Jump to content

-Archive- Macefix86 2006 -Archive-


bofors
 Share

443 posts in this topic

Recommended Posts

no, not till we say so lol, no point buying something incompatible and its a G965 btw not a P965

 

although my DP965LT lists the DG965RY on the motherboard right by the cpu slot, which makes me think that its identical except for the onboard video. I also compared the specs on the intel site and they are the same except for the GMA3000! so here's hoping :)

Link to comment
Share on other sites

"i have to attach a flash chip on my nforce4 with some glue first biggrin.gif"

I think it could be a real story in near future (not a joke anymore). As I have said on IRC: I have found a board that have EFI activated from ECS (a 865G one) (the 945 will appear on their site soon) which have Insysde FW inside.

There is non-EFI version. I have check the board images from their site, they are almost identical, exept the flash chip socket is empty (so I guess that they just change the flashchip).

To Kiko: I have offlined for so long, feel sorry to miss interesting news like iflash patched or attempt to boot up 915 with hacked FW.

Link to comment
Share on other sites

Hi, I've written a little python script (thank you intel/tiano for the python port!) that searches the output of dh -v for ImageBase/ImageSize and then outputs a .nsh that uses dmem to dump the entries to disk, then converts the canonical hex output to binary grabbing the names from PdbFileName. So far I've come across 106 .efi files in my macbook pro's EFI env. Of most note are probably these:

 

abomination:~/src/efirip/out anomaly$ ls *Legacy*

AppleLegacyStartup.efi Legacy8259.efi LegacyBootFS.efi LegacyMetronome.efi

 

However a couple others look like they may be pertinant to the osx86 project.

 

(Is it possible the CSM consists solely of these files or would portions of it still reside elsewhere? I've dumped the 'bios' and 'vbios' from inside bootcamp but can't find anything matching them in either the dumped .efi's or a complete firmware image dump. Also bootcamp appears to set the first 'boot option' entry to point to the start of the firmware with the filename of '?', for example using bcfg to list boot options has "01. MemMap(11:FFE00000-FFF9FFFF)/? "Mac OS X" OPT" (yes, it says osx in there when booting xp, swaps the memmap reference for a partition reference when booting osx), how does this tell EFI to init the CSM?)

 

I'll attach a full bootcamp'd dh listing aswell as a file listing of the dumped efis. Let me know if anyone wants these files to look at, or the script itself (real messy and jury rigged right now but I'll clean it up if people want it). Also, does anyone know how to start these up manually? AppleLegacyStartup appears to be a driver but looks for a LegacyBoot handle that doesn't appear to exist on any devices... can you force this protocol to appear bound to a device even when it's not?

list1.txt

dh.txt

Link to comment
Share on other sites

I want to get into the record a slightly different twist to this, which I have been discussing with some people on the IRC channel this week.

 

Basically, Chapter 3 of EFI Spec. (link above), implies that the native EFI has a "boot order list" for EFI images (like a BIOS has a boot order for partitions or drive volumes). This means that we can instruct the native EFI to "boot" an EFI image other than the firmware flashed onto the EEPROM by inserting (or otherwise forcing the selection of) a new "boot" option. This means we do not have worry about stoping ExitBootServices() from being called in the firmware nor flashing such a modified version of the firmware.

 

The "boot order list" is an EFI global NVRAM variable. Chapter 6 details OS runtime methods which are normally availible to get and set EFI NVRAM variables like the "boot order list" by using the EFI "pointer" we have been discussing above. However, the question now is how do we access the native EFI NVRAM since it appears that the methods detailed may not work for BIOS booted OS's.

 

We may have access to EFI NVRAM if it is mapped into "normal" memory somehow, we also may be able to look at the implementations of the get and set methods in Chapter 6 in the OS X kernel and add a kext to bring that functionality back into BIOS OSx86, or we may be able to use something like the non-native EFI Shell to access native EFI NVRAM.

Link to comment
Share on other sites

  • 2 weeks later...
Guest
This topic is now closed to further replies.
 Share

×
×
  • Create New...