Pxtools for Linux : first version, please try it !

Hi all,

For those intrested in using their PX-712/716 drives under Linux and take advantage of the advanced check functions, here is the very first draft of the port of pxscan to Linux. Be aware this version is just a draft and has rough edges. The output is still text based, but I wrote little Python scripts to display the results in Gnuplot.
So maybe there were people who wanted to play with it during the week-end. For the moment, I have ported the PISum8/PIF checks and the TA check.


1) Download pxlinux-latest.tar.gz and decompress it somewhere.
cd pxlinux

2) make
You'll get three executable files (you need the c++ compiler and development libraries)
ta
pisum8
pif

3) Commands: Note: you need to be ROOT to send commands to the drive. Use sudo or su. I have to play with the setuid in the code in the future. There is no harmful code in the different programs but be aware that using programs under root is bad.


- PISum8

./pisum8 /dev/dvd [interval]
[interval] is equivalent to the accuracy under windows. By default, it is 1 (takes 30min to scan a DVD), the higher it is, the faster it is, because it swaps the test every [interval] blocks.

For a graphical output, use ./pisum8 /dev/dvd [interval] > pisumfile to save the result in the file pisumfile. Then use the python script:
./plotpisum8.py [-l] pisumfile
and then you'll get a graphical output. The X axis represents the number of blocks.
Note you need Python and Gnuplot installed. Usually they are available in many distros or even installed by default.


- PIF

./pif /dev/dvd [interval]
The same than pisum8.
For a graphical output, use ./pif /dev/dvd [interval] > piffile to save the result in the file piffile. Then use the python script:
./plotpif.py piffile
and then you'll get a graphical output.

-TA

./ta /dev/dvd [layers]
[layers] is the number of layers, 1 by default. Use 2 if you have a DVD+R DL (this will be automatised in the future).
For a graphical output, use ./ta /dev/dvd [interval] > tafile to save the result in the file tafile. Then use the python script:
./plotta.py [-l] tafile
and then you'll get a graphical output in this order : inner, middle, outer. If there is no -l switch, pits are displayed. If there is the -l switch, the land are displayed.
Note that the file tafile contains an analysis of the shift (cat tafile).

Finally, these are some pictures:
- PISum8

Here we see PoweRec in action at 80% of the length. Note the Y scale is logarithmic, so few errors are made more visible.

- PIF

No PIF higher than 2.

- TA pits and lands



OK, now the TODO:
- solve the root problem
- automatic support for DL
- tests for CD
- a real GUI
- a clean code
- the other stuff (jitter/beta, gigarec, varirec, powerec,...)

Also, the legal prints: In accordance with Alexander Noe, the code is under GPL. For people who don't know what it means:
- you can run, analyse, modify, copy, redistribute and even sell the code freely (this is FREE software)
- but any change in the code source MUST be made available FOR FREE (actually the cost of the media) to anyone receiving the binaries, and any derivative work has to be GPL. You cannot put the code in any proprietary application, the whole application has to be GPL.
In other terms: you don't pay with money, but with sharing the modifications you would do to this code. Everybody wins in that deal.

The code will be deposited to Sourceforge, but I thought some of you would like to try it this week-end Note also that the hyperlinks there won't be available after I put the code in Sourceforge. But you will know

Cheers all

---

Last edited by G@M3FR3@K; 18-02-2005 at 16:01.

Wow, this is amazing.

ET

amazing work....hats off to both Alex and Zeb

I will test it on fedora core 3 and redhat 9.

Hi,
new version: pxlinux-0.0.2.tar.gz

This is a minor correction of the ta test. The graph was shifted for the middle and outer peaks because of incorrect use of # symbol in the ta result data file.


Note to the moderator: I cannot edit my first post. Is it normal ? I wanted to update the link to the new version. Could it be possible to edit just once, and to replace pxlinux-0.0.1.tar.gz by pxlinux-latest.tar.gz ? And I will create a symbolic link of pxlinux-latest.tar.gz to the latest version. Many thanks

Great Job!
It will be easy for me to implement a gnome based frontend.
I will be very pleased to contribute to your job!

Hi,

thanks SlammerA
Next week I am putting the code on Sourceforge. This way, we will have CVS.

BTW, we will have to rethink the way data are saved. I had to use dirty things (like # symbols) to be able to use gnuplot. But if you write a GUI, you may need the data in a more formal format. We will see this later.

I have to disconnect until tuesday. I'll post again then.

Cheers

Since these features are in the Plextor software, has anybody given a thought to maybe the commands are proprietary? Is it legal for Zeb to post all this stuff?

it's a port of Alex Noe's pxScan and pxView which have been around for some time. to my knowledge, Alex hasn't had any "legality" issues. personally, i commend the two of them for their efforts and think Plextor should too.

I don't know if those commands are propriatary, and I don't care. Before I've used PxScan/PxView there, I've discussed about legal aspects with the one responsible for legal crap of that page, and we came to the conclusion that Plextor can't really own those commands, even if they want to.

