriverSATA disk (GPT) 0:926USB disk (MBR) 0:926 VBox HFSSATA disk (GPT)0:9261:118 on system partitionUSB disk (MBR)1:100 on system partition So none is especially snappy. Grub EXFAT is bundled into boot[367].  You can disable by./ebuild.sh -D NO_GRUB_DRIVERS  I said MSR (Microsoft Reserved Partition) not MBR. Those 2 partitions are exfat. (Data 1 and Data 2)The were created recently to replace 2 NTFS. So I guess the slow down happened after that.  The only efi drivers I have are fsinject and hfsplus.

#2239287 EDK2 (UEFI) Shell issues/bugs?

Posted by Zenith432 on 20 May 2016 - 07:16 AM

These patches have been committed to EDK2 in github - but ShellBinPkg not rebuilt yet. cecekpawon - Your problems with shell from ShellBinPkg are not universal.  The newer shell for me has always worked, though may be bugs in it that I didn't notice in features I don't use.It could be problem in the Windows build.  For instance, the shell converts device-paths to text on startup (for the map), which can cause potential hangs with use of va_list.

#2238842 Clover Bug/Issue Report and Patch

Posted by Zenith432 on 18 May 2016 - 07:28 AM

Slice, You're right (and Micky1979 too in post #831).  The bug is that argument TagPtr dict to GetTagCount is NULL.I got confused about this.Because the parameter dict is passed in ECX, and ECX in the dump is 0x3C8CF0 - looking valid.So I mistook this pointer for dict.But now I look at the disassembly again.At the crash pointESI holds the value of dict->tag, which was taken from NULL->tag,and ECX holds  the value of ESI->tagNext, which corresponds to statement tagList = tag->tagNext.So value of parameter dict is not in a register anymore at the crash point.  dict is forgotten.  The onlyplace it's used later in the func is dict->type, which is remembered in register EDX (it's kTagTypeArray,but really taken from NULL parameter....)  It's a freak accident that NULL->type ended up as kTagTypeArray.Otherwise GetTagCount would be exited with 0 as it should.

#2238585 Clover Bug/Issue Report and Patch

Posted by Zenith432 on 17 May 2016 - 08:20 AM

calibre, it's not the same as Micky1979's crash.Page fault, instruction "cmp dword [esi], byte +0x2" and esi = 0xFFFFFFFF.I'll try to locate it later on.

#2238556 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 17 May 2016 - 06:25 AM

crazybirdy, Please try attached fake_efi.c with trunk 2815.  Overwrite i386/libsaio/fake_efi.c and make a clean build of boot.  Check if it prevents the freeze, or if the workaround causes a noticeable delay.  If it's ok I'll commit it. Thanks. Edit: Attachment removed, committed to r2817.

#2238548 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 17 May 2016 - 05:43 AM

And I found the root cause is (2807) trunk/i386/libsaio/fake_efi.c, it didn't boot with my old MB asus P5B. After replace to old fake_efi.c (2790-2806) and rebuild boot file, the kext reading works fine and boot to 10.11.4 system now.(patched kernel) So I can't use Enoch branch 2816 now.It's because of this // // FIXME: PM Timer is usually @ 0x408, but its position is relocatable // via PCI-to-ISA bridge. The location is reported in ACPI FADT, // PM Timer Block address - zenith432 //Your PM Timer is not at 0x408, so esi always gets the same value, and the continue statement there generates an infinite loop. As a temporary workaround, comment out the continue statement a few lines below this comment. I need to think of a permanent solution. The fix to the kernel path for Enoch (or this problem) has not been checked in yet.

#2238364 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 16 May 2016 - 03:26 PM

crazybirdy, Shanee --> trunk@2815 both issues.

