unofficial LaunchELF v4.42

[zabolyx]

beautiful... just beautiful.... This is great... I can finally do everything that I need to my HDD with one app... no more resetting every five minutes while installing the expliot and HDD in someone elses PS2...

any chance that it could handle basic operations to the HDL partitions? such as renaming and deleting? I know, but using the keyboard to rename games after install would be nice....

Now we need a forth programmer to jump in and help include the text/hex editor into the app.... nothing fancy (no need for a clipboard or anything) just backspace, delete, position cursor, write, and insert..... already have the keyboard support....


@Dlanor

What about the error with the 'do you wish to overwrite?' dialog.... has that been corrected to only display the folder name?.... I think it should show both... foldername [spacing] save name.... make it a little easier to make sure what you are overwritting.....


@Polo35

what about shrinking a partition? has that thought been chewed on yet? And a possible defragmenter... although it seems clear that two different types of defragmenting would need to be implemented.... one for files withing a partition... and one for the scattered subpartitions.... is it an idea that is worth tackling....

[Jones23]

Find a new Bug in 3.71 (not a real bug, but you´ll see, what I mean ).

This "Quick-Jump to Path"-Function (<- Was it "PathPad-String"-Function?), where I can define a Path in CNF, has 1 Bug.

When I tipe in a path in the CNF + the ELF-Name it will handle the ELF as a folder.

Maybe that can cause in problems. :-/

What the heck are talking about?
You said it yourself: It's a "Jump to Path"-function! Not a "Launch ELF"-function. So it will only jump to a Folder.

About the PowerOff-Function-Bug I have.

I started UL-ELF through compressed modified (<- Removed Slash-Bug) DMS-MC-Explorer, from a 32 MAX-MC.

Maybe the MC is the Problem.

I´ll try to reproduce it today.

And will test it again with 3.72.

Quit using those crappy MAX MCs. They're the worst thing you can do to your PS2. The official $ony MC is the only reliable one. Get them here for 18$ and free shipping worldwide!

I don´t know, if there is still the bug, when I browse a device (like MC), when FTP reads/writes to it, that one of the 2 apps (UL-ELF, or the FTP-Client) will crash and maybe data-corruption occurs. <- If yes, maybe it could prevented, just by an easy if call function, that ask´s the FTP-Module (<- if it can handle this), if it uses the device.

If yes, you couldn´t access the device and a message will pop up like "This device is currently used by...", or something like that.
If no, you could access the device.

The whole FTP-thing isn't very stable. Better use host and ps2client or RHC.
Read my post here for better explanation.

And please standartize "Swap Buttons" as on.

I hate swapping it @first configuration.

Why don't you simply keep your old config, when updating to a new version?
Then you don't have to configure it all over again every time.

regards Jones23

[bootsector]

Hi Polo!

Good news you have fixed the problems with the format function from libhdd in the current version of PS2SDK.

Do you have write access to the PS2SDK SVN? If so, could you please fix the sources for us?

This way, I can successfuly close my thread at forums.ps2dev.org:

http://forums.ps2dev.org/viewtopic.php?t=5457

Best regards,

bootsector

HEY, HEY, HEY,


You're definitly not the one who can tell me something, because you don't have any hdd in your ps2. ( It's a joke. lol )


There is no problem, because ULE use allready patched modules for 48bits hdd.

So, if your hdd is well reconize by ULE for now, everything will work fine.


No problem for me to add some new setting for graphical colors, but i think dlanor is the master for that sort of thing.


So, i fix format problem with the line i mentioned before:
http://psx-scene.com/forums/showpost...postcount=1971 (unofficial LaunchELF v4.42)

Old lib is libhdd1.2 found at ps2dev.org site, and new lib is libhdd from ps2sdk.

The fix is to add to ps2sdk/iop/hdd/apa/src/hdd_fio.c file, in hddFormat fonction, in setup mbr section.

There is no more size report problem, just a cleanning, recompiling problem in ps2sdk, when i apply the fix.

