I wrote an article about Loongson and Free Software

[Zhang Le]

This article will be published on China Computer World magazine.
( http://www.ccw.com.cn ) So it is written in Chinese.

Because I have to submit it today, I am afraid there is no time to translat
it to English. I am sorry I should've posted it here earlier.

But, anyway, before it gets published, I hope everyone could take a look and
give me some advices if you have any.

Thanks!

--
Zhang, Le
Gentoo/Loongson Developer
http://zhangle.is-a-geek.org
0260 C902 B8F8 6506 6586 2B90 BC51 C808 1E4E 2973

[yajin]

An excellent article on the current situations and issues of the community of loongson.

It seems that lemote or ICT has no interest in cooperating with the community. They prefer playing with loongson their own than sharing with the community. In fact it is the community who promote the software of loongson, but lemote/ICT does not want to hear the sounds from the community.

There should be a person in lemote/ICT who is responsible communicating with and listening from the community. They should donate or help the open source projects related with loongson. Moreover, they should push their code into the upstream, not letting these codes aside with the main line.

The N32 port of debian should be a big gift from the community. I hope lomote/ICT can learn the importance of the free software in loongson's development from it.


yajin

http://vm-kernel.org

2009/6/29 Zhang Le <r0b...@gentoo.org>

[Ruan Beihong]

Some pieces of advise:
I think the whole article is talking about current situation about the
free software and the Loongson development, but the title is pointing
to the strategies. So I think more attention should be paid to the
comparisons of the current situation and the ideal strategies, in
which way Ubuntu goes. And these comparisons would better be
enumerated.
e.g. 'About the Code Survival' can talk about the customs in free
software development (coding style, why and how codes get into the
upstream);
'About the Community' can talk about the community stuff (how big
power is in store in the community, how can a company get well
involved in the community even when the company itself is the power
pushing the development forward).
Based on what I said above, I think a reorganization is needed.

After all, the points you've pointed is really important, I hope they
could be presented in a better way.

If any mistakes in my English, please correct them for me. I've just
done the CET-4, hoping to pass it.

2009/6/29 yajin <yaji...@vm-kernel.org>:

--
James Ruan

[Zhang Le]

First of all, thanks for your advice.

On 17:27 Mon 29 Jun , Ruan Beihong wrote:
>
> Some pieces of advise:
> I think the whole article is talking about current situation about the
> free software and the Loongson development, but the title is pointing
> to the strategies. So I think more attention should be paid to the
> comparisons of the current situation and the ideal strategies, in
> which way Ubuntu goes. And these comparisons would better be
> enumerated.
> e.g. 'About the Code Survival' can talk about the customs in free
> software development (coding style, why and how codes get into the
> upstream);
> 'About the Community' can talk about the community stuff (how big
> power is in store in the community, how can a company get well
> involved in the community even when the company itself is the power
> pushing the development forward).
> Based on what I said above, I think a reorganization is needed.

I think I just did that. I summarized 3 points in the second half of the
article. The first is about community, the second is about making use of
existing free software, the third is about working with upstream.

I just didn't go very deep because the audience may not be able to understand
some technical details. Also I think the first step to go is to make lemote
understand what to do. After that, we can talk in more detail about how to do.

So what kind of reorganization are you suggesting exactly? However, given the
time constraint, I am afraid it is too late for some big modifications.

[Ruan Beihong]

Let me show you an example:
第三，龙芯相对欠缺与国际自由软件社区主动交流的行动和意识。自由软件的一个好处是谁都可以得到源代码来学习和改进。例如逸珑8089笔记本的温度与风扇监控模块在最新版本的Linux内核上工作不正常，我们只有自己去做相应的修改。如果模块进入了内核，我们就不用做这个事情，当初引入这个改动的人会负责把内核中所有需要改动的地方进行修改。

You'd better say the importance of going upstream first than expend
the example. Many people may doubt about '但改进只有被上游所收录之后才
能真正的存活下去'. It should be better explained at first shot. Many people
can't understand the productive cycle in free software.

To many people who really understand the free software, what you
talked is good, but to many other people, who can't understand some
basic concepts of the free software, the first sight of 'Community',
'Upstream', 'Contribute' is very weird. So I think it should be talked
in more understandable way: show the concepts and explain it well,
than do the comparison , talk about what Loongson is now doing and
what should be done.
We understand things in the order of What, How, Why, and the order
should be apply to introduction.

