(UEFI) Dell XPS 15z L511z modded BIOS - and HOWTO

[nebster]

@TimeWalker
Hmmm, that is really weird...
I've just tried patching it and it patched it fine. (Andy's tool doesn't complain and still lists the modules as there)
Can you send me a patched version which it generates that is wrong?

[TimeWalker]

  BIOS1_A13_MOD.rar (Size: 1.92 MB / Downloads: 12)

@TimeWalker
Hmmm, that is really weird...
I've just tried patching it and it patched it fine. (Andy's tool doesn't complain and still lists the modules as there)
Can you send me a patched version which it generates that is wrong?

Maybe I'm just doing something wrong .. but it won't for me..

>main.release.exe BIOS1_A13_ORIG.WPH BIOS1_A13_MOD.WPH patchset.bin

WARNING: Exception caught (harmless): Enforcement failed (Line: 570)
WARNING: Unknown type 2: 00504624-8A59-4EEB-BD0F6B36E96128E0
WARNING: Exception caught (harmless): Enforcement failed (Line: 570)
OS X Power Management - Patching PowerManagement2.efi...
OS X Power Management - Done
Unlock Advanced Tiano Setup - Patching PlatformSetupAdvancedDxe.efi...
Unlock Advanced Tiano Setup - Done
Remove Tiano 'Reserved' String - Patching PlatformSetupAdvancedDxe.efi...
Remove Tiano 'Reserved' String - Done

Ok, now I get why yours opens with AnyP's tool .. you probably using Version 2.11 right?
Well, try with version 2.02 (which is more stable and has proven to be producing 100% working output when doing mods .. read @kasar's post above to see what I'm talking about).
The version 2.02 won't open the modded bios due to obvious chechksum issues. It opens up pathced bioses from PMpatcher from @CodeRush fine though.

2.11 output (notice how no compression is used because checksum was incorrect):

The DUMP folder contains only the PADDING-00000098_0_4.ROM 4.5 Mb
How in the world are pubkey and marker found in the padding rom ? FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
2.02 will say the following:

And the DUMP folder will contain the padding as well, but this time it's PADDING-00000000_0_3.ROM 1Kb

[nebster]

Hmmm, that's really weird...
My output is ever so slightly different to yours.
Here's the diff from my version to yours:

Code:

- 00000a0 9c b7 af a6 3d cc ee c9 db 40 01 40 18 10 42 f8
+ 00000a0 9c b7 af a6 3d cc ee c9 b1 4d 01 40 18 10 42 f8
- 0180100 b2 eb 48 8b 23 65 70 22 bc 3f 0b 40 6c 80 19 f8
+ 0180100 b2 eb 48 8b 23 65 70 22 af 3f 0b 40 6c 80 19 f8

With mine, it works perfectly for some reason but yours says no compression! (with 2.11)
Could you attach version 2.02 (or a link to it) since I can't seem to find it floating around the net?

EDIT: Try main.debug.exe instead, just in case I screwed up somewhere!
EDIT2: Ok, I found the bug... There's a bug with the experimental 64bit windows compiler for D. I've recompiled it with -O1 instead of -O2 for the release version
EDIT3: Uploaded to github

[TimeWalker]

Yup, thanks ! That was it, all fine now! +1
Applying a custom patchset is a cool and handy (but not really end-user-friendly hehe) feature.
V2.02 in the attachment.

[nebster]

Applying a custom patchset is a cool and handy (but not really end-user-friendly hehe) feature.

Yeah, I was thinking of extracting all of the EFI files (correctly named) and allow modifications to them. Then, when the modifications have been done, create a new WPH file and a patch file (text and compiled) to go with it!

[jkbuha]

Great work nebster! +1 rep!

Agreed with TimeWalker - this would be a really good feature to have - that way we could test upcoming patches quickly. I know I'd need this feature as I have a lot of patches I'd love to test at the moment, such as our DSDT hacks

[TimeWalker]

I'm actually thinking of swapping the model number of the touchpad, because Dell has "accidentally" forgot to include support for older 7.5 Synaptics in their Windows 8 drivers .. so I need to use something from GROUP2 or GROUP5 range..

Code:

Device (PS2M)
                {
                    Method (_HID, 0, NotSerialized)  // _HID: Hardware ID
                    {
                        If (LEqual (BRID, One))
                        {
                            Return (0xC5048C11)
                        }
                        Else
                        {
                            Return (0xD9048C11)
                        }
                    }

DLL0595 (G2 - 0x95058C11) or DLL05A3 (G5 - 0xA3058C11) to mimic the supported models and those allow for Disable_Enable_TP_By_HotKey to be executed, otherwise Fn+F3 does nothing.. because even though the scancode it expects (0xE01E) is emitted the driver lacks support for my 0xD9048C11 model ..

[nebster]

Ok, I've implemented an extractor and injector so that it should be easy to generate modded WPH files. It doesn't deal with modified file sizes since I haven't got any examples to test that with.
Uploaded to GitHub

[TimeWalker]

A basic howto when the program is ran without arguments wouldn't hurt .. I'm getting lost with all these CLI utils.
The scenario is the following:
- To avoid messing with GUIDs I run the extractor on a BIOS.WPH and it extracts all the modules by their respective names (also what about modules with no names .. like ACPI tables/firmwares/vbioses ?)
- I do necessary mods to these modules and run injector which assembles a patched BIOS.WPH back together ?

Well, extractor won't run fro me .. "Stack trace invalid for MinGW", it decompresses only DxeMain.efi and that's it. Tried both debug and release versions.

[nebster]

Sorry, I coded that way too quickly...
Did you know if you run the program manually it will ask for all command line parameters? I'll put in a message saying about this and checks for "/?", "-h" and "--help".

Scenario is correct. That is how it works (it doesn't work for files of different sizes yet as I don't have an example though)
Modules with no names currently won't be extracted as I didn't know whether every module would have a name or not!
I'll extract modules with no names to their GUID's.