So, now, hdd is reconize formated by $ony tool, and DM$ tool, without any problem.


Exactly, problem solve by creating a 2048Mb part ( if requested size is > to 2048Mb ) and expanding it to requested size.


No link problem here, just a mistake from my part.

I forgot to add fileXioUmount pfs0:, pfs1: and fileXioDclose hddFs, after getting hdd info.

So, the last listed part was all time open, and that prevent, removing, renaming, expanding it.


I found an interesting line in ps2sdk/iop/hdd/apa/src/hdd_fio.c file ( same file i fix ) : // do the renameing / note: subs have no names!!

I've to make some test and to see with $ony tool which show all part and sub part...


In conclusion : ULE HDD MANAGER is almost completed, and work fine with 5 tested hdd, i just have to make some cosmetic changes for UI design, according to dlanor agreement ( colors )


Best regards

Polo

[spedmetal]

@spedmetal: He mean´s the Slash-Bug.

ahhhh ok thanks

[dlanor]

FTP note 2: There is an issue that I noticed where you place a file of 'x' size on the hdd. When you overwrite that file with a smaller one of the same name, the file is copied over the original but the original size remains. I did some hex compares and saw that it leaves the rest of the original file intact. In other words it doesn't terminate at the files end point as it should.

Ooops! That's a very bad error indeed when updating stuff...
But perhaps it can be fixed simply by adding "|O_TRUNC" somewhere ?

@dlanor in regards to fsclient.
It's known that you have reissue the "fsclient -h ps2 dir hdd0:" command in order for the partitions to show up. I saw that mentioned over a ps2dev forums.

Yes, I've seen that too, but it was never fully explained or investigated.

The whole partition sub-partitions listing thing is what ps2ftpd used to do until I changed it to ignore sub-partitions. Unfortunately in order to calculate the entire size of a partition you really need to add all the sub parts and the main partition sizes together.

Sure, but doing that properly could be a one-time investment. Then we just reuse the same algorithm where needed.

You can see all the chunks or segments and fragmentation better with hdl_dump's toc command.

Thanks I'll try that out too then.

Another issue with fsclient is that it doesn't calculate partitions properly because it uses 32 bit int so you end up with negative sizes for some large partitions.

Do you mean when using it to access PFS paths ? (Thus accessing the 'high-level' partition, so to speak)

Because for the "hdd0:" path, sign and truncation problems seem impossible, since large subpartitions simply don't exist, and for this path the sums for 'high-level' partitions are never shown.

I mentioned this to radad after some changes were made to the hdd driver. Of course, it's not such a big deal as it's only a listing. Although I too would much rather prefer a more accurate one.

Well, that is always best of course, but the only stuff I was interested in here was the subpartition usage, and those numbers never got close to the sign bit problem, so I think we can rely on them.

Best regards: dlanor

[Polo35]

When will it be released (a.k.a. "It will be released, when it´s bean ready." *lol*)? ^^

It will be released, when it´s bean ready. *lol*

So with this new hdd manager, all the stuff about how you shouldn't create partitions >4GB can be forgotten ?

Yep, exactly.

As well as manipulating partitions that had to be reduced to a minimum ?

Don't really understand what you want to mean ??? Sorry.

any chance that it could handle basic operations to the HDL partitions? such as renaming and deleting?

Not for the moment, sorry, i've to study a little bit more hdl partitions before thinking to manipulate them.

what about shrinking a partition? has that thought been chewed on yet?

About shrinking, i found a way to remove last sub-partition, so that decrease the size, but every sub-part have a fixed size so no way to remove a requested size. ( Or for example: remove last 4096M, and add 3072M, to remove only 1024M, but probably lost of file in partitions ).

And a possible defragmenter...

We have allready talk about defrag, and i maintain there is no real need of that sort of tool.

Good news you have fixed the problems with the format function from libhdd in the current version of PS2SDK.

That seem to work fine now, but i've to make some more test.

Do you have write access to the PS2SDK SVN? If so, could you please fix the sources for us?

