ProjectOSX

Welcome Guest!

Returning User? Login here.

Want membership privileges? Register here.

67 Pages V  « < 49 50 51 52 53 > »   
Start a new topic Add Reply
> How To Make Efi Bootloader
STLVNUB
post Apr 10 2012, 07:09 PM
Post #1001
May be time to add another new feature
Slice has memorydump kext
Can this be used for something useful ?
At work at moment will checkout when home
Good job blackosx

Azrock Z77 Pro3, G1610 AMD HD7750 Clover/Ozmosis
Toshiba A660-0MR00R I7-7740QM GT330M 8 Gig Clover

Slice
post Apr 11 2012, 01:35 PM
Post #1002
New DarwinDumper works flawlessly. What may be added here?
1. DeviceTree
ioreg -lxw0 -pIODeviceTree >ioregDT.txt
2. VideoBios dump?

Really I have some test version of boot1h2 so forget it. I will reinstall to actual version boot1h2.

@crazybirdy
Can you explain me five different installation of Clover by your installer? I am not understanding what is what.
What is boot0hfs? I have several partitions with FAT32, NTFS, HFS+. So why boot0hfs?
Пожалуйста, прочитайте ЧаВо!
i3-2120 GA-H61M-S1, Radeon HD6670, ALC887(VoodooHDA 2.8.4), OS⌘10.9.2, OS⌘ 10.7.5 Clover FakeSMC_plugins_3.3.1 Realtek LAN v3.1.2
blackosx
post Apr 11 2012, 01:52 PM
Post #1003
QUOTE (Slice @ Apr 11 2012, 02:35 PM) *
Really I have some test version of boot1h2 so forget it. I will reinstall to actual version boot1h2.

No problem. Thanks for confirming.

I've since fixed a small bug where the printing of disk numbers and partition names may sometimes be incorrect.
DarwinDumper_v1.2k.zip

EDIT: Link removed as newer version is posted here

This post has been edited by blackosx: Apr 18 2012, 09:33 AM
10.9.4 | Asus Maximus IV Gene-Z (PMpatched BIOS 3603) | i7-2600 3.40GHz | SSDT from Sam's ssdtPRGen.sh | Radeon 5770 1GB | 4GB DDR3
Booting with either Clover 2716 | Chameleon r2375 | RevoBoot v1.5.40 | Ozmosis 0894M
dmazar
post Apr 12 2012, 09:58 AM
Post #1004
Some presentation on Win UEFI booting:
http://www.uefi.org/events/UEFI-Plugfest-W...Environment.pdf

Again: loader sets kernel virtual address space and then calls SetVirtualAddressMap().
HW: Asus P8P67-M, Intel Core i5-2300, 4GB, XFX HD-567X-ZHH3 SW: SL, L, ML: Clover UEFI boot
crazybirdy
post Apr 12 2012, 10:13 AM
Post #1005
QUOTE (Slice @ Apr 11 2012, 01:35 PM) *
New DarwinDumper works flawlessly. What may be added here?
1. DeviceTree
ioreg -lxw0 -pIODeviceTree >ioregDT.txt
2. VideoBios dump?

Really I have some test version of boot1h2 so forget it. I will reinstall to actual version boot1h2.

@crazybirdy
Can you explain me five different installation of Clover by your installer? I am not understanding what is what.
What is boot0hfs? I have several partitions with FAT32, NTFS, HFS+. So why boot0hfs?


boot0 = always search the first active primary HFS partition on disk1 only.
boot0hfs = always search the first primary HFS partition on disk1 only (not need to be active, it could keep active on win partition for sleep issue).
boot0md = will search the first active primary HFS partition works like boot0 but could search for disk1 or disk2....(as I know.)
boot0md works fine with boot1f32alt after some test here, boot0+boot0hfs not work with boot1f32alt here.
my pkg will install boot1f32alt with boot0md under process.

