k3flasher, a USB download mode utility

[knc1]

Originally Posted by geekmaster

I did not read the license (yet). Can the modified source be published openly or not (with their license included)? Or is there some restriction about registering at their website and downloading the original source from them and modifying it yourself?

Can not publish the source.
Exception: *.h files, no re-distribution restrictions.

Can only publish object code (original or derived) as part of an application for a licensed i.MX device (the kindles are).
May not publish object code "free standing" whatever Freescale legal department thinks that is.

Which is why I suggested to Hawhill he "build it into" K3flasher as a byte array with bin2c.

I'll push the two *.PDF files to the project directory (once I find where I put them - I have 1.5Tbytes here, 2Tbytes on-line and a bad memory in my head).

[geekmaster]

Originally Posted by knc1

... Can only publish object code (original or derived) as part of an application for a licensed i.MX device (the kindles are).
May not publish object code "free standing" whatever Freescale legal department thinks that is. ...

That sounds to me like you can publish a linked module but not linkable .o or .obj files. So it should be safe to distribute both the .exe and the RAM kernel file (especially when packaged together in a .zip file as suggested by the freescale agreement). Embedding a binary into an executable (even if encoded) could violate their "no GPL" clause, but a freestanding binary should be fine. As far as GPL cares, the RAM kernel is just data, but if embedded it could be interpreted as code.

Proprietary and GPL programs can only communicate through STDIO, as I interpret the various legal decisions. So a GUI wrapper around GPL is okay, especially if the end user downloads the GPL part separately from an authorized website or you provide it on separate media (or if the program downloads it automatically on request by the end user).

It would be best to avoid GPL altogether if using any freescale source code here, but LGPL is fine if you do not static link to it.

There is a Lawyer who frequents the parent forum (Harmon). He might give us an informed (non-binding) free opinion on this matter if asked nicely.
https://www.mobileread.com/forums/sho....php?p=2195551

[knc1]

Yeah, a person could mentally choke to death on those license terms.
Plus they where written before the days of collabortive development teams.

My position (until Freescale locks me up):

I only distribute username/password access to those who "need to know" for the purpose of supporting an i.MX licensed device.

The people having a valid username/password pair, by its use, is the written acknowledgement that they recognize the materials are of a "limited distribution" nature.

That should cover the general "Protect our IP as you would protect your own" clause.

One of the problems (which Freescale said they would fix in the next release) is that this material is covered by two (2) of the Freescale licenses but you can only see one of them when you sign for the material on their site.

They **should** be including both licenses in the archive bundle next time.
They **should** be including the sources to the GPL's binaries they are shipping in the ATK source bundle next time.
**should**
But they also said: "Thank you" for bring those matters to their attention. Nice folk.

[knc1]

@GM : I only wrote that the Freescale license is "GPL like" in the sense that it is viral.
I did not (intend to) make any suggestion that it be combined with GPL code.

PS: Hawhill's K3flasher is not GPL code, he can embed the RAM kernel if he wants to.

Plus: the only communication with the RAM kernel is over a communications link, it isn't even as advanced as RPC.

[geekmaster]

Originally Posted by knc1

...
But they also said: "Thank you" for bring those matters to their attention. Nice folk.

Yeah, amazon support staff told me "it should be fine for you to distribute kindle firmware images for repair purposes", but when I asked for that in writing, they said "our legal staff will contact you in a few days". But they never did.

They also failed to get back to me "in a few days" on other technical issues as well. Apparently "in a few days" is a euphamism for "when hell freezes over".

Originally Posted by knc1

@GM : I only wrote that the Freescale license is "GPL like" in the sense that it is viral.
I did not (intend to) make any suggestion that it be combined with GPL code.

PS: Hawhill's K3flasher is not GPL code, he can embed the RAM kernel if he wants to.

Plus: the only communication with the RAM kernel is over a communications link, it isn't even as advanced as RPC.

I understood that the first time I read it. Hawhill has used GPL in the past, which was my only concern. It is good that he has forsaken GPL in this particular instance, to simplify things. I do not think it is required to embed the RAM kernel in his executable, but I approve in that it is one less thing for an end-user to keep track of. I like monolithic executables with minimal external dependencies that can break later.

[knc1]

Originally Posted by knc1

I'll push the two *.PDF files to the project directory (once I find where I put them - I have 1.5Tbytes here, 2Tbytes on-line and a bad memory in my head).

Done.

Take note of the src_license.pdf where it says that any changes, including but not limited to, suggestions and bug-fixes become Freescale's property.

Those two documents are home-made *.pdf files, made from the link targets of the two links that the Freescale License Compliance department sent me.
They only differ by changes in whitespace during the conversion process.

[knc1]

All of this "license chatter" is an important part of "real world programming" - -
Respecting other people's IP rights is a sensitive topic to myself.