2009/6/29 Zhang Le <r0b...@gentoo.org>:

--
James Ruan

[Zhang Le]

On 18:52 Mon 29 Jun , Ruan Beihong wrote:
>
> Let me show you an example:
> 第三，龙芯相对欠缺与国际自由软件社区主动交流的行动和意识。自由软件的一个好处是谁都可以得到源代码来学习和改进。例如逸珑8089笔记本的温度与风扇监控模块在最新版本的Linux内核上工作不正常，我们只有自己去做相应的修改。如果模块进入了内核，我们就不用做这个事情，当初引入这个改动的人会负责把内核中所有需要改动的地方进行修改。
>
> You'd better say the importance of going upstream first than expend
> the example. Many people may doubt about '但改进只有被上游所收录之后才
> 能真正的存活下去'. It should be better explained at first shot. Many people
> can't understand the productive cycle in free software.
>
> To many people who really understand the free software, what you
> talked is good, but to many other people, who can't understand some
> basic concepts of the free software, the first sight of 'Community',
> 'Upstream', 'Contribute' is very weird. So I think it should be talked
> in more understandable way: show the concepts and explain it well,
> than do the comparison , talk about what Loongson is now doing and
> what should be done.
> We understand things in the order of What, How, Why, and the order
> should be apply to introduction.

Actually, what you now see is a compromise between me and the editor. And the
compromise is necessary because there are some problems of personal taste and
there is constraint on the number of words.

I did exactly some point which you have made, but it was later modified by the
editor. To show some respect, I'd better not reverse the changes the editor has
made, as long as there is no obvious errors. After all, this article is going to
be published on a magazine, not on a personal blog. What do you think?

I have attached my original article for your information.

[Ruan Beihong]

It will be a hard problem if there is constraints on the number of words.

But I think using the comments and links in original article is a good
way to provide information. I feel sorry that it didn't appeal the
editor. (Something too technical is improper though, the comments is
well-focused.)

2009/6/29 Zhang Le <r0b...@gentoo.org>:

--
James Ruan

[Ruan Beihong]

appeal the editor => appeal to the editor :)

2009/6/29 Ruan Beihong <ruanb...@gmail.com>:

--
James Ruan

[Deng Xiyue]

As I understand it, it is still too large a topic to be fit within the
word limit :) The majority are yet to understand the free software
community, the Loongson and MIPS issue, and hence has question about the
future of the combination. Such information may well fill a whole
stand-alone website. IMHO, Zhang's article is an introduction, and the
beginning. What I suggest is Mr. Zhang may provide several link to
related resources, so interested users may find additional information.

Regards,
Deng Xiyue

[Eric Fisher]

2009/6/29 yajin <yaji...@vm-kernel.org>:

> It seems that lemote or ICT has no interest in cooperating with the
> community. They prefer playing with loongson their own than sharing with the
> community. In fact it is the community who promote the software of loongson,
> but lemote/ICT does not want to hear the sounds from the community.

> There should be a person in lemote/ICT who is responsible communicating with
> and listening from the community. They should donate or help the open source
> projects related with loongson. Moreover, they should push their code into
> the upstream, not letting these codes aside with the main line.

Actually, I'm from ICT, compiler group. I'm glad to discuss with us
about free software projects.
And hope I can give any help.

> The N32 port of debian should be a big gift from the community. I hope
> lomote/ICT can learn the importance of the free software in loongson's
> development from it.

I will let our group and hardware group to know what is going on in
the community. :-)

Eric

[Wu Zhangjin]

Hi,

On Mon, 2009-06-29 at 16:13 +0800, yajin wrote:
> An excellent article on the current situations and issues of the
> community of loongson.
>

That is!