This post has been edited by crazybirdy: Apr 12 2012, 11:08 AM
STLVNUB
post Apr 13 2012, 07:47 AM
Post #1006
New CloverGrowerV0.5
Uses 'new' efi folder structure in Clover.
Builds 32/64 bit Clover and package as well dry.gif
[attachment=4271:CloverGrowerV0.5.zip]

XCode command line tools must be installed.
Builds on SnowLeopard, Lion, MountainLion using gcc4.63

edit: please replace the CloverGrower.sh with this one.
[attachment=4274:CloverGrower.sh.zip]
Slight bug in the svn update.


This post has been edited by STLVNUB: Apr 13 2012, 05:30 PM
Azrock Z77 Pro3, G1610 AMD HD7750 Clover/Ozmosis
Toshiba A660-0MR00R I7-7740QM GT330M 8 Gig Clover

blackosx
post Apr 13 2012, 04:56 PM
Post #1007
Just tested CloverGrowerV0.5 on my iMac running 10.6.8 with Xcode 3.2.3 installed.
It ran faultlessly leaving me with a 2.23GB folder, 32/64bit complied builds and the package too.
Good Job STLVNUB - that was effortless.

I've just run the package installer to a fresh MBR/FAT32 USB flash memory drive and will test booting from it later this evening.
For reference, at first glance I thought the installer's 'Clover EFI' options looked a little confusing with the different option. How about just offering to install the following for the Clover EFI option?
- boot0md (the latest version that's in the Chameleon trunk)
- boot1h2 / boot1f32alt (from Clover)
- Then maybe the 32/64bit boot files and the order required as it's currently offered.
10.9.4 | Asus Maximus IV Gene-Z (PMpatched BIOS 3603) | i7-2600 3.40GHz | SSDT from Sam's ssdtPRGen.sh | Radeon 5770 1GB | 4GB DDR3
Booting with either Clover 2716 | Chameleon r2375 | RevoBoot v1.5.40 | Ozmosis 0894M
STLVNUB
post Apr 13 2012, 05:28 PM
Post #1008
QUOTE (blackosx @ Apr 14 2012, 02:56 AM) *
Just tested CloverGrowerV0.5 on my iMac running 10.6.8 with Xcode 3.2.3 installed.
It ran faultlessly leaving me with a 2.23GB folder, 32/64bit complied builds and the package too.
Good Job STLVNUB - that was effortless.

I've just run the package installer to a fresh MBR/FAT32 USB flash memory drive and will test booting from it later this evening.
For reference, at first glance I thought the installer's 'Clover EFI' options looked a little confusing with the different option. How about just offering to install the following for the Clover EFI option?
- boot0md (the latest version that's in the Chameleon trunk)
- boot1h2 / boot1f32alt (from Clover)
- Then maybe the 32/64bit boot files and the order required as it's currently offered.

Thanks blackosx, there is slight problem in svn update Clover, I'm replacing build.sh in Clover folder, need to change
Line 157 in script to cp -R "${WORKDIR}"/CopyToClover/cbuildClover.sh "${cloverDIR}"/

You will need to delete Clover/.patchdone to get it to re-copy
The options where from the build as is, I haven't touched that part,yet.
Will take a look.
Only thing I DON'T like about the install, is why 2 boot64?? ( boot and boot6), need to replace the booth with my version, press 'c' for Chameleon (boot) or press '3' for Clover32 (boot3) or default boot6.
To me it seems logical to leave boot as Chameleon etc.

I'll have a look at the package making as see what gives.

edit;
@blackosx/Slice
Would like to include my boot1altV2 in the build
[attachment=4278:boot1altV2.zip]
This boots Chameleon (boot) if 'c' pressed OR boot3 if '3' pressed, any other key OR timeout boot6.

I think this is the way it should be.
Installer should be able to compile the source so if user wants default as 32 then swap in source code NOT swap as file.
Seems silly to me to have two same files with different names on HD.

