Re: [coreboot] [GSOC 2016] Coreboot panic room

[Denis 'GNUtoo' Carikli]

On Thu, 10 Mar 2016 22:34:45 +
Antonello Dettori  wrote:

> Hello everyone,
Hi,

> I would like to concentrate most of my efforts towards improving and 
> upstreaming the previous efforts, implementing a way to easily access 
> the recovery mode when needed and further the integration between 
> coreboot, serialICE and flashrom for this use-case.
> 
> Regarding the existing patches I would like to know if they would
> need to stay romc-compatible or should the scope be limited to CAR
> boards?
How was it implemented?

I see several building blocks:
- nvramtool and the fallback/normal mecanism.
- Many hardware also do have watchdogs, but I believe it's disabled by
  coreboot.
- SerialICE: I'm not sure it can work from coreboot, or how to
  integrate it in the same flash. Patches to add some kind of support
  for it in coreboot might exist in gerrit.
- libhwremote: works with SerialICE, and was made to work with pciutils
  and superiotool
- Coreboot also has gdb support, and even something like gdbwait, but
  only in ramstage.
- Coreboot also has some code to write to some flash chips, in order to
  store its logs there.

Denis.


pgpuEkLZ9nWfp.pgp
Description: OpenPGP digital signature
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[ron minnich]

Actually, this seems a pretty good proposal in its current state, but not
my one comment.

thanks

ron

On Wed, Mar 23, 2016 at 12:14 PM ron minnich  wrote:

> Thanks, now I understand, and I will try to get to this tonight. Sorry to
> be causing your worries.
>
> ron
>
> On Wed, Mar 23, 2016 at 12:09 PM Yurii Shevtsov  wrote:
>
>> I'm not complaining. I'm just worrying, only two days left till
>> submission deadline, and, who knows, maybe people forget. I'll be patient
>>
>> 2016-03-23 20:55 GMT+02:00 ron minnich :
>>
>>> you need to be patient. All the reviewers have day jobs and there is a
>>> lot of work in all this.
>>>
>>> There's no advantage to you or your proposal in complaining about the
>>> reviewers, I think.
>>>
>>>
>> Thanks for you reply. I'm not complaining. I'm just worrying, only two
>> days left till submission deadline, and, who knows, maybe people forget.
>>
>>> ron
>>>
>>>
>>>
>>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[ron minnich]

Thanks, now I understand, and I will try to get to this tonight. Sorry to
be causing your worries.

ron

On Wed, Mar 23, 2016 at 12:09 PM Yurii Shevtsov  wrote:

> I'm not complaining. I'm just worrying, only two days left till submission
> deadline, and, who knows, maybe people forget. I'll be patient
>
> 2016-03-23 20:55 GMT+02:00 ron minnich :
>
>> you need to be patient. All the reviewers have day jobs and there is a
>> lot of work in all this.
>>
>> There's no advantage to you or your proposal in complaining about the
>> reviewers, I think.
>>
>>
> Thanks for you reply. I'm not complaining. I'm just worrying, only two
> days left till submission deadline, and, who knows, maybe people forget.
>
>> ron
>>
>>
>>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[Yurii Shevtsov]

I'm not complaining. I'm just worrying, only two days left till submission
deadline, and, who knows, maybe people forget. I'll be patient

2016-03-23 20:55 GMT+02:00 ron minnich :

> you need to be patient. All the reviewers have day jobs and there is a lot
> of work in all this.
>
> There's no advantage to you or your proposal in complaining about the
> reviewers, I think.
>
>
Thanks for you reply. I'm not complaining. I'm just worrying, only two days
left till submission deadline, and, who knows, maybe people forget.

> ron
>
>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[ron minnich]

you need to be patient. All the reviewers have day jobs and there is a lot
of work in all this.

There's no advantage to you or your proposal in complaining about the
reviewers, I think.

ron

On Wed, Mar 23, 2016 at 11:37 AM Yurii Shevtsov  wrote:

> Seems like making draft public doesn't make any difference
>
> 2016-03-17 23:26 GMT+02:00 Yurii Shevtsov :
>
>>
>>
>> 2016-03-17 23:00 GMT+02:00 David Hendricks :
>>
>>>
>>> On Thu, Mar 17, 2016 at 1:39 PM, Yurii Shevtsov 
>>> wrote:
>>>
 Hello. Three days passed, but unfortunately I haven't got any feedback
 on my draft yet.