> It seems that lemote or ICT has no interest in cooperating with the
> community. They prefer playing with loongson their own than sharing
> with the community. In fact it is the community who promote the
> software of loongson, but lemote/ICT does not want to hear the sounds
> from the community.
>

Sorry, I can not completely agree with you on this aspect :-)

Lemote really have interest in cooperating with the community and have
done some jobs:

1. loongson2e support have already been in the official community.
2. about 4 different versions of loongson2f support have been pushed
into the community, several patches are accepted, although not
completely accepted yet.
3. STD(Hibernation) support for mips have already been in linux-mips
4. about 3 different versions of ftrace(kernel tracing) support for mips
have been released out(hope the 4th version will be accepted).
5. several different versions of RT_PREEMPT(real time patch) for mips
have been released, and it will be accepted in not long future.

and currently, several reasons impact on the cooperation between Lemote
and community:

1. as a not-big company, in the short term, it mainly focus on products,
make them more stable and acceptable to the users, so, it do not have
enough extra employees working on cooperating with community.

but, in the long term, the situation should be changed, because the
cooperation with the community will also help to improve the products,
and it is obvious that as a company using free softwares, it should
contribute it's own source code to the community, not only demanding,
but also, contributing!

2. as a young company, it do not have enough experience on cooperating
with community, but this seems a pretext: why not employ some guys who
have experiences? this may increase the cost of the company. and why not
let the current employees communicate with the community? perhaps this
is a good idea, there is a need to encourage them changing working
approaches: not only play with the source code, but also communicate
with other 'programmer'(hackers!) in the community, then they will grow
up quickly and the company will also benefit from it!

So, I think some plans should be made to improve the current situation:
"no plan no gain", and "no action no gain"!

> There should be a person in lemote/ICT who is responsible
> communicating with and listening from the community. They should
> donate or help the open source projects related with loongson.
> Moreover, they should push their code into the upstream, not letting
> these codes aside with the main line.

for the present, I'm working on pushing the loongson patches to the
upstream.

>
> The N32 port of debian should be a big gift from the community. I hope
> lomote/ICT can learn the importance of the free software in loongson's
> development from it.
>

That is! and I will launch another N32 project this summer holiday:
archlinux for loongson:
http://groups.google.com/group/archlinux-for-loongson

this is also a free software project from community:
http://oss.lzu.edu.cn

Regards,
Wu Zhangjin

[yanhua]

Wu Zhangjin 写道:

> Hi,
>
> On Mon, 2009-06-29 at 16:13 +0800, yajin wrote:
>
>> An excellent article on the current situations and issues of the
>> community of loongson.
>>

What is the article? where can we get it?

[Zhang Le]

On Jun 30, 9:09 am, Eric Fisher <joefoxr...@gmail.com> wrote:
> I will let our group and hardware group to know what is going on in
> the community. :-)

Everyone who's work has anything to do with Loongson would be a member
of the community.
Sadly, now we don't have any well-known, official, authoritative
channel to publish news in the community.

However, I think this mailing list has the potential to become one.

So, if any of you get any news or new progress on your work, please
send an email here to share with all others.

And if you don't want to miss any such news, please join this mailing
list.

Treat yourself as a member of the COMMUNITY.

Thanks!

[Zhang Le]

On Jun 30, 10:01 am, Wu Zhangjin <wuzhang...@gmail.com> wrote:
> > It seems that lemote or ICT has no interest in cooperating with the
> > community. They prefer playing with loongson their own than sharing
> > with the community. In fact it is the community who promote the
> > software of loongson, but lemote/ICT does not want to hear the sounds
> > from the community.
>
> Sorry, I can not completely agree with you on this aspect :-)
>
> Lemote really have interest in cooperating with the community and have
> done some jobs:
>
> 1. loongson2e support have already been in the official community.
> 2. about 4 different versions of loongson2f support have been pushed
> into the community, several patches are accepted, although not
> completely accepted yet.
> 3. STD(Hibernation) support for mips have already been in linux-mips
> 4. about 3 different versions of ftrace(kernel tracing) support for mips
> have been released out(hope the 4th version will be accepted).
> 5. several different versions of RT_PREEMPT(real time patch) for mips
> have been released, and it will be accepted in not long future.

This is great demonstration of the consequence of lack of
communication with community.
As a Linux/MIPS subscriber, I surely know what lemote (especially Wu
Zhangjin) has done.
But others may not know, unless they have subscribed to Linux/MIPS
mailing list, too.

So, please, whenever you have done something or encounter any
questions, please share it with others.
This is what free software all about.