This post has been edited by STLVNUB: Jul 5 2012, 09:32 PM
Azrock Z77 Pro3, G1610 AMD HD7750 Clover/Ozmosis
Toshiba A660-0MR00R I7-7740QM GT330M 8 Gig Clover

dmazar
post Apr 13 2012, 10:20 PM
Post #1009
Hi guys. An update on booting on AMI Aptio UEFI board ...

Managed to add virtual mappings for run time areas and SetVirtualAddressMap works now!!!

Kernel loads, no KP Yeeaaaahhhh!!! tongue.gif
but stops on waiting for DSMOS - so, desktop is not reached yet. sad.gif

Anyway, managed to start kernel tongue.gif

Current, very ugly, code attached if somebody will be interested. Will clean it in next few days. This is just some kind of proof of concept of how some issued can be resolved. And it is with a lot of limitations: 64bit code only, Microsoft compile only, Lion kernel with 0x02b8000 as entry point, only kernelcache, requires entering of GraphicsOutputProtocol data into BootArgs.c, no memory protection so it's possible that the code will overwrite itself with kernel on some boards ...

But it does: stack relocation, fooling boot.efi by returning available memory instead of requested, patching kernel entry with jump back to code to be able to continue when boot.efi finishes and then fixing after it: fixing boot args, some values in device tree, runtime services with virtual mem mapping and jumping back to kernel.

I am starting the thing from the shell with "uefi.nsh", then "load OsxRelocationOverride.efi" and then "boot.nsh".
Attached File  UefiBootTest.zip ( 543.23K ) Number of downloads: 23
HW: Asus P8P67-M, Intel Core i5-2300, 4GB, XFX HD-567X-ZHH3 SW: SL, L, ML: Clover UEFI boot
dmazar
post Apr 13 2012, 10:35 PM
Post #1010
Kernel log from this UEFI boot attached. Will appreciate if somebody will be able to give some advice what should be fixed here.
Attached File  UEFIFirstBootLog.txt.zip ( 2.36K ) Number of downloads: 13


Note: no ACPI patching here and SMBIOS probably wrong.
HW: Asus P8P67-M, Intel Core i5-2300, 4GB, XFX HD-567X-ZHH3 SW: SL, L, ML: Clover UEFI boot
THe KiNG
post Apr 13 2012, 10:54 PM
Post #1011
QUOTE (dmazar @ Apr 14 2012, 01:35 AM) *
Will appreciate if somebody will be able to give some advice what should be fixed here.

Hope I can give you some help, my APTIO board should arrive soon, then I can play.
Thanks for the development an for using Windows on this...
Slice
post Apr 14 2012, 06:21 AM
Post #1012
Hi dmazar, THeKiNG and all interesting.
I think this thread became for compilations methods and it will be better if you started new development thread devoted to UEFI boot.
I will be glad if some of my Clover sources will be useful for the new project.
May be in future I also will have a board with UEFI BIOS.
Пожалуйста, прочитайте ЧаВо!
i3-2120 GA-H61M-S1, Radeon HD6670, ALC887(VoodooHDA 2.8.4), OS⌘10.9.2, OS⌘ 10.7.5 Clover FakeSMC_plugins_3.3.1 Realtek LAN v3.1.2
dmazar
post Apr 14 2012, 08:42 AM
Post #1013
QUOTE (Slice @ Apr 14 2012, 08:21 AM) *
Hi dmazar, THeKiNG and all interesting.
I think this thread became for compilations methods and it will be better if you started new development thread devoted to UEFI boot.
I will be glad if some of my Clover sources will be useful for the new project.
May be in future I also will have a board with UEFI BIOS.

You are right. I'll open a new thread next week.