If any contributors or consultants having access to the topic webdav Kloud don't agree with Freescale's terms - just destroy your username/password pair.

This is, and must remain, Freescale's code we are playing with according to their license terms.

[knc1]

I added a bit of structure to the topic webdav://kloud

Note:
There are no read/write permission controls here, only an access control. If you have access, you have read/write/create/delete access.

The only thing keeping people spread across 8 time zones from stepping on each other's toes here is common sense.
No hardware, software, or human "god" to keep things under control. Everyone is an equal.

If you accidentally delete someone else's favorite file, it is gone.

[hawhill]

Originally Posted by geekmaster

I understood that the first time I read it. Hawhill has used GPL in the past, which was my only concern. It is good that he has forsaken GPL in this particular instance, to simplify things.

In fact, it was merely kind of a renewal of a funny license joke I've played in another instance. It was meant to scare newbies enough to not let them try it unprepared for havoc - the software was only tested on my own device at that point. I'm open to license it under any license. As personal preferences go, I'm a fan of the MIT/BSD license family. However, I understand the reasons to use GPL, and I don't like arguing about them.

So: if needed, I'll happily relicense to what is requested. Just drop a note.

[geekmaster]

Originally Posted by hawhill

In fact, it was merely kind of a renewal of a funny license joke I've played in another instance. It was meant to scare newbies enough to not let them try it unprepared for havoc - the software was only tested on my own device at that point. I'm open to license it under any license. As personal preferences go, I'm a fan of the MIT/BSD license family. However, I understand the reasons to use GPL, and I don't like arguing about them.

So: if needed, I'll happily relicense to what is requested. Just drop a note.

I try to avoid GPL whenever possible, but sometimes a piece of GPL is just too good to bother trying to reinvent it from scratch. I suspect that is why your muPDF project is GPL. It appears that MIT and BSD are compatible with the Freescale license, if you do not openly publish their modified source modules that you link to your program.

[knc1]

Whoops !
There is a restriction on what licenses the over-all work can be licensed under - - -
The layman's translation is: "May not be licensed under terms that would require revealing and/or providing the Freescale source code."

I.E: @Hawhill : Please do not re-license you k3flasher as GPL, it would prevent us from distributing the modified RAM kernel with it.

The current license or MIT or BSD or anything that does not **require** the end user be given access to the Freescale source is just fine. Access to the Freescale object code as embedded data should be alright.

[knc1]

Originally Posted by knc1

I added a bit of structure to this topic's webdav://kloud

Old School: All projects start with a trip to the library.

New School:
webdav://kloud/docs - reference materials about the software to be modified.
webdav://kloud/DX - i.MX31 hardware reference materials
webdav://kloud/K3 - i.MX35 hardware reference materials
webdav://kloud/K4K5 - i.MX50 hardware reference materials.

K3Flasher is only for use with the K3.
The DX, K4, and K5 materials are there because they may someday be relevant.

Going by the file dates in the ATK-1.71 source bundle, the files are older than the introduction date of the i.MX508.
Which is probably why the ATK does not mention support for that SoC.
Since this stuff is a very low level function of the silicon in all Freescale SoCs, I am assuming it is in the silicon even though it is not yet mentioned in the (older) ATK collection of loader scripts.

First question: Is that watchdog time-out reasonable for all conditions found in the field?
read, read, read, read, rtfm.

[knc1]

I have been hoping to find a good reason to NOT rebuild the Freescale RAM kernel.
Not having much luck with reaching that goal.

New K3 problem in another thread here:
How well does this binary work under "low clock speed, low voltage, low power" mode of operation (set by the SoC hardware, outside of the control of the current RAM kernel)?

@Hawhill ::
I wrote earlier I wasn't going to change the protocol.
But it is looking more and more like we need an enhanced "get status or get info" command so that the host software can tell if it is safe to proceed.

Speed bump: The i.MX35 "manual" is over 3,000 pages long. It will take awhile to read it.

[qlob]

Originally Posted by qlob

Hrm......
Got it to work now, but this could be a bad sign:

[root@frankenboxen k3flasher]# sudo ./k3flasher mx35to2_mmc.bin info
I: found suitable device
E: wrong transfer length, wanted to receive 8 bytes but received 4 bytes.
I: above error can be ignored, it's due to the device being in ROM kernel mode
I: RAM kernel should be running now. Trying to re-open device: .
I: got it.
I: read info
<long pause>
E: reading data from device: LIBUSB_ERROR_NO_DEVICE
E: aborting. It is suggested you power-cycle the device.

Got the same results running it with the other RAM kernel (mx35to1.bin).....

[knc1]

Thanks for the report.
I was just hoping it would be something as simple as using a different version binary.