Submitted by ReactOS, the open source implementation of a Windows XP/2003 compatible operating system, just published a new interview in their series of interviews with ReactOS developers. Today’s interview features the most active kernel developer Alex Ionescu.
Interview: Alex Ionescu, ReactOS Kernel Developer
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
its nice to see that people working on the project seem to be enjoying themselves. I really hope that ReactOS continues its development at a good pace. I agree with Alex that after 2.0 I to got a lot more excited abotu the project. keep up the good work guys
You can find out more information about Alex and his reverse engineering projects on his blog at OpenRCE:
http://www.openrce.org/blog/browse/AlexIonescu
“My name is Alex Ionescu, and I’m a reverse engineer”
and http://www.reactos.org/archives/public/ros-dev/2006-January/007393….
He’s very good at translating x86 assembly language into C!
Thanks for the promotion, but please use http://www.alex-ionescu.com for my blog instead; my OpenRCE blog will soon be linked to it as soon as Pedram finishes the site move.
Best regards,
Alex Ionescu
My favorite part of the interview is near the end:
“How hard was it to convince your boss to let you contribute to free software?
I’m working as a Microsoft Student Partner so I’ll let you guess about that one.”
Ah, the conspiracy theories I could come up with!
Alex is one of the most inspirational people I have been in contact with, infact he is one of heros.
Very inspirational!
“All my kernel code is reversed and doesn’t have a single magic constant.”
http://www.reactos.org/archives/public/ros-dev/2006-February/007879…
It’s very sad but at the same time quite amusing that Mike McCormack feels the needs to ridicule any new mention of ReactOS on the internet.
He is the ultimate troll, a blueprint for new trolls to come.
Back to the topic in hand. Alex is a great programmer, with a rare skill. It’s worth keeping an eye on his blog.
Edited 2006-11-22 08:16
There’s no doubt Alex is great with a disassembler. The problem is ofcourse that he uses that skill to translate disassembled Microsoft code to C and then commits that to ReactOS, thereby violating Microsofts copyrights. An example of that is the fastcall entry code (referenced by mike_m above).
A quote from the TinyKrnl (one of his projects) FAQ at http://www.tinykrnl.org/about_faq.htm
TinyKRNL is a research project as well as contains possibly Microsoft copyrighted or patented technologies or code
Another quote, from the interview:
It’s expected that it [TinyKRNL] will share a lot of its kernel and HAL code with ReactOS
Now, put one and one together and tell me if you still think Alex is the best thing that happened to the ReactOS project.
Alex seems to be very confused about copyright and licenses. The TinyKRNL about page says TinyKRNL will be released under the BSD license but then goes on to state that commercial use is not allowed. The BSD license allows unlimited commercial use. Besides, how you can release code copyrighted by Microsoft under the BSD is beyond me.
There is no doubt that GvG has a lot of trouble reading English, so perhaps someone should clarify things for him.
Merging HAL and Kernel code from ReactOS to TinyKRNL is appropriate use of the GPL, because that code will remain GPL licensed.
Secondly, the fast entry call is already written in assembly. Any competent OS developer should know that such low level code cannot be written in C, because registers mush be manually manipulated and precise trap frames created. It would be a pointless excercise to convert assembly to C when this C code is unusable. Since ReactOS does indeed boot, then I guess it’s not using any magic C code.
Finally, TinyKRNL does not use the BSD license, as the FAQ page clearly states. It uses its own license which is obtainable from the SVN server in the root folder, and all code specifies this. If GvG had actually done research (instead of another envious attempt at discreditting a developer), he would’ve seen this file. But perhaps he has and it’s not matching with his theories.
If there is a mention of the BSD license on the about page, it is a mistake and probably has already been fixed. As the frontpage mentions (perhaps GvG has not read that either), the webpages are still under construction, and our webmaster has been hospitalized for over a month now. Making sure he recovers from his injuries is more important to me then nagging him about a mistake on the about page, which is meaningless since the important place — the code — has the right license.
Once you get popular enough in life, it’s normal for you to start having envious people who try to discredit you. Mike_m and GvG are prime examples of this, as their comment history shows. It’s unfortunate that their inability to make a name of themselves must result in their attempts to destroy another’s name.
Best regards,
Alex Ionescu
There is no doubt that GvG has a lot of trouble reading English
Well, English isn’t my native language, that’s for sure.
Secondly, the fast entry call is already written in assembly
Yeah, I messed up there. In the case of the fast entry call you copied the disassembled code to asm instead of converting it to C. Minor mistake on my part, still a huge mistake on your part. Or are you now claiming the ReactOS fast entry code was written independently by you? Just a fair warning, I still have an email around where you admit to copying the code from Windows.
Merging HAL and Kernel code from ReactOS to TinyKRNL is appropriate use of the GPL, because that code will remain GPL licensed.
Finally, TinyKRNL does not use the BSD license, as the FAQ page clearly states.