I hope you'll get the UEFI board soon - I have a selfish reason: you'll be able then to speed this UEFI booting thing up tongue.gif . The difference in using DUET and some UEFI is like running the code in some simulator (slower) and running native (faster). And I hope in the future we will not use some of your sources, but full Clover (UEFI part). We just need a separate patch for UEFI board (what I am trying to do) and modifications to Clover to work on the board's UEFI. Writing some separate bootloader is not an option to me - this requires a lot of time (as you know from your experience) and would be complete waste of time when UEFI booter already exists.

THe KiNG: Win compile is for historical reasons - I did not have OSX 64bit tools at the time I was working on the asm code for 32-64-32 bit switch. And writing that code in MS 64bit assembler is a nightmare because it can produce only 64bit code. blink.gif
HW: Asus P8P67-M, Intel Core i5-2300, 4GB, XFX HD-567X-ZHH3 SW: SL, L, ML: Clover UEFI boot
STLVNUB
post Apr 16 2012, 06:13 AM
Post #1014
QUOTE (dmazar @ Apr 14 2012, 06:42 PM) *
THe KiNG: Win compile is for historical reasons - I did not have OSX 64bit tools at the time I was working on the asm code for 32-64-32 bit switch. And writing that code in MS 64bit assembler is a nightmare because it can produce only 64bit code. blink.gif

Try Fasm, their is an OS X version in the Tools folder in CloverGrowerV0.5, I can also give you efi version as well

This post has been edited by STLVNUB: Apr 16 2012, 06:13 AM
Azrock Z77 Pro3, G1610 AMD HD7750 Clover/Ozmosis
Toshiba A660-0MR00R I7-7740QM GT330M 8 Gig Clover

modbin
post Apr 16 2012, 05:14 PM
Post #1015
QUOTE (dmazar @ Apr 14 2012, 08:42 AM) *
You are right. I'll open a new thread next week.

We just need a separate patch for UEFI board (what I am trying to do) and modifications to Clover to work on the board's UEFI. Writing some separate bootloader is not an option to me - this requires a lot of time (as you know from your experience) and would be complete waste of time when UEFI booter already exists.


Hey there, on Insyde based UEFI firmware and on Intel UEFI Firmware one can just free the required memory pages. Then boot.efi will load up on the unmodified firmware. I use that on my DX58SO. I added a new boot option that loads RefiT and ebl.efi by iNDi to do the patches needed for OS X (dsdt, dev-props, extra kext...)
On aptio firmware I still think a PEI driver that allocates the memory early is the best option.

greets,
modbin
dmazar
post Apr 16 2012, 08:17 PM
Post #1016
QUOTE (modbin @ Apr 16 2012, 07:14 PM) *
Hey there, on Insyde based UEFI firmware and on Intel UEFI Firmware one can just free the required memory pages. Then boot.efi will load up on the unmodified firmware. I use that on my DX58SO. I added a new boot option that loads RefiT and ebl.efi by iNDi to do the patches needed for OS X (dsdt, dev-props, extra kext...)
On aptio firmware I still think a PEI driver that allocates the memory early is the best option.

I need some basic patches for testing. Can I use that ebl.efi? Where can I find it? Thanks in advance.
HW: Asus P8P67-M, Intel Core i5-2300, 4GB, XFX HD-567X-ZHH3 SW: SL, L, ML: Clover UEFI boot
Slice
post Apr 17 2012, 04:47 AM
Post #1017
Hi blackosx,
QUOTE (blackosx @ Apr 16 2012, 10:09 AM) *
Hi Slice


Why not?
The stage 0 boot code (be it boot/boot0hfs/boot0md) lives in the MBR of either an MBR or GPT/MBR hybrid disk.


So it has to be boot0hfs.


It works, just like boot0 does, except it looks for the boot1 code on the first HFS partition, rather than the first active partition.