>>>
>>> If you make the doc public then you might get more responses.
>>>
>>
>> David, you convinced me. Thanks for your advice. Now it's public!
>>
>> https://docs.google.com/document/d/1rxKXzbaq9w4EthVIP2O5mqXaTJTzK3IKsnbVOWvstVI/edit?usp=sharing
>>
>>>
>>> Martin and Ron are both very busy people and might simply have been too
>>> busy with work this week.
>>>
>>> GSOC is all about open source coding and community engagement. Keeping
>>> your doc private is not a good way to start...
>>>
>>>
 Although, I can see Martin Roth is watching my proposal. Also I'm still
 not assigned to any task. Spare me some time, please :-)
 Thanks in advance

 2016-03-14 23:11 GMT+02:00 Yurii Shevtsov :

> I know we haven't discussed project properly yet, and I understand my
> vision of project may vary from yours. So feel free to point me on a right
> way :-)
> Here is the link:
>
> https://docs.google.com/document/d/1OWOYfKMUSZTi4tTHmGIIY27KAbM5MJVV0DdirjpB2As/edit?usp=sharing
>
> I made my proposal draft private. As for now I shared it with Ron
> Minnich and Martin Roth. If you part of a GSoC commission, I'll add you 
> too
> (head to the doc link).
>
> Unfortunately, I haven't submitted any patch yet. Please, assign me to
> a task.
>


 --
 coreboot mailing list: coreboot@coreboot.org
 https://www.coreboot.org/mailman/listinfo/coreboot

>>>
>>>
>>>
>>> --
>>> David Hendricks (dhendrix)
>>> Systems Software Engineer, Google Inc.
>>>
>>
>>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Denis 'GNUtoo' Carikli]

On Wed, 09 Mar 2016 04:24:04 +
ron minnich  wrote:

> yeah, we had something like this in the linuxbios days. I think you
> don't want to build it on demand, but rather have a bunch of
> pre-built images that are known good.
Once we get reproducible builds, we could:
1) Store an (SHA) hash of images in board-status.
2) Make the rom-o-matic build (with the board-status configs) and
   verify(checksums) the image of a known good state.

That would still not guarantee that the given image would work on every
variation of a given device[1], but that would still be better than the
current status today.

References:
---
[1]I'm thinking of mainboards supporting several CPU families: In that
   case the rom-o-matic would somehow need to inform the user about the
   tested configuration. RAM sizes and modules also often varies
   between users.

Denis.


pgpN2funkbgoX.pgp
Description: OpenPGP digital signature
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

On Thu, Mar 17, 2016 at 2:33 PM Yurii Shevtsov  wrote:

>
> I can't see any other way, it's all about running the commands. Browsers
> can't do this. Certain actions are still required from user. Some shell
> script, which will run commands and then send stdout to the server, can be
> developed though.
>
>>
>>

the board-status script can send the ref and the configuraiton.

That's not part of rom-o-matic.

Further, tools at coreboot.org and create the tables whih rom-o-matic uses
to present choices to the user.

That's what I was wondering, anyway.

ron
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

2016-03-17 22:13 GMT+02:00 ron minnich :

>
>
> On Thu, Mar 17, 2016 at 1:08 PM Denis 'GNUtoo' Carikli 
> wrote:
>
>>
>> Once we get reproducible builds, we could:
>> 1) Store an (SHA) hash of images in board-status.
>> 2) Make the rom-o-matic build (with the board-status configs) and
>>verify(checksums) the image of a known good state.
>>
>>
>>
>
> I really like this idea. I wonder if we could add additional information
> for a board, derived from, say, pci tree, dmidecode, that kind of thing,
> that would be used to characterize a board.
>
> ron
>

I like it too. Maybe we can ask users on rom-o-matic to provide neccesary
info for further project improvement
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

On Thu, Mar 17, 2016 at 1:08 PM Denis 'GNUtoo' Carikli 
wrote:

>
> Once we get reproducible builds, we could:
> 1) Store an (SHA) hash of images in board-status.
> 2) Make the rom-o-matic build (with the board-status configs) and
>verify(checksums) the image of a known good state.
>
>
>

I really like this idea. I wonder if we could add additional information
for a board, derived from, say, pci tree, dmidecode, that kind of thing,
that would be used to characterize a board.

ron
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[ron minnich]

I can't access the document

On Thu, Mar 17, 2016 at 1:40 PM Yurii Shevtsov  wrote:

> Hello. Three days passed, but unfortunately I haven't got any feedback on
> my draft yet. Although, I can see Martin Roth is watching my proposal. Also
> I'm still not assigned to any task. Spare me some time, please :-)
> Thanks in advance
>
> 2016-03-14 23:11 GMT+02:00 Yurii Shevtsov :
>
>> I know we haven't discussed project properly yet, and I understand my
>> vision of project may vary from yours. So feel free to point me on a right
>> way :-)
>> Here is the link:
>>
>> https://docs.google.com/document/d/1OWOYfKMUSZTi4tTHmGIIY27KAbM5MJVV0DdirjpB2As/edit?usp=sharing
>>
>> I made my proposal draft private. As for now I shared it with Ron Minnich
>> and Martin Roth. If you part of a GSoC commission, I'll add you too (head
>> to the doc link).
>>
>> Unfortunately, I haven't submitted any patch yet. Please, assign me to a
>> task.
>>
>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Denis 'GNUtoo' Carikli]

