Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Alex G.]

On 02/01/2016 11:36 AM, Martin Roth wrote:
> What I don't read in that blog post is anything about the
> coreboot project being a democracy.

I'd normally find such a statement very disappointing, but I'm pretty
certain you're off on that one. I'm not sure how involved you've been in
coreboot public before leaving SAGE -- I've mostly known your name from
patches Marc submitted. What I can tell you with great confidence is
that the picture you try to portray of coreboot has historically not
been true.

I've been around for a long time, maybe too long, and I remember many an
issue having been publicly discussed. I remember cases where things were
brought up for discussion, and the so-called leadership team decided the
exact opposite of what it had intended. I remember the switch to git and
gerrit, and I remember seeing quite some back and forth on the issue.

I'm not sure why you have such a dark picture of coreboot in your mind.
I've haven't heard anyone from the so-called "coreboot leadership" team
make the claims you do. I'm sorry you feel that way, but that shouldn't
be a reason to shut down Paul's proposal

> Please note that these are just my own personal thoughts on these
> subjects

Exactly!

Alex

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Martin Roth]

I meant to send this to the mailing list on Saturday, but ended up just
sending it to Paul by accident.

Paul,
  This thread quickly followed one point that you commented on, but
has so far not touched much on one of your other issues.  This is the
issue of leadership of the coreboot community.  You might want to look
at Stefan's blog post "on coreboot leadership" from May of last year -
http://blogs.coreboot.org/blog/2015/05/11/on-coreboot-leadership/

The way that I read it is that anyone is welcome to contribute ideas
and opinions.  Stefan may poll people for their opinions at times as
well.  What I don't read in that blog post is anything about the
coreboot project being a democracy.  I think that Stefan makes it
pretty clear that the coreboot project has a leader, even though he
believes in a hands-off approach.

If people think that there are things that need to be discussed, I've
found that Stefan, Ron and the rest are very welcoming of reasonable
and constructive discussion.

Regarding announcements of changes to policy, maybe something could be
set up so that changes to certain wiki pages would automatically be
posted to the mailing list.

Please note that these are just my own personal thoughts on these
subjects, and are in no way trying to state any official coreboot
policy.

Martin


On Mon, Feb 1, 2016 at 11:30 AM, Alex Gagniuc  wrote:
> On Sun, Jan 31, 2016 at 2:24 PM, Stefan Reinauer
>  wrote:
>>
>> That is a really surprising statement coming from you, Alex, as you and I
>> have discussed this very topic in person several times
>
> And as I have said in those very same discussions, decisions about
> coreboot shold be done publicly. You're also portraying a distorted
> picture of what was actually discussed, but it was still a private
> conversation and has no bearing towards what Paul pointed out.
>
> Going back to Paul's proposal, he noted that an official project
> guideline had been modified with neither public mail discussion, nor
> community oversight. This was not a "minor typo fix" or
> "clarification". I count three different changes which were made this
> way. And I think Paul nailed it with his proposal.
>
> I fully support Paul's proposal.
>
> On Sat, Jan 30, 2016 at 6:54 PM, Martin Roth  wrote:
>> There was significant discussion in several
>> meetings about the reasons for and against standardizing on AT
>> syntax.
>
> As I've explained above, private conversations are not the proper
> forum to make decisions related to coreboot. I realize you and others
> involved are wearing two hats, and sometimes it's hard to tell which
> hat you're wearing, either for you, or observers. Please consider the
> image portrayed on your employer, when a group of its employees
> unilaterally discusses, changes, then enforces rules in a public
> project. I think Paul's proposal fixes this issue.
>
>> If you have a reason for using Intel syntax that is really more
>> persuasive than keeping the asm code in the project consistent, feel
>> free to state it.
>
> While I have a lot to say of the matter, this is not the appropriate
> place. This discussion is about a change to a policy, not a
> development guideline. Let's focus on what Paul is saying here.
>
> Alex

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Alex Gagniuc]

On Sun, Jan 31, 2016 at 2:24 PM, Stefan Reinauer
 wrote:
>
> That is a really surprising statement coming from you, Alex, as you and I
> have discussed this very topic in person several times

And as I have said in those very same discussions, decisions about
coreboot shold be done publicly. You're also portraying a distorted
picture of what was actually discussed, but it was still a private
conversation and has no bearing towards what Paul pointed out.