>
> and currently, several reasons impact on the cooperation between Lemote
> and community:
>
> 1. as a not-big company, in the short term, it mainly focus on products,
> make them more stable and acceptable to the users, so, it do not have
> enough extra employees working on cooperating with community.
>
> but, in the long term, the situation should be changed, because the
> cooperation with the community will also help to improve the products,
> and it is obvious that as a company using free softwares, it should
> contribute it's own source code to the community, not only demanding,
> but also, contributing!

Focusing on products is not mutually exclusive with cooperating with
the community.
There is a good example I mentioned in my article, the LXDE example.

And it seems to me that lemote always worries about the company
itself, not thinking of the big picture of the community.
Note, it is just my feeling. If it is not the case, I apologize.

>
> 2. as a young company, it do not have enough experience on cooperating
> with community, but this seems a pretext: why not employ some guys who
> have experiences? this may increase the cost of the company. and why not
> let the current employees communicate with the community? perhaps this
> is a good idea, there is a need to encourage them changing working
> approaches: not only play with the source code, but also communicate
> with other 'programmer'(hackers!) in the community, then they will grow
> up quickly and the company will also benefit from it!
>
> So, I think some plans should be made to improve the current situation:
>        "no plan no gain", and "no action no gain"!

Encouraging employees to communicate with the community surely is the
right way to go.
But then what? The community may find that the employees may not be as
smart/powerful as they would've imagined before.
They get discouraged.

A healthy free software community (backed by a commercial company)
works as follows (at the very least IMO):

The company provide all the necessary platforms for the community to
work on: mailing list, planet, bug tracker, wiki, forum, irc, and make
them well known to everyone in the community, make them play the role
they are supposed to play.

The company should help the community to function properly: establish
an administration team, do proper marketing, make the achievement
accomplished in the community widely known to everyone, attract new
members to join

The company should hire employees from the community, only hire the
smartest one, the one who many people dreamed to work with. So that
every potential employees could see a bright future, and work hard
towards their goals.

[Zhang Le]

On Jun 30, 10:44 am, yanhua <y...@lemote.com> wrote:
> Wu Zhangjin 写道:> Hi,
>
> > On Mon, 2009-06-29 at 16:13 +0800, yajin wrote:
>
> >> An excellent article on the current situations and issues of the
> >> community of loongson.
>
> What is the article? where can we get it?

Please check the attachment of my first email in this thread.

[Wu Zhangjin]

On Tue, 2009-06-30 at 00:27 -0700, Zhang Le wrote:
> A healthy free software community (backed by a commercial company)
> works as follows (at the very least IMO):
>
> The company provide all the necessary platforms for the community to
> work on: mailing list, planet, bug tracker, wiki, forum, irc, and make
> them well known to everyone in the community, make them play the role
> they are supposed to play.
>
> The company should help the community to function properly: establish
> an administration team, do proper marketing, make the achievement
> accomplished in the community widely known to everyone, attract new
> members to join
>
> The company should hire employees from the community, only hire the
> smartest one, the one who many people dreamed to work with. So that
> every potential employees could see a bright future, and work hard
> towards their goals.

Very good suggestion, this is really a good guideline for building a
FOSS community.

Regards,
Wu Zhangjin

[Frederic Muller]

There are actually at least 2 commercial companies using the Loongson 2F chip at the moment and more coming. Hopefully in the future we'll see a diversity of hardware platforms with all different implementations and features.

For what it's worth, OLPH already has all those community elements (mailing list, planet, bug tracker, wiki, forum, irc) and closely collaborate and gets support from Lemote and ICT when needed. In fact we have worked very closely on many items such as bi-lingual documentation for Pmon, STR and STD, kernel fixing, frequency scaling. All the results of this work are already available to the community in different products (read different OS) and source code is of course available on at least 2 repositories.

In fact both companies do provide hardware to developers at cost or for free based on a few factors, and also provide compiling farms for those who do not own the platform (or cannot afford it), technical support, forums. The only thing you could argue about is that Lemote is using a non-Open Source forum platform with a buggy RSS feed system. Fuxin, the engineer in charge of maintaining their forum and myself have personally discussed about migrating to phpBB and there are 2 issues: moving the Discuz! posts history in phpBB and training of the engineer to a new tool. This takes time.

Never the less, we always welcome criticism and suggestion on how to make our communication and support better for the community members.I hope that shows in the response from Lemote and myself.