On Thu, 17 Mar 2016 15:56:13 -0600
Martin Roth  wrote:

> As the project proposal says, first we want to just build roms that
> are already present in the board status database.  We have a config
> already present for these, so we don't need to ask any configuration
> options at all, just give them a list of boards that they can build a
> rom for.  We also don't need to give users every option available in
> Kconfig.  Maybe we give them an edit window that they can paste a
> config into or edit an existing config.
I would see ROM-O-Matic for coreboot working this way:

The users would have a list of boards, and could then select the given
known-good configuration that they want.

The configuration would list the hardware variations used in the test,
such as RAM, CPU when relevant[3].
- It would also warn the user when additional binaries have to be
  added to cbfs[1]. The user would then need to keep that in mind when
  consulting the installation documentation.
- It would warn against or filter out configurations with unknown
  payloads, or payloads with unknown revisions.
- It would warn against or filter out boards and configurations not in
  board-status.

Then an image would be produced, but also its hash and signature.
The website would also use https (everywhere).

Along with the image (and its hash and signature), some information
would appear:
- A big warning if some are lacking in cbfs[2].
- If the files matches, but the hash differ from the one in
  board-status[4], then a big warning will appear too.
- A link to the relevant documentation explaining how to install/flash
  this image for the given board.

I'm undecided on the idea of permitting users to do dangerous things,
such as editing the configuration, building for master revisions of
payloads, building for untested boards:
- One one hand, it would permit users to test that configuration, and
  then to upload the result on board-status more easily.
- On the other hand, we might rather want the given configuration to be
  uploaded on board-status first. This would be to reduce bricking
  probabilities.
- Another option would be to have both. In that case, we might want to
  very strongly warn the user if they want to have their own
  configuration built.

References:
---
[1]Part the necessary binaries can be deduced from coreboot's .config.
[2]That can be deduced from board-status: it has the list of CBFS files.
[3]This would have to be deduced from the logs. It's not trivial, so
   maybe there is some better way. Doing it at board-status log
   submission time is probably a better idea. Its scripts would have to
   be modified for that.
[4]board-status scripts will need to be modified to add such feature.
   Coreboot build scripts will also need to be modified to make sure
   that builds are reproducible. In coreboot,
   util/genbuild_h/genbuild_h.sh still leaks some non-reproducible
   information. Fixing it would also be needed to give users proper
   feedback.

Denis.


pgpfZmQMq8gCo.pgp
Description: OpenPGP digital signature
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Martin Roth]

Sorry it's taken a while to get back to you on this.

The build needs to happen with the correct toolchain to generate a rom that
is similar (or identical) to the original rom that was tested in the board
status.  That would be done with with 'make crossgcc'.  If I try to build a
rom that was submitted 8 months ago, I'd want to use the toolchain that was
in use 8 months ago, not today's toolchain.

Boards that are in the board status repo:
https://review.coreboot.org/gitweb/cgit/board-status.git/tree/amd
These are boards that have been tested.  These submissions are known to at
least boot (with the configuration that was used to test the rom).

Any valid board:
https://review.coreboot.org/gitweb/cgit/coreboot.git/tree/src/mainboard/amd
The source code for these boards is in the codebase, but we have no way of
knowing whether it's currently booting or not until we actually test it.

As the project proposal says, first we want to just build roms that are
already present in the board status database.  We have a config already
present for these, so we don't need to ask any configuration options at
all, just give them a list of boards that they can build a rom for.  We
also don't need to give users every option available in Kconfig.  Maybe we
give them an edit window that they can paste a config into or edit an
existing config.

The preference would be to do everything on the server or in the browser
though.  We would really prefer not to have any scripts that need to be run
on the user's system.


Martin


On Thu, Mar 17, 2016 at 3:35 PM, ron minnich  wrote:

>
>
> On Thu, Mar 17, 2016 at 2:33 PM Yurii Shevtsov  wrote:
>
>>
>> I can't see any other way, it's all about running the commands. Browsers
>> can't do this. Certain actions are still required from user. Some shell
>> script, which will run commands and then send stdout to the server, can be
>> developed though.
>>
>>>
>>>
>
> the board-status script can send the ref and the configuraiton.
>
> That's not part of rom-o-matic.
>
> Further, tools at coreboot.org and create the tables whih rom-o-matic
> uses to present choices to the user.
>
> That's what I was wondering, anyway.
>
> ron
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[David Hendricks via coreboot]

On Thu, Mar 17, 2016 at 1:39 PM, Yurii Shevtsov  wrote:

> Hello. Three days passed, but unfortunately I haven't got any feedback on
> my draft yet.
>

If you make the doc public then you might get more responses.