Going back to Paul's proposal, he noted that an official project
guideline had been modified with neither public mail discussion, nor
community oversight. This was not a "minor typo fix" or
"clarification". I count three different changes which were made this
way. And I think Paul nailed it with his proposal.

I fully support Paul's proposal.

On Sat, Jan 30, 2016 at 6:54 PM, Martin Roth  wrote:
> There was significant discussion in several
> meetings about the reasons for and against standardizing on AT
> syntax.

As I've explained above, private conversations are not the proper
forum to make decisions related to coreboot. I realize you and others
involved are wearing two hats, and sometimes it's hard to tell which
hat you're wearing, either for you, or observers. Please consider the
image portrayed on your employer, when a group of its employees
unilaterally discusses, changes, then enforces rules in a public
project. I think Paul's proposal fixes this issue.

> If you have a reason for using Intel syntax that is really more
> persuasive than keeping the asm code in the project consistent, feel
> free to state it.

While I have a lot to say of the matter, this is not the appropriate
place. This discussion is about a change to a policy, not a
development guideline. Let's focus on what Paul is saying here.

Alex

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Stefan Reinauer]

* Alex G.  [160131 00:05]:
> I conclude from these, that your assertion that this has been an
> unspoken rule, is not true. Furthermore, I believe that this arbitrary
> change was done as an act of spite towards a set of engineers. Also,
> since you, and the rest of the coreboot leadership have shown a pattern
> of first discussing changes to the project publicly, on the mailing
> list, I conclude that a discussion about this issue was deliberately
> avoided in order to prevent those engineers, as well as other coreboot
> community members, from presenting their arguments.

That is a really surprising statement coming from you, Alex, as you and I
have discussed this very topic in person several times, and I have
explained to you why I initially checked in Scott's patches with mixed
syntax and we both had agreed that in the long run the code has to be
rewritted in AT syntax. So now you are using the fact that there is no
log on the mailing list about this discussion to hurt the project and
create some sort of artificial divide in the community once again.

I am really disappointed in you.

  Stefan

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[ron minnich]

The requirement for ATT syntax was set informally in 2001, when I had some
partners from U. Md. who were advocating for Intel syntax. We discovered
that having two syntaxes is unworkable.

>From that time we assumed that everyone would use a common syntax. It
certainly never occurred to us that we had to say it explicitly, since it
seems obvious that uniform assembly syntax is a good idea, and that the
syntax we currently use, and that our tools use, should determine the
assembly syntax for new code.

The change to the guidelines is hence a codification of a practice that
goes back to the project's beginnings.

thanks

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Patrick Georgi via coreboot]

2016-01-30 15:26 GMT+01:00 Paul Menzel :
> But then, I wondered why I was not aware of that section in the
> development guidelines [2], and wanted to read up on it. While at it, I
> also looked through the history, and there I see, that it was only
> added [3] on the same day.
We had mentions of intel assembler syntax on the list over the decades
(all 1.5 of them, proposed starting point for web based searches:
site:www.coreboot.org/pipermail/coreboot/ -gerrit intel syntax), and
it was always clear that coreboot uses AT syntax for x86.
The collective consciousness (with apologies to Durkheim) was pretty
consistent here.

The reason why this was added now is that people insisted that as long
as there's no written down hard rule about it, everything is fair
game.
Personally, I'm not a huge fan of this model, but after the power
games surrounding this topic already took forever (may include traces
of hyperbole), it was the most economic decision to just nail it down.