No, i don't, and maybe it will never be fix in ps2sdk from svn.

Because ULE release a ps2sdk folder with fix for various thing.
So when you want to compile ULE, you copy this folder in ps2sdksrc and recompile completly ps2sdk before working on ULE.

But the fix is not a big deal, look some post before to fine the file and the line to fix.

Best regards

Polo

[bootsector]

Great Polo!!!!

Could you please post this solution for the hdd formatting in the ps2dev forums? This way I can finally close my thread there

Regards,

bootsector

[dlanor]

Find a new Bug in 3.71 (not a real bug, but you´ll see, what I mean ).

This "Quick-Jump to Path"-Function (<- Was it "PathPad-String"-Function?), where I can define a Path in CNF, has 1 Bug.

When I tipe in a path in the CNF + the ELF-Name it will handle the ELF as a folder.

You can abuse many of the CNF variables by editing them to contain garbage, but that doesn't mean the program is bugged (rather the user ). Problems that can only be caused by incorrect direct editing are not to be considered as program bugs.

The PathPad array is only intended for pure browsing paths, not for ELF launching, and that is why the user interface only allows you to set pure paths.

About the PowerOff-Function-Bug I have.

I started UL-ELF through compressed modified (<- Removed Slash-Bug) DMS-MC-Explorer, from a 32 MAX-MC.

Maybe the MC is the Problem.

Hardly. I use two of those myself, as my main MCs.

I´ll try to reproduce it today.

And will test it again with 3.72.

Good. Reports should always be for the latest version, whenever possible, so we avoid wasting time on old issues already fixed.

@dlanor: Hope you get the PM from EEUG, that I send to him instead of you or EP.

Yes, he did pass that on to me.

This is a little feature-list. Hope it would be addet also. ^^

So please move the Button-Configurations to it´s own Sub-Menu in Config.

Maybe eventually, but everyone is used to having it on the first CONFIG page, so we won't change that until we really need the space. OK?

As for your other wishes:
The 'PathPad-like' launch list you wanted may become implemented eventually (possibly in different form though), but not right now.

Integrating HDL_Dump server is not on our agenda at all.
It would open a huge 'can of worms' that we'd rather do without.

Your idea for auto-mounting multiple partitions for FTP would be too inconsistent with current methods. We prefer to keep that as it traditionally has been for all PS2 FTP servers, except for the simplest method of auto-mounting by browsing, since that is a very intuitive method, which can be used without conflict with the other current methods.

I don´t know, if there is still the bug, when I browse a device (like MC), when FTP reads/writes to it, that one of the 2 apps (UL-ELF, or the FTP-Client) will crash and maybe data-corruption occurs.

If you report bugs without even bothering to test if they exist, why do you expect me to bother with debugging them ?

Also, the fact that you can perform file operations while the server is also doing so does not mean that you should do it excessively.

The PS2 is a fairly simple system, as compared to most computers, so it doesn't have all the safeguards you might expect of a fullfledged operating system. Simultaneous access to the same devices by different programs (here FTP server and uLE) can cause conflicts, and those may very well crash the programs, and of course a crash during write operations may leave garbage in its wake.

This has nothing to do with any bugs in either program, but is rather due to a usage inappropriate for systems of this kind. We'll do what we can to fix such things, but in some cases that's simply not up to us. (The MC drivers are not homebrew, but Sony originals from the PS2 ROM.)

If yes, maybe it could prevented, just by an easy if call function, that ask´s the FTP-Module (<- if it can handle this), if it uses the device. If yes, you couldn´t access the device and a message will pop up like "This device is currently used by...", or something like that.
If no, you could access the device.

Communication between the main program and the FTP server is not as simple as you think, since they execute in different address spaces using different processors.

And please standartize "Swap Buttons" as on.

I hate swapping it @first configuration.

The default already is standardized, though clearly not to your liking.
(But the standard we use here is the original standard of the PS2.)