Martin and Ron are both very busy people and might simply have been too
busy with work this week.

GSOC is all about open source coding and community engagement. Keeping your
doc private is not a good way to start...


> Although, I can see Martin Roth is watching my proposal. Also I'm still
> not assigned to any task. Spare me some time, please :-)
> Thanks in advance
>
> 2016-03-14 23:11 GMT+02:00 Yurii Shevtsov :
>
>> I know we haven't discussed project properly yet, and I understand my
>> vision of project may vary from yours. So feel free to point me on a right
>> way :-)
>> Here is the link:
>>
>> https://docs.google.com/document/d/1OWOYfKMUSZTi4tTHmGIIY27KAbM5MJVV0DdirjpB2As/edit?usp=sharing
>>
>> I made my proposal draft private. As for now I shared it with Ron Minnich
>> and Martin Roth. If you part of a GSoC commission, I'll add you too (head
>> to the doc link).
>>
>> Unfortunately, I haven't submitted any patch yet. Please, assign me to a
>> task.
>>
>
>
> --
> coreboot mailing list: coreboot@coreboot.org
> https://www.coreboot.org/mailman/listinfo/coreboot
>



-- 
David Hendricks (dhendrix)
Systems Software Engineer, Google Inc.
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[Yurii Shevtsov]

Hello. Three days passed, but unfortunately I haven't got any feedback on
my draft yet. Although, I can see Martin Roth is watching my proposal. Also
I'm still not assigned to any task. Spare me some time, please :-)
Thanks in advance

2016-03-14 23:11 GMT+02:00 Yurii Shevtsov :

> I know we haven't discussed project properly yet, and I understand my
> vision of project may vary from yours. So feel free to point me on a right
> way :-)
> Here is the link:
>
> https://docs.google.com/document/d/1OWOYfKMUSZTi4tTHmGIIY27KAbM5MJVV0DdirjpB2As/edit?usp=sharing
>
> I made my proposal draft private. As for now I shared it with Ron Minnich
> and Martin Roth. If you part of a GSoC commission, I'll add you too (head
> to the doc link).
>
> Unfortunately, I haven't submitted any patch yet. Please, assign me to a
> task.
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

2016-03-17 22:57 GMT+02:00 ron minnich :

>
>
> On Thu, Mar 17, 2016 at 1:30 PM Yurii Shevtsov  wrote:
>
>>
>>
>>>
>> I like it too. Maybe we can ask users on rom-o-matic to provide neccesary
>> info for further project improvement
>>
>
> no, it has to be automated or it won't work.
>

I can't see any other way, it's all about running the commands. Browsers
can't do this. Certain actions are still required from user. Some shell
script, which will run commands and then send stdout to the server, can be
developed though.

> ron
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] Proposal review

[Yurii Shevtsov]

2016-03-17 23:00 GMT+02:00 David Hendricks :

>
> On Thu, Mar 17, 2016 at 1:39 PM, Yurii Shevtsov  wrote:
>
>> Hello. Three days passed, but unfortunately I haven't got any feedback on
>> my draft yet.
>>
>
> If you make the doc public then you might get more responses.
>

David, you convinced me. Thanks for your advice. Now it's public!
https://docs.google.com/document/d/1rxKXzbaq9w4EthVIP2O5mqXaTJTzK3IKsnbVOWvstVI/edit?usp=sharing

>
> Martin and Ron are both very busy people and might simply have been too
> busy with work this week.
>
> GSOC is all about open source coding and community engagement. Keeping
> your doc private is not a good way to start...
>
>
>> Although, I can see Martin Roth is watching my proposal. Also I'm still
>> not assigned to any task. Spare me some time, please :-)
>> Thanks in advance
>>
>> 2016-03-14 23:11 GMT+02:00 Yurii Shevtsov :
>>
>>> I know we haven't discussed project properly yet, and I understand my
>>> vision of project may vary from yours. So feel free to point me on a right
>>> way :-)
>>> Here is the link:
>>>
>>> https://docs.google.com/document/d/1OWOYfKMUSZTi4tTHmGIIY27KAbM5MJVV0DdirjpB2As/edit?usp=sharing
>>>
>>> I made my proposal draft private. As for now I shared it with Ron
>>> Minnich and Martin Roth. If you part of a GSoC commission, I'll add you too
>>> (head to the doc link).
>>>
>>> Unfortunately, I haven't submitted any patch yet. Please, assign me to a
>>> task.
>>>
>>
>>
>> --
>> coreboot mailing list: coreboot@coreboot.org
>> https://www.coreboot.org/mailman/listinfo/coreboot
>>
>
>
>
> --
> David Hendricks (dhendrix)
> Systems Software Engineer, Google Inc.
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

On Thu, Mar 17, 2016 at 1:30 PM Yurii Shevtsov  wrote:

>
>
>>
> I like it too. Maybe we can ask users on rom-o-matic to provide neccesary
> info for further project improvement
>

no, it has to be automated or it won't work.

ron
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

Thanks for your reply, it clarifies the main idea for me.