>4. If there is objection, and no agreement can be reached, the
>   proposal(s) should go up for a vote. The vote period should be at
>   least one week.
There exists no electorate (or to state it more constructively: who
gets to vote?)

> No idea, but I’d think it’d be only fair to revert the changes made
> this months, and discuss them first.
With all the time already wasted on the assembly syntax issue (and no
chance for a different outcome), and everything else looking like
context-preserving clarifications (eg. GPL -> GPLv2) or harmless
updates (bug tracker) I don't see a need for that.
This also assumes both that your proposal takes effect, and is
effective retroactively.

The main problem here (if any) was the lack of notification after
changing the document. Given the limited impact (there were no other
instances of .intel_syntax out there, and folks generally knew to use
AT syntax), it probably wasn't considered a big deal.


Patrick
-- 
Google Germany GmbH, ABC-Str. 19, 20354 Hamburg
Registergericht und -nummer: Hamburg, HRB 86891, Sitz der Gesellschaft: Hamburg
Geschäftsführer: Matthew Scott Sucherman, Paul Terence Manicle

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

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Alex G.]

On 01/30/2016 11:30 AM, ron minnich wrote:
> The change to the guidelines is hence a codification of a practice that
> goes back to the project's beginnings.

I find that statement inaccurate. This practice had not been an issue in
the past, on patches that you yourself approved. Mixed syntaxes have
been present in the codebase for a while. See EXHIBIT A.

EXHIBIT B shows a patch, mid last year, when a new intel_syntax code was
introduced. There was only one comment (following) about the syntax. The
patch was approved by you, Ron, without mentioning what you claim has
been a practice.

> Patrick Georgi, Jul 13, 2015
> follow up patch: we don't really want two _syntaxes in one file_, right?
(emphasis added)

The comment clearly addresses the issue of mixing two syntaxes within a
file. It does not mention "a practice that goes back to the project's
beginnings", nor does it mention that either "AT syntax is to be used
through-out the project", or "No new Intel syntax code is allowed into
the project."


EXHIBIT C, shows another such patch that, you, have also approved, Ron.
The comment got no replies.

> Ronald G. Minnich, Oct 30 10:43 AM
> why intel syntax _in a file that is att syntax_? it's a real
oportunity for a buggy time.+2'ing because you promised me you're
rewrite it.
(emphasis added)

Notice how the issue here is strictly the mixing of syntaxes within the
same file.

EXHIBIT D shows another such patch, this time approved by me. I have
made a comment about the issue of mixed syntax.

> Alexandru Gagniuc, Jul 17, 2015
> Really? Mixed syntax in the same file?

This also addresses strictly the issue of mixing of syntaxes within the
same file.

EXHIBIT E shows another patch, that you yourself approved. This time,
there was not even a mention of mixed syntaxes.


I conclude from these, that your assertion that this has been an
unspoken rule, is not true. Furthermore, I believe that this arbitrary
change was done as an act of spite towards a set of engineers. Also,
since you, and the rest of the coreboot leadership have shown a pattern
of first discussing changes to the project publicly, on the mailing
list, I conclude that a discussion about this issue was deliberately
avoided in order to prevent those engineers, as well as other coreboot
community members, from presenting their arguments.

Alex


### EXHIBIT A: Presence of .intel_syntax ###

* 4a30ab9 cpu/amd: remove .intel_syntax
* 0302b06 arch/x86: remove .intel_syntax
* c7b2b7c cbfstool: Fix potential error when using hash attribute

$ git reset --hard c7b2b7c67dc8afb52c3dc8e9297e5ed81fa22674


$  git grep intel_syntax
src/arch/x86/boot.c:".intel_syntax noprefix\n\t"
src/arch/x86/c_start.S:.intel_syntax noprefix
src/arch/x86/wakeup.S:  .intel_syntax noprefix
src/cpu/amd/agesa/cache_as_ram.inc:  .intel_syntax noprefix
src/cpu/amd/pi/cache_as_ram.inc:  .intel_syntax noprefix