Well, no, it clearly states TinyKRNL DOES use the BSD license. At this time (22 nov 18:30 GMT) it says “Because our components are linked with static libraries from the DDK as well as because some of our driver components are based on Microsoft Distributable Code, whose license prohibits usage of an Excluded License (such as the GPL or LGPL), these TinyKRNL components are licensed under the BSD license.”.
I have no reason to doubt that the license in the svn tree is different if you say so, but I never made claims about the license in the svn tree, only about the license mentioned on the FAQ.
Well, no, it clearly states TinyKRNL DOES use the BSD license. At this time (22 nov 18:30 GMT) it says “Because our components are linked with static libraries from the DDK as well as because some of our driver components are based on Microsoft Distributable Code, whose license prohibits usage of an Excluded License (such as the GPL or LGPL), these TinyKRNL components are licensed under the BSD license.”.
You really can’t read can you? “some of our driver components.” “these TinyKRNL components”. Yes, DDK samples are licensed BSD because this is 100% ALLOWED BY THE DDK EULA. Companies all over the WORLD are using samples for their commercial products even. ATI, NVDIA, VMWare, Maxtor, etc. Only the WDK license changed this but we don’t use WDK samples.
The rest of the paragraph you quoted says:
“[…]This restriction has led to the creation of the TinyKRNL Shared Source License, which is available in the root folder of the SVN server. Always check the source file to verify which license it belongs to.”
Thank you for spreading FUD!
Best regards,
Alex Ionescu
> Thank you for spreading FUD!
Its not FUD. The legality of using the DDK drivers in a non-windows project has always been in question. The DDK EULA seems to make the usage quite clear which did not allow for development for non-windows systems.
In any case, the reason why the former ReactOS developers and Wine project members are not happy with the ReactOS project it because the project demographics have changed. The old ReactOS crew wanted a truely clean room, independent implementation. The current group of developers have no objection to not operating in a clean-room. These are different standards which has lead to animosity. It makes the older developers in the Wine camp feel like all of their work could be for nothing as at some point in the future under US copyright law it could be regarded a derived work.
The question about rather the clean-room standard Wine, Samba and other projects apply is really ethically necessary is a debate worth having, however your not going to change what has been an accepted standard in FreeSoftware development.
Steven Edwards
Former ReactOS Project Leader and Wine Developer
Its not FUD. The legality of using the DDK drivers in a non-windows project has always been in question. The DDK EULA seems to make the usage quite clear which did not allow for development for non-windows systems.
Maybe it’s been in question for you, but not for me. We are not violating the EULA. (We would be violating the WDK 6000 EULA however, which is why we don’t use it). TinyKRNL drivers, unlike ReactOS drivers, are actually pure NT drivers that work under Windows 2003. This exempts them from “non-windows systems” clauses. What you state applies to ReactOS, which is why ReactOS does not use DDK code. Therefore, you simply have no point to make, sorry. Additionally, we heavily re-write and re-use DDK samples, so it’s not just simply re-releasing code.
If I were you, I’d be more careful about the various Win32 API patents that Wine is violating, as well as bits of CRT source code that match MSVC’s implementation line for line, before complaining about EULAs which are not been violated.
Best regards,
Alex Ionescu
“If I were you, I’d be more careful about the various Win32 API patents that Wine is violating, as well as bits of CRT source code that match MSVC’s implementation line for line, before complaining about EULAs which are not been violated.”
If you have an allegation about any line of code in Wine that may be viewed as tainted the Wine developers would have no problem asking for independent legal review, removal if needed and having someone independently re-implement it.
Hmmm, seems your position changed from “TinyKRNL does not use the BSD license, as the FAQ page clearly states” to “some of our driver components use the BSD license”. The way I read the FAQ page was: our components … these components which would include all components.
But you’re only responding to minor points, I really couldn’t care less about TinyKRNL (yeah, I know, my own fault that I brought it up). The important points as far as I’m concerned are:
a) did you write the fast entry call from scratch, without referencing disassembled Windows code?
b) if that code wasn’t written independently, is it still present in the current (HEAD) ReactOS tree?
Dear mike, why do you insist with your hate towards reactos and Alex in general and use your every chance to flame them when there is a news item about it on osnews?
I wonder if codeweavers pay you for flaming a wine friendly project, or you are just scared that reactos would put you out of business?
I can’t wait for a release I can install as my primary OS at home!
This is getting boring. Mike and GvG are flaming against ReactOS and tell loads of bullshit and everyone tries to shoot back at them by telling the truth. Things like these done by Mike and GvG are a crime in Germany and I’m very sure not only here. We call it “Rufmord” Just sue them to shut up and flame somebody else. Add a nice fee they have to pay if they will do it again.
@ GvG: Once you were one of my Idols in the ReactOS Project. Once… Quite sad that you were assimilated by the enemy (Hi Mike!) Because of some lines that were quite similar to MS Code. Really sad.
