Print

[mav]

Just a note, iirc GH,IJ,KL,MN are reserved for memory cards (controler 1 slot1=G controler 1 slot2=H etc...)
so if your going to use another letter there might be issues there
xyz are cache of course so maybe O->W. i might be wrong just thought i would through that in.

[heinrich]

QUOTE (heinrich @ Aug 29 2003, 12:24 AM)

Makes sense, and in some cases, it is needed to use an older bios in order to get around certain errors (like no system time), so i would think that the best way to do it would be to have that 1 extra partition for everything above the 137gig mark.  Since I will admit right now that i have NO clue if this is do-able (to have C E F Y Z, then F from the 8 gig mark to 137, and a G from 137 and on), I will leave it to you, but if sing an older bios later on would currupt data, i can see this as being the only feasible way to do it.

Another time when you would need to extra F drive, but couldnt use past the 137gig mark is any debug bios.  While i suppose that >137 support could be put into lets say... the TATX bios, or just 5558.  Its great that it *works* but implementation i think will take a bit of work, and you will need the help and support of other homebrew devs.

[fixxxer]

WAY TO GO sweetness. cant wait for a release of this biso. so i can mess around my self  

[oz_paulb]

QUOTE (mav @ Aug 29 2003, 05:35 AM)

Just a note, iirc GH,IJ,KL,MN are reserved for memory cards (controler 1 slot1=G controler 1 slot2=H etc...) so if your going to use another letter there might be issues there xyz are cache of course so maybe O->W. i might be wrong just thought i would through that in.

The partition table in the bios will only assign a partition number (7, for example) - not a drive letter.  It's up to applications (like EvoX dash) to decide what partition should be mapped to what driie letter.

C/D/E have 'standard' assignments in Xbox, F: was added to map to the hard drive's "partition6", and has become a new 'standard' drive letter.  If more partitions are created in the future, new 'standard' drive letters will probably emerge.

- Paulb

[oz_paulb]

QUOTE (Flagg3 @ Aug 28 2003, 11:59 PM)

You asked about the source for Slayers Auto install.  He is no longer actively updating it but that's not what you'd need either way.  If you look at it, all it is is a very cleverly written evox.ini.  The auto installer doesn't actually add any code to evox, it just takes advantage of the ability to heavily customize Evox.  If you look at the included evox.ini, you'll see that the auto-installer simply calls the format functions from evox to prepare a new drive.  So basically, if your having a problem with formatting all of the partitions in Slayer's Auto-install, you'd need the source for the evox dashboard to determine where the problem is.

I've figured out where the problem is with the Slayer EvoX installer (compatibility with my BIOS).

It's nothing to do with partition tables (I thought it was this before).  It's the fact that I've got LBA48 support in the BIOS (if I disable my LBA48 'hooks', everything works OK).

The EvoX.ini entry "ConfigSector" is the culprit.  If I remove that line, the rest of the format of all other partitions works fine (with LBA48 support turned ON).

Does anyone know what happens on the ConfigSector command?  Is that code built-into EvoX, or is it executing an external app (like ConfigMagic) to do the work?

I noticed that the source to much of ConfigMagic is available online at team-assembly.com - maybe some of that has been put into EvoX.  The source code does direct access to IDE command/status registers.  I can't think why this would be a problem (if they're doing all of their IO separate from the KERNEL, and not trying to mix the two, for example).

I know that when you issue an LBA48 command, the subsequent 'status' that you get back is in a different format than for LBA28 commands (because the status registers need to show you what LBA address failed, and have to be modified for LBA48 in a way like the commands).  But, if you issue LBA28 commands, you get LBA28 status results.

I'd like to understand what's going on here before I release the BIOS mod to the world - I don't want to find that there are lots of programs doing direct IDE access that will start crashing (I think it's probably only utilities - not games/etc).  I don't think there's anything I can do in the KERNEL to fix this - I think some of these utilities may need to be re-written with the knowledge that LBA48 may be in effect.  I'm not sure, though - as I said, I don't really understand why it fails.

Can anyone confirm that the "ConfigSector" command is based on the code @ team-assembly.com?  I'm guessing that it is, just because it looks like it makes use of a "disk.bin" file that was the result of a ConfigMagic "backup".  But, they could be unrelated.  (come to think of it, I've never used "ConfigMagic", and I'm not sure it's the tool that creates the "backup" stuff that I've seen reference to before - so I could be totally confused here (it is getting late)).

Well, I'm about to head to bed, and have to get up early for work in the morning.  Since I'll be @ work all day tomorrow, I may not have time to check in to the forums too often.

- Paulb

[fixxxer]

sleep well man i think you just made some of the biggest new yet in the xbox scene.

[starwarznut]

I just wanted to say that I've been following the thread since I got home from work today (didn't get a chance to log in at work today....crazy ass week...) and wanted to say thank you, and good job on this.  This is big news and we all appreciate your efforts.

[focusracer1]

QUOTE

Sorry to write here, paul but i hear this discussion now since weeks. and it is false. First, you should tell all people that if you enable lba48 (which is simple) you can not access the data with an normal kernel anymore From a Windows Support Site >If you enable 48-bit LBA support on a system that does NOT have a 48-bit LBA- >compatible BIOS, and a hard disk larger than 137 GB, you will experience data >corruption. but as people sure are "more intelligent" as i, please consult ATA/ATAPI 6.0 specification. It is written very clear there, why the "using in both worlds" is not possible. Good luck for your non-working & incompatible system . franz

Well, at least we know Franz has his PHD (Playa Hater Degree)

BooyaKasha!!!!!

[feliperal]

pualdb, first I would like to commend you for all the hard work you are putting on this project.