2016-03-17 23:56 GMT+02:00 Martin Roth :

> Sorry it's taken a while to get back to you on this.
>
> The build needs to happen with the correct toolchain to generate a rom
> that is similar (or identical) to the original rom that was tested in the
> board status.  That would be done with with 'make crossgcc'.  If I try to
> build a rom that was submitted 8 months ago, I'd want to use the toolchain
> that was in use 8 months ago, not today's toolchain.
>

So I need to came up with some mechanism to store multiple toolchain
version.

> Boards that are in the board status repo:
> https://review.coreboot.org/gitweb/cgit/board-status.git/tree/amd
> These are boards that have been tested.  These submissions are known to at
> least boot (with the configuration that was used to test the rom).
>
> Any valid board:
> https://review.coreboot.org/gitweb/cgit/coreboot.git/tree/src/mainboard/amd
> The source code for these boards is in the codebase, but we have no way of
> knowing whether it's currently booting or not until we actually test it.
>

Did I understand correctly, that there is no possibility to automatically
test all these boards, because they require specific hardware, which can't
be emulated? So, if user wants to build and flash one, he takes all
responsibility?

> As the project proposal says, first we want to just build roms that are
> already present in the board status database.  We have a config already
> present for these, so we don't need to ask any configuration options at
> all, just give them a list of boards that they can build a rom for.  We
> also don't need to give users every option available in Kconfig.  Maybe we
> give them an edit window that they can paste a config into or edit an
> existing config.
>


> The preference would be to do everything on the server or in the browser
> though.  We would really prefer not to have any scripts that need to be run
> on the user's system.
>

I strongly agree about this.

>
> Martin
>
>
> On Thu, Mar 17, 2016 at 3:35 PM, ron minnich  wrote:
>
>>
>>
>> On Thu, Mar 17, 2016 at 2:33 PM Yurii Shevtsov  wrote:
>>
>>>
>>> I can't see any other way, it's all about running the commands. Browsers
>>> can't do this. Certain actions are still required from user. Some shell
>>> script, which will run commands and then send stdout to the server, can be
>>> developed though.
>>>


>>
>> the board-status script can send the ref and the configuraiton.
>>
>> That's not part of rom-o-matic.
>>
>> Further, tools at coreboot.org and create the tables whih rom-o-matic
>> uses to present choices to the user.
>>
>> That's what I was wondering, anyway.
>>
>> ron
>>
>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

I admit, I oversimplified. Build server have much more advanced features.
But user still needs to configure ROM, like nconfig/menuconfig prompts, as
a first step, isn't it? Also, why would server need to rebuild toolchain,
isn't crossgcc alone enough? What is the difference between "boards listed
in board-status repository" and "any valid board", what is meant by these
terms?

2016-03-12 7:41 GMT+02:00 ron minnich :

>
>
> On Fri, Mar 11, 2016 at 5:28 PM Yurii Shevtsov  wrote:
>
>> But personally I think it's fun. Also, now I understand, that rom-o-matic
>> should be web-version of nconfig/menuconfig
>>
>>
>
> I'm not sure that's quite right. But we can talk. You definitely know what
> you are doing at this point :-)
>
> ron
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

On Fri, Mar 11, 2016 at 5:28 PM Yurii Shevtsov  wrote:

> But personally I think it's fun. Also, now I understand, that rom-o-matic
> should be web-version of nconfig/menuconfig
>
>

I'm not sure that's quite right. But we can talk. You definitely know what
you are doing at this point :-)

ron
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

Proof of running rom file:
http://pastebin.com/P2WzyTQ8

While I was going through building process, I got quite a view of how
inconvenient it can be - setting the whole building environment, waiting
toolchain compilation. But personally I think it's fun. Also, now I
understand, that rom-o-matic should be web-version of nconfig/menuconfig

BTW, I occasionally found student requirements on GSoC mentors page. And
I'm asking for a simple task)

2016-03-09 22:57 GMT+02:00 ron minnich :