BTW, according to what I know, Plextor Europe would actually like to cooperate, but Plextor Japan dreams that even the Gigarec commands were secret. Well, Plextor Europe also were the ones to refuse the delivery of broken drives (TLA#00, #01 of PX716)... seems like Plextor Europe cares about their customers, unlike other "outposts" of Plextor

If you ask me, those Q-Check functions are the only *really* interesting feature of Plextor drives besides the 6x DL write capability (there are enough 16x writers with good write quality out there). Plextor was incredibly stupid not to help free developers to create Q-Check tools for their drives for other operating systems. However stupid those guys in japan are, they just *have* to realize that we are doing their work for nothing!

Since the SCSI commands are transmitted in clear to the device, there is no big secret about them. Actually, what Alexander is doing, is just to understand how by turning a key he starts a car's engine (and this is already a quite difficult job). Moreover, a lot of basic commands (like reading blocks or write) are standards for all devices. And the gigarec/varirec commands have already been ported into another recording software for Unix, Solaris and so on (cdrecord).

Also, I am based in Europe. If only the Plextor commands were concerned by the sotware patents or a DMCA-like law, they do not apply there. And even if there was something like the DMCA in Europe, first we don't break any encryption, second even in the US, those liberticide laws do not apply for interoperability purposes, which is exactly the case here, since we offer to use some hardware under another OS.

Finally, Plextor would be stupid to fight against this. We just do their job for free, and make their hardware more attractive for more users. Even the Plextor Europe website gives links to cdrecord and other Plextor utilities under Linux too. More users = more money and everybody is happy

Just a small post to warn people I submitted the PxScan project to Sourceforge.net. I am waiting for its review and hopefully its acceptance. As soon as it is accepted, we will be able to deposit files and use CVS.

Also I am working on porting the Beta/Jitter test (but I have little troubles, soon to be solved I hope).

Hi Zeb,

first many many thanks for this great tool.

It works great on my AMD64 machine (Gentoo 64bit, Kernel 2.6.10). I get some segfaults at the end of the scans. But nothing serious.

On an other machine (P4, 32bit Gentoo, Kernel 2.6.10) I always get segfaults in pisum8 and pif in
When I do a

just before the mentioned lines, the tools seem to work.
Is this result_map necessary at all ?

ciao
Christoph

"result map" is just an array containing all data after scanning. It is needed to create a file containing all the data (e.g. on windows, a pxd file to be read by pxview).

There are not many reasons why that code could crash. I could imagine that
- "capacity" is invalid there (what is the value of "capacity" when the crash occurs?)
- your compiler is broken

Hi,

My downloads are offline for the moment. I hope to solve the problem soon.

This rings a bell for me. I dont declare capacity as an (__int64) but a (long). Does that matter ?

Hi,

the value of capacity is 17921, the value of result_map is -1207997544.
I don't think the compiler is broken, because I use Gentoo, and there, everything is compiled from source. I tried gcc-3.3.4 and gcc-3.4.3.
I also get a segmentation fault at the end of the scans, but haven't investigated further, yet.

ciao
Christoph

This is the adress that result_map points to. The results are stored in an array *(result_map +n) or result_map[n].

Sorry, I'm not very familiar with that pointer stuff.
But where is this address initialised ?
Is it done in
?

Hi,
actually it is a pointer of pointer of int (because there are 2 stars).
So actually result_map is a 2 dimensional table. If you print result_map or result_map[0], you get an address. However data are stored in result_map[x][y] with x and y the indices of your table.
Here, result map is reinitialised with calloc with the size of capacity, but I don't remember exactly the code (and don't have access here).
You need to take a good C or C++ course and read the pointer stuff It is quite tricky at the beginning, but logical when you get used to it. Actually C does not come with complex data structures, like in Perl or Python. Using pointers are a way to create array, tables at however dimensions.

Actually my Linux code is a quick and dirty port. I shall take time to clean and make it more rigorous, but for the moment I cannot do it

New download address for the linux version and information about a troll who wants customers of plextor drives to be forced to use Windows here

Actually we don't know yet who takeshima, the guy who made the complain, is. The mail was truncated and transmitted by the sourceforge moderators. We shall get more information on Monday hopefully.
If the project cannot be hosted on Sourceforge, there may be other solutions though, such as using a similar European host, because in Europe there is nothing like DMCA - even if there is no encryption involved in the data transmitted from/to the drive - and interoperability work is clearly permitted. I know a French equivalent of Sourceforge that will be happy to host us.
Sorry for the fuss. But I would be very disappointed if the complaint came from Plextor officials. There is nothing illegal in what we do.

If that complaint turns out to come from Plextor officials, then i'll make sure that as many people as possible know that they sell expensive drives and then want to prevent users from making use of all the unique features...

EDIT grammar

Yep. Slashdot will be a good start.