I'm very interested in studying the work you have been doing.  Would it be possible if you could document or explain in more depth how you were able to do it?  I would also be interested in viewing the kernel modifications and some of the source code.

If it is so troublesome, can you perhaps give us a quick overview of what you did?  I would like to reproduce these steps and document it, giving credit where it belongs. Perhaps, this can spur new ways to circumvent this limitation for soft mods, also.

Thank Again for all your work.

[greengiant]

oz_paulb I was aked by heinrich to reply to your question as I had not seen it.

I wrote the NewHDEvox installer, witch was used for quite a while. The foundation where Slayers came from.


The backup you talk about has NOTHING to do with ConfigMagic.
Backup is done by EvolutionX for quite a few builds now, it creates a few files in
Cbackup
EEPROM .bin file and HD partition/other .bin file
as well as the HD key and a cpl other files.

In my installer I used the drive.bin file because at the time evox did not have any INI commands for formating the drive. The ConfigSector command was the preliminary way of doing it, using raw data from the bin file.
Later on the builds evox came with built in INI commands that would format the proper partitions.

Since your now using LBA48 the paritition table and a few other things will be much different that when the drive is accessed in LBA28.
The disk.bin was created using standard LBA28 commands and thus the data will be in a format that will probably not be compatible with LBA48.
I donot know much about the ATA standard so please bear with me.

I beleive, not for sure, the format drive commands use a much less 'raw' way of formatting and setting up the partitions than if you where you use a disk.bin file and do it the raw way.

Let me know if this helps you.



And many apps, like evox and utilities do access the hd drive and many other xbox functions directly via memory calls and not via the kernel, because of certain needs.
Some functions within those apps do make direct calls, and others do kernel calls.
So it may be nessary to provide a module for developers to incorporate into their programs to make them compatible with your new work.

[neomaytricks]

this is the best development for the xbox this year.  If not in the gaming community. If there is anything we can I imagine some of the XS folks are willing to help continue with your work.  

[xVanWildeRx]

wow. i am absolutely amazed. rock on paulb!

[SigTom]

Had some time while im setting up my Clark Connect Router( dis some cool shit) and popped in here and read the updates on teh forum.  GREAT JOB Paul.  Glad you got it to write and verify.  Hope you get all the partitioning wokring and the ability to access it 100% of the time.  Glad to see that we have more technical people adding in and helping you out with some stuff.  Heres another "Big Up, Much Respect My Hommie G-Dog".  Gotta work the new job tomorrow, so Ill check up in the evening and see whats going on.  Good luck man.

[oz_paulb]

QUOTE (greengiant @ Aug 29 2003, 07:45 AM)

The backup you talk about has NOTHING to do with ConfigMagic. Backup is done by EvolutionX for quite a few builds now, it creates a few files in Cbackup EEPROM .bin file and HD partition/other .bin file as well as the HD key and a cpl other files. In my installer I used the drive.bin file because at the time evox did not have any INI commands for formating the drive. The ConfigSector command was the preliminary way of doing it, using raw data from the bin file. Later on the builds evox came with built in INI commands that would format the proper partitions. Since your now using LBA48 the paritition table and a few other things will be much different that when the drive is accessed in LBA28. The disk.bin was created using standard LBA28 commands and thus the data will be in a format that will probably not be compatible with LBA48. I donot know much about the ATA standard so please bear with me. I beleive, not for sure, the format drive commands use a much less 'raw' way of formatting and setting up the partitions than if you where you use a disk.bin file and do it the raw way.

I don't think the 'hang' problem I'm seeing has anything to do with the fact that the "disk.bin" file is being written, or anything to do with the contents of "disk.bin".  I think it has to do with the method of writing the file to the hard drive's sectors.

If the "ConfigSector" command were to use KERNEL methods to read/write sectors, then I think it would work.  But, I'm guessing it's doing direct ATA/IDE commands, and that there is some sort of interaction between those (LBA28) commands and the KERNEL's LBA48.

Where is the underlying code for the "ConfigSector" evox.ini command?  Is it part of the EvoX executable (built-in to EvoX), or does it launch another standalone app that does the work?

Does the "ConfigSector" stuff do direct ATA/IDE access (direct to the I/O ports), or does it make use of KERNEL functions for reading/writing to 'raw' hard drive sectors?

When I saw 'KERNEL functions', I mean using functions like NtOpenFile/NtReadFile/NtWriteFile on the 'file' called "\Device\Harddisk0\partition0" (partition "0" is for raw access to the entire drive).

I'd really like to get to the bottom of this, to make sure I haven't missed some part of the KERNEL code that I need to update for LBA48.

Thanks,

- Paulb

[oz_paulb]

QUOTE (feliperal @ Aug 29 2003, 07:40 AM)

pualdb, first I would like to commend you for all the hard work you are putting on this project. I'm very interested in studying the work you have been doing.  Would it be possible if you could document or explain in more depth how you were able to do it?  I would also be interested in viewing the kernel modifications and some of the source code. If it is so troublesome, can you perhaps give us a quick overview of what you did?  I would like to reproduce these steps and document it, giving credit where it belongs. Perhaps, this can spur new ways to circumvent this limitation for soft mods, also. Thank Again for all your work.

I plan on publishing everything that I did, including source code to the new code I've added.  I won't be publishing a complete BIOS 'binary' (for legal concerns), but will release a method of patching an existing BIOS (hopefully via Xbtool).

The actual code I've added is very small - the hardest part was coming up with a method to 'expand' the KERNEL to make room for my patches, then tracking down all of the locations in the KERNEL that I needed to patch (finding all I/O instructions to ATA/IDE ports, deciding which ones needed to be converted to LBA48).

- Paulb