Should you need further information on where to get all the resources I think the best way is to ask what you need on this list.

Thanks and happy hacking!

Fred

[yanhua]

Frederic Muller 写道:

Yes. If you can provide all the information in a website, this will be
more convenient for us and the community.

Any way, I agree with what you say, and we will pay more attention to
the community.

>
> Thanks and happy hacking!
>
> Fred
>
>
>
> Wu Zhangjin wrote:
>> On Tue, 2009-06-30 at 00:27 -0700, Zhang Le wrote:
>>
>>> A healthy free software community (backed by a commercial company)
>>> works as follows (at the very least IMO):
>>>
>>> The company provide all the necessary platforms for the community to
>>> work on: mailing list, planet, bug tracker, wiki, forum, irc, and make
>>> them well known to everyone in the community, make them play the role
>>> they are supposed to play.
>>>
>>> The company should help the community to function properly: establish
>>> an administration team, do proper marketing, make the achievement
>>> accomplished in the community widely known to everyone, attract new
>>> members to join
>>>
>>> The company should hire employees from the community, only hire the
>>> smartest one, the one who many people dreamed to work with. So that
>>> every potential employees could see a bright future, and work hard
>>> towards their goals.
>>>
>>
>> Very good suggestion, this is really a good guideline for building a
>> FOSS community.
>>
>> Regards,
>> Wu Zhangjin
>>
>>
>> >>
>>
>>

--
晏华

[Zhang Le]

No wonder I didn't receive Fred's email, it is in HTML format, not plain text.
And I have a rule in my .procmailrc which filters HTML email, because normally
there are only plain text emails in open source project's mailing list.

I want to explain a little bit about what I wrote about how "A healthy free
software community (backed by a commercial company) works". I didn't really mean
any specific company here. You can replace "a company" with "a group of
companies".

Admittedly, Emtec is doing very well in building gdium's own community. However,
what I have in mind is something bigger. The community of all participants in the
Loongson ecosystem, including but not limited to: ICT, MTI, ST, lemote, Emtec,
gnewsense, debian, gentoo, archlinux, developers and contributors not in the
aforementioned organizations and end users of all kinds of Loongson based
products.

[Frederic Muller]

Dear Robert,

So what do you exactly expect from those players? What in your view is
missing for you to "do your community job"?

I'm genuinely curious, you know I am 100% a community person and Open
Source. Today considering the size of our (loongson) community I find
it's not doing bad. In fact I believe the participation/ownership ration
is pretty high in fact (as opposed to other communities in general).

Thank you.

Fred

[Zhang Le]

On 14:31 Thu 09 Jul , Frederic Muller wrote:
> Dear Robert,
>
> So what do you exactly expect from those players? What in your view is
> missing for you to "do your community job"?
>
> I'm genuinely curious, you know I am 100% a community person and Open
> Source. Today considering the size of our (loongson) community I find
> it's not doing bad. In fact I believe the participation/ownership ration
> is pretty high in fact (as opposed to other communities in general).

http://www.testlab.com.cn/html/qiyejisuan/20090702/1293.html

I've already summed up 3 points in my article, although 2 of them are actually
only for lemote. Honestly, the target audience of this article is ICT/lemote, or
specifically Hu Weiwu/Zhang Fuxin.

Taking the first point for example, I think we should have a roadmap, a todo
list, so that if someone want to help, he can know how to help[1]. In my
article, I already tried to summarize the current situation, although I forget
to mention the current situation of Xorg's support in both xorg-server and
video card drivers. Xorg-server now requires a patch which contains hardcoded
address, which in turn means it will never be accepted. And the most decent
solution may be to implement ioperm and/or iopl in glibc for MIPS[2]. Also in
lemote's yeeloong, the video card's driver shipped in the Loonux is from the
card manufacture, which is based on an out-dated xorg driver, which does not
support libpciaccess, and that means it can't be used with xorg-server-1.5 and
later version. xorg's silimotion driver previously does not work on yeeloong.
I helped to fix it[3]. And yet it still has another problem, i.e. dual head not
working because 16bit depth could cause X to freeze under certain circumstance[4].