#2238352 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 16 May 2016 - 02:06 PM

csrutil shows you the right status, as it gets the value by doing a system call to get it from the kernel.The source code for KCPM is unavailable, but if it gets the value from NVRam variable csr-active-config, it is doing the wrong thing.boot.efi reads this nvram variable, and puts the value in kernel struct boot-args.  Chameleon sets this value directly in boot-args without bothering with NVram, which it does not support.  The value used by the kernel is the one in the struct, and can be obtained from the kernel with a system call.If it makes you feel better, I'll add a printout into bdmesg to show the value it is setting.  This won't alter the behavior of csrutil or KCPM.  Booting with r2812 I'm still having issues with CsrActiveConfig. I have it set in my boot.plist as follows, <key>CsrActiveConfig</key> <string>103</string>csrutil status results in, System Integrity Protection status: disabled. But for example KCPM U...

#2238316 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 16 May 2016 - 10:04 AM

This is probably due to the fix of checkOSVersion() in r2806. Before, it was sometimes mis-identifying Yosemite as ElCapitan or vice versa. Now, it identifies correct OS version, which may expose bug in selection of kernel somewhere else. I'll check it out when I have time. Enoch 2812 (download section) didn't work here as pics, same as svn trunk 2807. All work fine with Enoch 2795 (download section). FYI.

#2237284 Clover Bug/Issue Report and Patch

Posted by Zenith432 on 11 May 2016 - 06:46 AM

looks like a bug with my PC but unfortunately he worked previouslyCould be a detachable USB device or bios update that changed bios behavior.  PS: could also be new PCI adapter with option rom that hooks and extends int 13. I may propose a simple workaround if (!BlockSize) BlockSize = 512; We should check alignment produced by different compilers.Sound good. I suggest Micky1979 test it before commit to see what it does to this device. At first he said it's compiler-dependent (which it shouldn't be). But now it looks like crash is also on GCC. So just device-dependent or bios-dependent.

#2236940 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 09 May 2016 - 02:24 PM

Great.   If you're making commit may I suggest breaking up initialize_runtime() into 2 parts and in boot() do void boot(int biosdev){ // Enable A20 gate before accessing memory above 1Mb. // Note: malloc_init(), called via initialize_runtime() writes // memory >= 1Mb, so A20 must be enabled before calling it. - zenith432 zeroBSS(); enableA20(); malloc_init(0, 0, 0, malloc_error); common_boot(biosdev);}As I said above, I believe calls to delay() inside enableA20() work today even without BSS being initialized, but it's not 'good form' to rely on this. If you think it's ok as is then leave it.

#2236819 Chameleon 2.3svn Official PKG Installer & Binaries

Posted by Zenith432 on 08 May 2016 - 07:40 PM

This is for the record.  I doubt any1 is listening... I believe the problem with commit 2602 was that some bioses don't support int 0x15/func 0x2401 for enabling A20, and were returning an error (CF set).  The function didn't check for this error.I've reinstated the bugfix from commit 2602 in commit 2807 - but a little different.  The code for enableA20 via PS/2 controller that is known to work is used.  The enabling of A20 must be done before malloc_init() - because this function references memory >= 1MB.The code in working enableA20() calls biosfn delay(), which is used in this instance before BSS section is initialized.  I believe this is ok, as whatever random data in bb struct is unused registers and does not interfere. Additionally, fixed some bugs with va_list, and bug in random generator in fake_efi introducted in commit 2697. Yep I know... But Bronya try to fix the problem w/o "lost" the 2602 commit... That's why I'm asking... I'm...

#2236817 Clover Bug/Issue Report and Patch

Posted by Zenith432 on 08 May 2016 - 07:32 PM

Just note about the nasm patch that I had some problem with the tabulation in the patch.  At first it wasn't exactly like the nasm sources, so patch was failing.  Make sure the tabulation is exactly like in buildnasm.sh - and check that patch is working and not outputting error message.

#2236740 Clover Bug/Issue Report and Patch

Posted by Zenith432 on 08 May 2016 - 02:44 PM

I already posted scripts in post #706.  Apple has published cctools 877.8, so buildmtoc.sh can be updated. The warnings are from edk2/BaseTools, which are part of edk2 - not Clover.  They're only built once, so ignore them.

© 2016 InsanelyMac  |   News  |   Forum  |   Downloads  |   OSx86 Wiki  |   Mac Netbook  |   PHP hosting by CatN  |   Designed by Ed Gain  |   Logo by irfan  |   Privacy Policy