Surely you understand that if I were to remove this option, the mode that would then become mandatory is the original and default mode (which I also use myself). And then you couldn't even get the mode you prefer...
So I think we'd better leave this as-is. OK?

Best regards: dlanor

[E P]

Hmm MyPS2 works with Ubuntu and Mac Samba, I do not have Windows...

Good then so if it works with Ubuntu than there is a good chance it will work with Slax for me then. I just got so much to learn about the networking side of linux. It's going to take some time.

Ooops! That's a very bad error indeed when updating stuff...
But perhaps it can be fixed simply by adding "|O_TRUNC" somewhere ?

It only appears to occur with the hdd. Never experience this with the memory card or even usb mass storage that I can remember. Most of the time files are overwritten with files that are larger than the original. It's easy to see how something like this was overlooked for so long. (Note: host doesn't have this problem)

Do you mean when using it to access PFS paths ? (Thus accessing the 'high-level' partition, so to speak)

No just the hdd0: listing. You're right though as size isn't real important here because what matters is the actual size when a given partition is mounted to PFS.

Well, that is always best of course, but the only stuff I was interested in here was the subpartition usage, and those numbers never got close to the sign bit problem, so I think we can rely on them.

Yeah, I understand. I still distinctly remember my 4096MB +NEW2 partition showing up as size 0 when it should show 4294967296. The size shows up as 0 in the hdd0: listing because it is using a signed 32 integer value instead of a 64 bit value. This was changed in ps2ftpd but we don't use that anymore in the FTP hdd/0 listing. However, it's still useful for large files of sizes equal to or greater than 4294967296 bytes.

[dlanor]

No problem for me to add some new setting for graphical colors, but i think dlanor is the master for that sort of thing.

Well, the current four colours are mainly intended for normal/selected text, frames, and backgrounds. The basic purpose of graphical representations like pie charts make the use of any of those colours inappropriate, as we want such a graph to 'stick out' as compared to the other material on screen.

So adding some new entries to the colour array seems a better choice here, to be used exclusively for graphs of various kinds. At least two seem needed, for colouring the 'main body' of the pie, and a selected 'slice' of it. But with the space available in the color setting dialog, we can just as easily add up to four new colours, which should suffice for more advanced needs too.

So, i fix format problem with the line i mentioned before:
-----
[There is no more size report problem, just a cleanning, recompiling problem in ps2sdk, when i apply the fix.

I'm not sure exactly what you meant here.
Is there a problem with "make clean" for PS2SDK ?

Either way, I assume it's easy to get around, since you don't seem worried about it

So, now, hdd is reconize formated by $ony tool, and DM$ tool, without any problem.

Excellent. Seems like you're getting close to something that can be released.

Exactly, problem solve by creating a 2048Mb part ( if requested size is > to 2048Mb ) and expanding it to requested size.

I assume that even such a 2G partition will consist of 2 subpartitions. Right ?

I forgot to add fileXioUmount pfs0:, pfs1: and fileXioDclose hddFs, after getting hdd info.

So, the last listed part was all time open, and that prevent, removing, renaming, expanding it.

Ok then. One thing less to worry about.

note: subs have no names!!

I suspected as much but I wasn't sure as I haven't inspected the HDDs with any low-level tools. The ones I've used simply repeat the same name for each subpartition, even if it only exists in storage for the main entry.

I've to make some test and to see with $ony tool which show all part and sub part...

Unfortunately I don't have any working Sony tools for HDD, so I can't try this. The only utility disc I ever found refuses to boot, even with ATAD patching, and regardless of connected equipment (none, NA-only, NA+HDD).

In conclusion : ULE HDD MANAGER is almost completed, and work fine with 5 tested hdd,

That's great! I really look forward to trying it myself.

i just have to make some cosmetic changes for UI design, according to dlanor agreement ( colors )

IMO you can use up to four new colours as you please, and don't worry about how you specify them. Just use constants if that's easier for you. I'll fix the details and setup their config entries when merging your stuff in with my sources.

Best regards: dlanor