But even after I did all these, my work still can't get any recognition. At
least this is my feeling. I am nobody except a "龙芯爱好者", or maybe "资深龙芯
爱好者". I am just a volunteer. Nobody would listen to me. I am not asking for
any money compensation. I just want my not-so-trivial work would be recognized
by the beneficiary of my work. Taking the siliconmotion's driver for example,
I have tried two different ways to solve it. The first is make the shipped
driver support libpciaccess, but failed. The half work is here[5]. The second is
make xorg's driver work, and I succeeded. This work took me several weeks. Just
like other works I have done in the past several months, e.g. N32 mozilla
firefox.
http://r0bertz.blogspot.com/2009/01/mips-n32-patch-for-xulrunnerfirefox.html
http://r0bertz.blogspot.com/2009/03/firefox-on-mips.html
http://r0bertz.blogspot.com/2009/04/mips-n32-firefox-in-acid3-test.html
http://r0bertz.blogspot.com/2009/05/mips-n32-firefox-35-beta4.html

Frankly speaking, I am now very discouraged. And I will need to find a full time
job. But don't worry, I will not give up on Loongson, I just can't promise as
much time as before.

[1] http://www.gentoo-cn.org/trads/
[2] http://www.linux-mips.org/archives/linux-mips/2009-02/msg00040.html
[3] http://bugs.freedesktop.org/show_bug.cgi?id=21528
[4] http://bugs.freedesktop.org/show_bug.cgi?id=21622
[5] http://www.gentoo-cn.org/gitweb/?p=siliconmotion.git;a=summary

[Frederic Muller]

Hi!

I wrote a 10 pages answer but that was way too long, so the short version:
1. Roadmap of?
2. Community help? my experience is it is hard to tell volunteer what to
do, they always want to go their own way (which I understand - and which
helps to balance visions)
3. SMI drivers: not ICT/Lemote's fault. We've all (Lemote, Mandriva,
Dexxon) been asking for a solution for over a year, meet so many times,
emails, phone calls. And we're still doing it.
4. A lot of xorg things don't work as well on x86. Do I blame Intel (I
have an ATI card...)?
5. Lemote (and Dexxon) ships a platform with an OS to do a specific job.
Arguably I'd say it works ok to do that job. Now you want to change
things, and Linux, the documentation provided and etc allows you to do
so and it's great! Your changes might not work but you shouldn't really
blame us for that (especially if it's a video driver issue....)

I think you're confusing responsibilities and feel frustrated because
things move slowly. Well we're even more frustrated but we've learned to
be patient and definitely appreciate the work and the participation of
the community following us.

Wishing you luck finding a job.

Fred

[Zhang Le]

On 16:28 Thu 09 Jul , Frederic Muller wrote:
> I wrote a 10 pages answer but that was way too long, so the short version:
> 1. Roadmap of?

GNU/Linux system support for Loongson:
1. kernel
2. glibc
3. xorg
4. multimedia libraries and applications.

For xorg, I still want to add something. Now intel and ati has got KMS support.
What about Loongson system's video cards?

Loongson 2G has video chips from Vivante, do they have kernel/xorg drivers?
Will they push their drivers to upstream? Or again we need to do this ourselves,
just like the smi7xx frame buffer driver?

> 2. Community help? my experience is it is hard to tell volunteer what to
> do, they always want to go their own way (which I understand - and which
> helps to balance visions)

It is true. But the premise is volunteer should already know what needs to be
done. I am afraid for Loongson this is not the case. I just saw a post on
lemote's bbs related to this: http://www.lemote.com/bbs/viewthread.php?tid=24073
"每年公布软件发展路线图,有了这个图,广大工程师就是业余抽空也能参与."

> 3. SMI drivers: not ICT/Lemote's fault. We've all (Lemote, Mandriva,
> Dexxon) been asking for a solution for over a year, meet so many times,
> emails, phone calls. And we're still doing it.

What not use xorg's siliconmotion driver, given that it is already working
in most situations?

> 4. A lot of xorg things don't work as well on x86. Do I blame Intel (I
> have an ATI card...)?

I am not blaming anyone for Xorg not working here. I just wanted to sum the
current situation, pointed out what needs to be done, complement my article.

> 5. Lemote (and Dexxon) ships a platform with an OS to do a specific job.
> Arguably I'd say it works ok to do that job. Now you want to change
> things, and Linux, the documentation provided and etc allows you to do
> so and it's great! Your changes might not work but you shouldn't really
> blame us for that (especially if it's a video driver issue....)