> On Wed, Mar 9, 2016 at 4:16 AM Yurii Shevtsov  wrote:
>
>> I looked at sites you mentioned. I haven't any configuration feature on
>> johnlewis.ie Instead it provides an instruction for running special
>> shell script. But I much more liked original Rom-o-matic. I want do develop
>> same thing, but with fancier design, if you mind) I have more questions:
>>
>> How important this project for coreboot community is?
>>
>
> well, that's a tough question. Back in 2000 when we first did it, it was
> very important. At this point, coreboot is mostly two user communities:
> people who use it and don't know and don't care (chromebooks); and people
> who are dedicated hackers and know the insides so well they don't need
> rom-o-matic. Those who don't know anything and don't care they're using
> coreboot probably outnumber knowledgable people by about 10,000 to 1 at
> least [based on the 10m+ systems shipped at this point with coreboot, and
> my guess that the coreboot hacker community is unlikely to be as many as
> 1000 people).
>
> The number of people who don't know anything and can use a rom-o-matic is
> probably numbered in single digits, because even to use rom-o-matic you
> have to be knowledgable enough that you might as well build your own
> coreboot. You certainly have to have a path out if something goes wrong,
> and at that point you are cracking open your laptop. A failed coreboot
> install is not like a failed OS install. It's more like destroying your
> mainboard.
>
> I hate to be discouraging but my guess is at present that what john lewis
> is doing is probably as much as is needed.
>
>
> Do I have to fix some bugs or make any other sort of contribution, before
>> submitting my proposal?
>>
>
> You should show that you know how to build and use coreboot from scratch.
> It makes no sense to talk about rom-o-matic otherwise.
>
>
>> Do you have a proposal template or some special requirments for it?
>> What do you think about nodejs, as a backend?
>>
>
>
> I suspect you know far more about writing such a tool than I ever will,
> but far less about coreboot than you need to know. Your first step should
> be to get it, build it, and boot it in qemu; bonus points for doing it on
> real hardware.
>
> I think the choice of node.js is not nearly as important as ensuing you
> give people images that won't brick their machine.
>
> ron
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

[coreboot] [GSOC 2016] Coreboot panic room

[Antonello Dettori]

Hello everyone,
I am currently a computing student at the University of Abertay in 
Scotland and I would like to take part in this year GSOC.


The proposal that I want to submit concerns the coreboot panic room idea 
since it looks like something that could really benefit the project as a 
whole and also because it seems like a really interesting technical 
challenge.


I already digged through the blog and the mail exchanges of the previous 
developer working on this, Tadas Slotkus, and took a look at the patches 
that he wrote so I already have a general understanding of their aim.


I would like to concentrate most of my efforts towards improving and 
upstreaming the previous efforts, implementing a way to easily access 
the recovery mode when needed and further the integration between 
coreboot, serialICE and flashrom for this use-case.


Regarding the existing patches I would like to know if they would need 
to stay romc-compatible or should the scope be limited to CAR boards?


Any suggestions on how to tackle this project or GSOC in general?

Thanks in advance.

--
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Peter Stuge]

Hi guys,

(I have a server problem but am sortof recovering.)

ron minnich wrote:
> Do we have any record as to how uniform, e.g., the x220 experience
> has been?

This is an important question. It comes down to the mainboard of course.

Judging from the schematics and hardware in the wild there's not too
much variation with ThinkPads; the *only* significant variation in the
(admittedly very) old 60 generation is whether a mainboard has discrete
graphics or not, and even then that only matters for using the correct
option ROM.

There might be a bit more variation with newer generations because
they are more complex, but not too much.


//Peter

-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Alexander Couzens]

On Wed, 09 Mar 2016 22:30:44 +
ron minnich  wrote:

> That's great to hear. So it sounds like rom o matic would be handy.
> 
> That said, w.r.t. features. I'm not big on features. coreboot is a
> chance to brick your laptop. That means
> rom o matic should never ever ever give you a bad coreboot image.

> What would be cool would be to have a program they can run that
> creates a string, and the either you paste
> that string into a form or it connects and pastes it in (but what
> about DDOS?) and you get back a URL for an image
> that is known to work. Could the coreboot board status repo be used to
> determine this?
> 
> I think the messy part is that just because you have, say, an "x220",
> doesn't mean coreboot works on it, because PCs
> are never the same, even given the same number. So you need a very
> long signature that can confirm the coreboot
> running on everyone else's "x220" will work on yours.
> 
> Do we have any record as to how uniform, e.g., the x220 experience has
> been? I know the chromebook uniformity is pretty good.
They are uniform, beside the case of the spi. Most x220 have a spi in a
SOIC8, but some have a WSON or mlp8 case.

I had problems with my x220 because I desolered the wson chip and
soldered a SOIC8 which wasn't quite supported by coreboot. I say the
"normal" coreboot user wouldn't do that ;). They only use a clip.

Best,
lynxis
-- 
Alexander Couzens

mail: lyn...@fe80.eu
jabber: lyn...@fe80.eu
mobile: +4915123277221
gpg: 390D CF78 8BF9 AA50 4F8F  F1E2 C29E 9DA6 A0DF 8604


pgp2WgmQ4Yhkr.pgp
Description: OpenPGP digital signature
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

That's great to hear. So it sounds like rom o matic would be handy.

That said, w.r.t. features. I'm not big on features. coreboot is a chance
to brick your laptop. That means
rom o matic should never ever ever give you a bad coreboot image.

What would be cool would be to have a program they can run that creates a
string, and the either you paste
that string into a form or it connects and pastes it in (but what about
DDOS?) and you get back a URL for an image
that is known to work. Could the coreboot board status repo be used to
determine this?

I think the messy part is that just because you have, say, an "x220",
doesn't mean coreboot works on it, because PCs
are never the same, even given the same number. So you need a very long
signature that can confirm the coreboot
running on everyone else's "x220" will work on yours.

Do we have any record as to how uniform, e.g., the x220 experience has
been? I know the chromebook uniformity is pretty good.

ron

On Wed, Mar 9, 2016 at 2:07 PM Alexander Couzens  wrote:

> On Wed, 09 Mar 2016 20:57:23 +
> ron minnich  wrote:
>
> > I suspect you know far more about writing such a tool than I ever
> > will, but far less about coreboot than you need to know. Your first
> > step should be to get it, build it, and boot it in qemu; bonus points
> > for doing it on real
>
> I know a lot people who don't know or have problems compiling
> coreboot on their own or they don't want to spent hours to compiling the
> toolchain + coreboot. For these people rom-o-matic is a nice thing.
>
> German hacker culture likes coreboot a lot. Last coreboot user
> group meeting a journalist showed up. Someone spent her a x60 and she
> would like to have coreboot on it. But after disassembling the hardware
> down to the board itself, it turns out, she already got coreboot. I
> never looked at the x60 bios booting up :).
>
> There are also the cryptoparty people, they also support coreboot.
> zaolin is doing installation partys at his hackerspace, too. And may
> others do the same.
> --
> Alexander Couzens
>
> mail: lyn...@fe80.eu
> jabber: lyn...@fe80.eu
> mobile: +4915123277221
> gpg: 390D CF78 8BF9 AA50 4F8F  F1E2 C29E 9DA6 A0DF 8604
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Alexander Couzens]

On Wed, 09 Mar 2016 20:57:23 +
ron minnich  wrote:

> I suspect you know far more about writing such a tool than I ever
> will, but far less about coreboot than you need to know. Your first
> step should be to get it, build it, and boot it in qemu; bonus points
> for doing it on real

I know a lot people who don't know or have problems compiling
coreboot on their own or they don't want to spent hours to compiling the
toolchain + coreboot. For these people rom-o-matic is a nice thing.

German hacker culture likes coreboot a lot. Last coreboot user
group meeting a journalist showed up. Someone spent her a x60 and she
would like to have coreboot on it. But after disassembling the hardware
down to the board itself, it turns out, she already got coreboot. I
never looked at the x60 bios booting up :).

There are also the cryptoparty people, they also support coreboot.
zaolin is doing installation partys at his hackerspace, too. And may
others do the same.
-- 
Alexander Couzens

mail: lyn...@fe80.eu
jabber: lyn...@fe80.eu
mobile: +4915123277221
gpg: 390D CF78 8BF9 AA50 4F8F  F1E2 C29E 9DA6 A0DF 8604


pgpJWmFx26_Hk.pgp
Description: OpenPGP digital signature
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

On Wed, Mar 9, 2016 at 4:16 AM Yurii Shevtsov  wrote:

> I looked at sites you mentioned. I haven't any configuration feature on
> johnlewis.ie Instead it provides an instruction for running special shell
> script. But I much more liked original Rom-o-matic. I want do develop same
> thing, but with fancier design, if you mind) I have more questions:
>
> How important this project for coreboot community is?
>

well, that's a tough question. Back in 2000 when we first did it, it was
very important. At this point, coreboot is mostly two user communities:
people who use it and don't know and don't care (chromebooks); and people
who are dedicated hackers and know the insides so well they don't need
rom-o-matic. Those who don't know anything and don't care they're using
coreboot probably outnumber knowledgable people by about 10,000 to 1 at
least [based on the 10m+ systems shipped at this point with coreboot, and
my guess that the coreboot hacker community is unlikely to be as many as
1000 people).

The number of people who don't know anything and can use a rom-o-matic is
probably numbered in single digits, because even to use rom-o-matic you
have to be knowledgable enough that you might as well build your own
coreboot. You certainly have to have a path out if something goes wrong,
and at that point you are cracking open your laptop. A failed coreboot
install is not like a failed OS install. It's more like destroying your
mainboard.

I hate to be discouraging but my guess is at present that what john lewis
is doing is probably as much as is needed.


Do I have to fix some bugs or make any other sort of contribution, before
> submitting my proposal?
>

You should show that you know how to build and use coreboot from scratch.
It makes no sense to talk about rom-o-matic otherwise.


> Do you have a proposal template or some special requirments for it?
> What do you think about nodejs, as a backend?
>


I suspect you know far more about writing such a tool than I ever will, but
far less about coreboot than you need to know. Your first step should be to
get it, build it, and boot it in qemu; bonus points for doing it on real
hardware.

I think the choice of node.js is not nearly as important as ensuing you
give people images that won't brick their machine.

ron
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

I looked at sites you mentioned. I haven't any configuration feature on
johnlewis.ie Instead it provides an instruction for running special shell
script. But I much more liked original Rom-o-matic. I want do develop same
thing, but with fancier design, if you mind) I have more questions:

How important this project for coreboot community is?
Do I have to fix some bugs or make any other sort of contribution, before
submitting my proposal?
Do you have a proposal template or some special requirments for it?
What do you think about nodejs, as a backend?

Thanks in advance!

2016-03-09 6:24 GMT+02:00 ron minnich :

> yeah, we had something like this in the linuxbios days. I think you don't
> want to build it on demand, but rather have a bunch of pre-built images
> that are known good.
>
> Really, look at johnlewis.ie, that's the best thing I've seen. Have you
> also seen the original Rom-o-matic from the etherboot (now gpxe or ipxe)
> project?
>
> ron
>
> On Tue, Mar 8, 2016 at 7:35 PM Yurii Shevtsov  wrote:
>
>> Hello) I want to be a GSoC 2016 student. I really liked the idea of
>> ROM-O-Matic project. Also it's a nice name. So, as I understood, the end
>> user should get a website with a ROM configurator and a big 'Download'
>> button, right?
>>
>> I'm student of Odessa polytechnic university, computer science. I have a
>> good as for student experience in system and web programming
>> --
>> coreboot mailing list: coreboot@coreboot.org
>> https://www.coreboot.org/mailman/listinfo/coreboot
>
>
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] [GSoC 2016] ROM-O-Matic project

[ron minnich]

yeah, we had something like this in the linuxbios days. I think you don't
want to build it on demand, but rather have a bunch of pre-built images
that are known good.

Really, look at johnlewis.ie, that's the best thing I've seen. Have you
also seen the original Rom-o-matic from the etherboot (now gpxe or ipxe)
project?

ron

On Tue, Mar 8, 2016 at 7:35 PM Yurii Shevtsov  wrote:

> Hello) I want to be a GSoC 2016 student. I really liked the idea of
> ROM-O-Matic project. Also it's a nice name. So, as I understood, the end
> user should get a website with a ROM configurator and a big 'Download'
> button, right?
>
> I'm student of Odessa polytechnic university, computer science. I have a
> good as for student experience in system and web programming
> --
> coreboot mailing list: coreboot@coreboot.org
> https://www.coreboot.org/mailman/listinfo/coreboot
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

[coreboot] [GSoC 2016] ROM-O-Matic project

[Yurii Shevtsov]

Hello) I want to be a GSoC 2016 student. I really liked the idea of
ROM-O-Matic project. Also it's a nice name. So, as I understood, the end
user should get a website with a ROM configurator and a big 'Download'
button, right?

I'm student of Odessa polytechnic university, computer science. I have a
good as for student experience in system and web programming
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

Re: [coreboot] GSoC 2016

[Stefan Reinauer]

On 03/06/2016 10:35 AM, Tahir Ramzan wrote:

Respected Sir,

I am a MS CS scholar of Virtual University of Pakistan, I want to 
participate in GSoC 2016 for coreboot. Data Science, Networks, 
Information security, digital forensics and ethical hacking are my 
core areas of interest.


Currently, I am working on a research project on live forensics of GPU 
and volatile memories like RAMs and Caches.


I am looking forward your guidance to start my contribution for 
coreboot, thanks in anticipation.


Regards
Tahir Ramzan


Hi Tahir,

Thank you for your interest. Your background looks very interesting!

did you take the time to determine what you might be interested in 
working on, already?
Please take a look at our GSoC page at https://www.coreboot.org/GSoC and 
the ideas pages of the three projects we are hosting this year 
(coreboot, flashrom, serialice) and let us know if you have any questions


Stefan

--
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot

[coreboot] GSoC 2016

[Tahir Ramzan]

Respected Sir,

I am a MS CS scholar of Virtual University of Pakistan, I want to
participate in GSoC 2016 for coreboot. Data Science, Networks, Information
security, digital forensics and ethical hacking are my core areas of
interest.

Currently, I am working on a research project on live forensics of GPU and
volatile memories like RAMs and Caches.

I am looking forward your guidance to start my contribution for coreboot,
thanks in anticipation.

Regards
Tahir Ramzan
-- 
coreboot mailing list: coreboot@coreboot.org
https://www.coreboot.org/mailman/listinfo/coreboot