GvG are flaming against ReactOS and tell loads of bullshit
Please point out any factual errors in my posts. I’ve already conceded the C/asm mistake for the fast entry call.
everyone tries to shoot back at them by telling the truth
No, the reactions I see are people sticking their fingers in their ears and shouting “la la la I can’t hear you”. Only Alex responded to my arguments.
Once you were one of my Idols in the ReactOS Project. Once… Quite sad that you were assimilated by the enemy
I wasn’t in the ReactOS project to be an Idol or to make a name for myself. I was in it for the fun. Alex took the fun out for me, so I left. Oh, BTW, I’m still hosting some (small) scripts on my webserver since the current developers seem to be unable to correctly configure DNS. If I really had a grudge against the ReactOS project I’d have stopped doing that months ago.
For the record, yes, I submitted some patches to Wine but I don’t consider myself a Wine developer.
Because of some lines that were quite similar to MS Code.
Far more worysome than “some lines” was the reaction of the (other) developers. You’d expect that these lines would have been taken out asap, but no, almost a year after Hartmuts message they are still in svn HEAD. Heck, the file containing them is even marked as “passed audit”!
Far more worysome than “some lines” was the reaction of the (other) developers. You’d expect that these lines would have been taken out asap, but no, almost a year after Hartmuts message they are still in svn HEAD. Heck, the file containing them is even marked as “passed audit”!
janderwald@#reactos [11/22/06 20:19 CET] wrote:
this is not true
the file is not in locked list cose the file was deleted in rev 20936 and the contents was moved to trap.S which is still locked
(part of the contents was moved to trap.S)
can someone with anosnews account please post that?
Weird….
svn info svn://svn.reactos.org/reactos/trunk/reactos/ntoskrnl/ke/i386/trap.s
Path: trap.s
Name: trap.s
URL: svn://svn.reactos.org/reactos/trunk/reactos/ntoskrnl/ke/i386/trap.s
Repository Root: svn://svn.reactos.org/reactos
Repository UUID: 97493ccd-5924-5043-b1f5-66cb403b36ce
Revision: 24803
Node Kind: file
Last Changed Author: ion
Last Changed Rev: 24395
Last Changed Date: 2006-10-04 18:00:36 +0200 (Wed, 04 Oct 2006)
This is what a file looks like which is locked:
svn info svn://svn.reactos.org/reactos/trunk/reactos/ntoskrnl/ke/profile.c
Path: profile.c
Name: profile.c
URL: svn://svn.reactos.org/reactos/trunk/reactos/ntoskrnl/ke/profile.c
Repository Root: svn://svn.reactos.org/reactos
Repository UUID: 97493ccd-5924-5043-b1f5-66cb403b36ce
Revision: 24803
Node Kind: file
Last Changed Author: audit
Last Changed Rev: 21252
Last Changed Date: 2006-03-08 21:03:59 +0100 (Wed, 08 Mar 2006)
Lock Token: opaquelocktoken:634ab742-840e-0410-a072-e659af5f3666
Lock Owner: audit
Lock Created: 2006-03-09 00:38:44 +0100 (Thu, 09 Mar 2006)
Lock Comment (1 line):
audit lock
The trap.s file is also not listed in http://www.reactos.org/generated/locked_files.log
FUD… FUD… more FUD…
This is how we count locked files:
http://www.alex-ionescu.com/fud.jpg
Notice that needs lock is ON.
Best regards,
Alex Ionescu
Seems the procedure for locking files has changed since it was introduced: http://reactos.org/archives/public/ros-dev/2006-March/007962.html
did call for a “svn lock” operation, in addition to the svn:need-lock property. Then again, initially it was also agreed that “locked” meant that no changes should be made to the file until it was audited. I guess that policy was changed too, there were 33 commits to trap.s after it was “locked”.
Ge, it’s possible the svn lockness was accidentally removed, it shouldn’t have been. If you find any more such files I’ll be glad to re-lock them with “svn lock”.
Also, we agreed that critical fixes should be allowed. Trap.s is a large file, and some of the code is required for other purposes. It has not yet been audited, however.
Best regards,
Alex Ionescu
It shouldn’t be too hard to write a Perl script to check, but I guess you can imagine I’m not motivated to write that script.
Ge, it’s possible the svn lockness was accidentally removed, it shouldn’t have been. If you find any more such files I’ll be glad to re-lock them with “svn lock”.
Any commit removes the lock. That does mean, the files from http://www.reactos.org/generated/locked_files.log weren’t changed so far. That does also mean, the audit is simply a big fake.
Changes, no. Fixes, yes.
And you are just an Idiot, who does not know what he is talking about. If you do not know how the audit works, just shut up and do not spread shitty lies.
If you do not know how the audit works, …
If you are an expert for the audit, you may explain for example the status of ntoskrnl/io/pnpmgr/pnpdma.c. The file was locked for the audit with rev 21252. This file was included in the commit of rev 22747 from Alex. And now, the file is unlocked.
And you are just an Idiot, …
…just shut up and do not spread shitty lies.
If yours ReactOS guys lose the arguments, yours become very childish…