### EXHIBIT B: src/arch/x86/[boot.c, c_start.S] ###

$ git blame src/arch/x86/boot.c |grep intel_syntax
9693885a src/arch/x86/boot/boot.c  (Stefan Reinauer 2015-06-18 01:23:48
-0700  63)  ".intel_syntax noprefix\n\t"

$ git blame src/arch/x86/c_start.S |grep intel_syntax
9693885a src/arch/x86/lib/c_start.S  (Stefan Reinauer
2015-06-18 01:23:48 -0700 403) .intel_syntax noprefix

$ git show 9693885a
commit 9693885ad88d21ead7bd9ebc32f3e4901841b18b
Author: Stefan Reinauer 
Date:   Thu Jun 18 01:23:48 2015 -0700

x86: Port x86 over to compile cleanly with x86-64

Change-Id: I26f1bbf027435be593f11bce4780111dcaf7cb86
Signed-off-by: Stefan Reinauer 
Signed-off-by: Scott Duplichan 
Reviewed-on: http://review.coreboot.org/10586
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand

Reviewed-by: Ronald G. Minnich 

### EXHIBIT C: src/arch/x86/wakeup.S ###

$ git blame src/arch/x86/wakeup.S |grep intel_syntax
ac901e6b src/arch/x86/wakeup.S   (Stefan Reinauer 2015-07-31
16:46:28 -0700  32).intel_syntax noprefix

$ git show ac901e6b
commit ac901e6bedc98d115ebb8089afd8f67aef39c000
Author: Stefan Reinauer 
Date:   Fri Jul 31 16:46:28 2015 -0700

wakeup: Switch back to 32bit mode first

On x86_64 we need to leave long mode before we can switch to 16bit
mode. Oh joy! When's my 64bit resume pointer coming?

Why didn't this get caught earlier? Seems the Asrock E350M2 didn't
do Suspend/Resume?

Yes, I know it's Intel syntax. Will be converted to AT syntax
as soon as the whole thing actually works.. 8)

Change-Id: Ic51869cf67d842041f8842cd9964d72a024c335f
Signed-off-by: Stefan Reinauer 
Reviewed-on: http://review.coreboot.org/11106
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich 


### EXHIBIT D: src/cpu/amd/agesa/cache_as_ram.inc ###

$ git blame 

Re: [coreboot] [RFC] Proposal for policy for changing the development guidelines

[Martin Roth]

Alex,
  Please stop already.  We know you don't agree with the decision.
Stefan has agreed privately that he shouldn't have submitted those,
and that it set a bad precedent.  As you say in your email, *YOU* even
questioned it when he did it, and he agreed that he would change them
from Intel syntax to AT syntax.

The "change" WAS to formalize an unwritten rule that had been broken a
few times in the past. There was significant discussion in several
meetings about the reasons for and against standardizing on AT
syntax.  Many of us, including myself, learned x86 assembly using
Intel syntax, and find it easier to read.  Mixing Intel syntax with
AT syntax is just confusing and seems like bad practice.  Because of
this, the decision was made to go with AT syntax only.   It was NOT
done to spite you, whatever you might believe.

If you have a reason for using Intel syntax that is really more
persuasive than keeping the asm code in the project consistent, feel
free to state it.  Otherwise, PLEASE let's move on.

Martin

On Sat, Jan 30, 2016 at 5:44 PM, ron minnich  wrote:
>
>
> On Sat, Jan 30, 2016 at 3:06 PM Alex G.  wrote:
>>
>>  Furthermore, I believe that this arbitrary
>> change was done as an act of spite towards a set of engineers.
>
>
> Well, wow. This just got weird. I think I'm done with this discussion.
>
> ron
>
> --
> coreboot mailing list: coreboot@coreboot.org
> http://www.coreboot.org/mailman/listinfo/coreboot

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