I am not famous with boot0hfs. Don't know who is the author and what is the idea for its creation.
I see description provided by crasybirdy:
CODE
"boot0hfs_description" = "Use fdisk440 to install boot0hfs to the first 440 bits of MBR without destroying any existing Windows installation if you have one, then boot0hfs will search first HFS partition for boot1h2.
It will not setup selected partition to be active.";

And I have USB stick MBR with one partition FAT32. According to decription it can't boot. Nontheless it works.

QUOTE
If you didn't install it manually then maybe you accidentally installed it using CrazyBirdy's installer?

Dunno how.

Thus I checked it is really that boot0hfs is installed on my stick and there is no question for the dumper.
This is a question for the boot0hfs.
Who is the author? Where is the sources? What is the functions?
Пожалуйста, прочитайте ЧаВо!
i3-2120 GA-H61M-S1, Radeon HD6670, ALC887(VoodooHDA 2.8.4), OS⌘10.9.2, OS⌘ 10.7.5 Clover FakeSMC_plugins_3.3.1 Realtek LAN v3.1.2
blackosx
post Apr 17 2012, 06:12 AM
Post #1018
QUOTE (Slice @ Apr 17 2012, 05:47 AM) *
This is a question for the boot0hfs.
Who is the author? Where is the sources? What is the functions?

boot0hfs is boot0 but Windows Vista/7 friendly.
It can still find and load all stage 1 Chameleon loaders - boot1h, boot1he, boot1hp, boot1f32.
The 'hfs' in the name indicates the only difference is it looks for the stage 1 loader on the first HFS partition and not the active partition.

Zef made boot0hfs after the discussion/work in that topic.

However, personally I no longer use boot0hfs since dmazar modded the boot0 sources and created boot0md which pretty much serves as a universal stage 0 loader.
10.9.4 | Asus Maximus IV Gene-Z (PMpatched BIOS 3603) | i7-2600 3.40GHz | SSDT from Sam's ssdtPRGen.sh | Radeon 5770 1GB | 4GB DDR3
Booting with either Clover 2716 | Chameleon r2375 | RevoBoot v1.5.40 | Ozmosis 0894M
blackosx
post Apr 17 2012, 06:17 AM
Post #1019
QUOTE (THe KiNG @ Apr 16 2012, 08:31 PM) *
Now a personal request for DarwinDumper:
On lspci to make more detailed and tree:
$lspci -nnvvbxxxx > $SYSinfoDir/lspci.txt
$lspci -nnvvt > $SYSinfoDir/lspcitree.txt

And include ioreg.app for full ioreg dump.

Thanks.

Shouldn't be a problem. Either I'll add it when I post the revised one I'm working on here for detecting stage 2 loaders, or maybe Trauma can add it with the other additions he's going to do?
10.9.4 | Asus Maximus IV Gene-Z (PMpatched BIOS 3603) | i7-2600 3.40GHz | SSDT from Sam's ssdtPRGen.sh | Radeon 5770 1GB | 4GB DDR3
Booting with either Clover 2716 | Chameleon r2375 | RevoBoot v1.5.40 | Ozmosis 0894M
crazybirdy
post Apr 17 2012, 07:00 AM
Post #1020
QUOTE (blackosx @ Apr 17 2012, 06:12 AM) *
boot0hfs is boot0 but Windows Vista/7 friendly.
It can still find and load all stage 1 Chameleon loaders - boot1h, boot1he, boot1hp, boot1f32.
The 'hfs' in the name indicates the only difference is it looks for the stage 1 loader on the first HFS partition and not the active partition.

Zef made boot0hfs after the discussion/work in that topic.

However, personally I no longer use boot0hfs since dmazar modded the boot0 sources and created boot0md which pretty much serves as a universal stage 0 loader.


I only use boot0hfs myself for a long time, that installer was created to be used myself, not for public purpose. tongue.gif
boot0md, it looks much better than boot0 and boot0hfs, maybe I should change to use the standard + EFI scripts of chameleon r1499 to install Clover.

67 Pages V  « < 49 50 51 52 53 > » 
Add Reply Start a new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members: