[U-Boot] [U-boot] fastboot command question
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
Best wishes,
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
Hi Tiger,
You bring up a question we're interested in as well.
It looks to me as if this died somewhere back in April of 2011 with this discussion:
http://lists.denx.de/pipermail/u-boot/2011-April/thread.html#90983
And a straw-man design document discussed here: http://lists.denx.de/pipermail/u-boot/2011-April/thread.html/#091597
And a message about licensing in May of 2012: http://lists.denx.de/pipermail/u-boot/2012-May/125304.html
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
Dear Eric Nelson,
Adding Wolfgang and Tom to CC.
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
Hi Tiger,
You bring up a question we're interested in as well.
It looks to me as if this died somewhere back in April of 2011 with this discussion:
http://lists.denx.de/pipermail/u-boot/2011-April/thread.html#90983
And a straw-man design document discussed here: http://lists.denx.de/pipermail/u-boot/2011-April/thread.html/#091597
And a message about licensing in May of 2012: http://lists.denx.de/pipermail/u-boot/2012-May/125304.html
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
Not that I am aware of, but I did not follow this discussion at all. What is fastboot anyway, is that the android thing?
Best regards, Marek Vasut
Hi, Eric : Based on 2011 maillist, linaro engineers wished to add fastboot feature into u-boot mainline code base, buf failed. But, i also could not find cmd_fastboot feature in uboot code base released by linaro.
For example: Linaro released uboot form Samsung's Origen board.
Best wishes,
Hi Tiger,
On 08/22/2013 02:52 AM, TigerLiu@viatech.com.cn wrote:
Hi, Eric : Based on 2011 maillist, linaro engineers wished to add fastboot feature into u-boot mainline code base, buf failed. But, i also could not find cmd_fastboot feature in uboot code base released by linaro.
Thanks. I saw that.
Linaro has lots of stuff on their plate, and it looks like this one didn't make the cut.
For example: Linaro released uboot form Samsung's Origen board.
Yep. And TI, Freescale, Samsung and others each have fastboot implementations, each in their own repositories.
This feature has been living around the periphery of U-Boot for years...
Regards,
Eric
Hi Marek,
On 08/21/2013 10:08 PM, Marek Vasut wrote:
Dear Eric Nelson,
Adding Wolfgang and Tom to CC.
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
Hi Tiger,
You bring up a question we're interested in as well.
It looks to me as if this died somewhere back in April of 2011 with this discussion:
http://lists.denx.de/pipermail/u-boot/2011-April/thread.html#90983
And a straw-man design document discussed here: http://lists.denx.de/pipermail/u-boot/2011-April/thread.html/#091597
And a message about licensing in May of 2012: http://lists.denx.de/pipermail/u-boot/2012-May/125304.html
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
Not that I am aware of, but I did not follow this discussion at all. What is fastboot anyway, is that the android thing?
Yeah. It's the Google/Android equivalent of DFU (plus some additional features).
Oddly, it has little to do with booting and isn't particularly fast ;)
Regards,
Eric
On Wed, Aug 21, 2013 at 05:30:52PM -0700, Eric Nelson wrote:
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
Hi Tiger,
You bring up a question we're interested in as well.
It looks to me as if this died somewhere back in April of 2011 with this discussion:
http://lists.denx.de/pipermail/u-boot/2011-April/thread.html#90983
And a straw-man design document discussed here: http://lists.denx.de/pipermail/u-boot/2011-April/thread.html/#091597
And a message about licensing in May of 2012: http://lists.denx.de/pipermail/u-boot/2012-May/125304.html
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
So, putting my TI hat on, I've poked one of our legal teams about one of the fastboot versions we've done (the shove everything into cmd_fastboot.c one, which I don't like as much as the split up ones, but, that's a technical thing not a license thing) to see if there's any problems or not.
Historically, Wolfgang objected, roughly, on the grounds of "great, Yet Another Standard by Bigcompany forcing things on us". Which I can understand, but frankly, it's more of an ABI than some of the "funny" things Android did within the kernel, so I'm willing to live with it, so long as the implementation is done well, and there's no legal hurdles. In fact, http://patchwork.ozlabs.org/patch/126797/ was "looks legally fine, needs technical changes".
Thanks Tom,
On 08/22/2013 06:59 AM, Tom Rini wrote:
On Wed, Aug 21, 2013 at 05:30:52PM -0700, Eric Nelson wrote:
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
<snip>
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
So, putting my TI hat on, I've poked one of our legal teams about one of the fastboot versions we've done (the shove everything into cmd_fastboot.c one, which I don't like as much as the split up ones, but, that's a technical thing not a license thing) to see if there's any problems or not.
Historically, Wolfgang objected, roughly, on the grounds of "great, Yet Another Standard by Bigcompany forcing things on us". Which I can understand, but frankly, it's more of an ABI than some of the "funny" things Android did within the kernel, so I'm willing to live with it, so long as the implementation is done well, and there's no legal hurdles.
Customers using Android ask for fastboot support, and this isn't a lot of code, so it seems worth pursuing IMHO.
In fact, http://patchwork.ozlabs.org/patch/126797/ was "looks legally fine, needs technical changes".
Hmmm. I missed this.
The only remaining objection seems to have been the commit message and attribution of the original source.
I'm CC'ing Aneesh to see if he can provide any guidance.
Regards,
Eric
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 08/22/2013 10:30 AM, Eric Nelson wrote:
Thanks Tom,
On 08/22/2013 06:59 AM, Tom Rini wrote:
On Wed, Aug 21, 2013 at 05:30:52PM -0700, Eric Nelson wrote:
On 08/21/2013 04:30 AM, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
<snip>
Do we need a clean-room implementation in order to comply with?
Is anybody working either the licensing or implementation of this?
So, putting my TI hat on, I've poked one of our legal teams about one of the fastboot versions we've done (the shove everything into cmd_fastboot.c one, which I don't like as much as the split up ones, but, that's a technical thing not a license thing) to see if there's any problems or not.
Historically, Wolfgang objected, roughly, on the grounds of "great, Yet Another Standard by Bigcompany forcing things on us". Which I can understand, but frankly, it's more of an ABI than some of the "funny" things Android did within the kernel, so I'm willing to live with it, so long as the implementation is done well, and there's no legal hurdles.
Customers using Android ask for fastboot support, and this isn't a lot of code, so it seems worth pursuing IMHO.
In fact, http://patchwork.ozlabs.org/patch/126797/ was "looks legally fine, needs technical changes".
Hmmm. I missed this.
The only remaining objection seems to have been the commit message and attribution of the original source.
I'm CC'ing Aneesh to see if he can provide any guidance.
He was another one of the folks looking to get things merged back then, and I don't think has time for it now.
- -- Tom
On Wed, Aug 21, 2013 at 07:30:15PM +0800, TigerLiu@viatech.com.cn wrote:
Hi, experts:
Why not include cmd_fastboot.c in common directory in 2013.07 release package?
Because of code license?
Or other reason?
It's technical reasons. I pointed our legal folks at: https://gitorious.org/rowboat/u-boot/blobs/am335x-master-android-ics/common/... as an example of the question about licenses, and the answer is:
"Yeah, this is an example of why I don't like mixing licensing in a file. It gets confusing. Here is the way I read it:
This file is GPLv2+ but contains BSD code. The only restriction is really on the BSD portion. Normally BSD would allow you to release object only and not have to provide the source. As there is a GPLv2 header on it the source would have to be provided. The + on the GPL really has no effect on the file per se.
I would say the license of this file is GPLv2+ and 2 Clause BSD like you said. Seems OK as long as you are using it somewhere where using GPL is OK with you.."
So, if someone would like to take one of the "fastboot" implementations that is GPLv2 and 2 Clause BSD and make it clean enough to merge, I'd be quite happy to review and get it in.
participants (4)
-
Eric Nelson -
Marek Vasut -
TigerLiu＠viatech.com.cn -
Tom Rini