Again, I am not blaming anyone for my changes not working.
I myself asked for all the troubles I have dealt and am dealting with.
I just want others to know what I have done and what I have get in return.

> I think you're confusing responsibilities and feel frustrated because
> things move slowly. Well we're even more frustrated but we've learned to
> be patient and definitely appreciate the work and the participation of
> the community following us.

Please, don't misunderstood me.
I will say it for one last time: I am not blaming anyone.
But I do feel frustrated because in some areas, things do not move at all or at
least not start moving and don't know yet whether it will ever start move or not.

> Wishing you luck finding a job.

Thanks.

[Zhang Le]

On 16:28 Thu 09 Jul , Frederic Muller wrote:
> 4. A lot of xorg things don't work as well on x86. Do I blame Intel (I
> have an ATI card...)?

About the Xorg's patch issue. What you said is true. And there is no point to
blame Intel. But the situation on Loongson is different from that on x86.

Although there are things that don't work on x86. But there is no basic support
problems.

But without this patch, X won't start on Loongson.
http://www.gentoo-cn.org/gitweb/?p=loongson.git;a=blob;f=x11-base/xorg-server/files/xorg-server-1.6.0-loongson.patch;h=4351f28fd22df46132a0fa5f3fcda1efc89201c6;hb=HEAD

Note there are two parts in the above patch, only the second part is necessary
for X to work. Also note the line 39:

39 + 0x1fd00000);

This address is Loongson specific. Since it is hardcoded, it won't be accepted
by upstream because other MIPS machine may use a different address.

In foreseeable future, upstream Xorg won't get support for Loongson, because
there is any plan or whatsoever ever been made.

[Zhang Le]

On 18:07 Thu 09 Jul , Zhang Le wrote:
> On 16:28 Thu 09 Jul , Frederic Muller wrote:
> > 5. Lemote (and Dexxon) ships a platform with an OS to do a specific job.
> > Arguably I'd say it works ok to do that job. Now you want to change
> > things, and Linux, the documentation provided and etc allows you to do
> > so and it's great! Your changes might not work but you shouldn't really
> > blame us for that (especially if it's a video driver issue....)
>
> Again, I am not blaming anyone for my changes not working.
> I myself asked for all the troubles I have dealt and am dealting with.
> I just want others to know what I have done and what I have get in return.

Here by "trouble", I only mean technical problems I have encountered so far on
Loongson.

[Zhang Le]

On 21:51 Thu 09 Jul , Zhang Le wrote:
> On 16:28 Thu 09 Jul , Frederic Muller wrote:
> > 4. A lot of xorg things don't work as well on x86. Do I blame Intel (I
> > have an ATI card...)?
>
> About the Xorg's patch issue. What you said is true. And there is no point to
> blame Intel. But the situation on Loongson is different from that on x86.
>
> Although there are things that don't work on x86. But there is no basic support
> problems.
>
> But without this patch, X won't start on Loongson.
> http://www.gentoo-cn.org/gitweb/?p=loongson.git;a=blob;f=x11-base/xorg-server/files/xorg-server-1.6.0-loongson.patch;h=4351f28fd22df46132a0fa5f3fcda1efc89201c6;hb=HEAD
>
> Note there are two parts in the above patch, only the second part is necessary
> for X to work. Also note the line 39:
>
> 39 + 0x1fd00000);
>
> This address is Loongson specific. Since it is hardcoded, it won't be accepted
> by upstream because other MIPS machine may use a different address.
>
> In foreseeable future, upstream Xorg won't get support for Loongson, because
> there is any plan or whatsoever ever been made.

Sorry, here "any plan" should be "no plan".

[yanhua]

Zhang Le 写道:

> On 16:28 Thu 09 Jul , Frederic Muller wrote:
>
>> 4. A lot of xorg things don't work as well on x86. Do I blame Intel (I
>> have an ATI card...)?
>>
>
> About the Xorg's patch issue. What you said is true. And there is no point to
> blame Intel. But the situation on Loongson is different from that on x86.
>
> Although there are things that don't work on x86. But there is no basic support
> problems.
>
> But without this patch, X won't start on Loongson.
> http://www.gentoo-cn.org/gitweb/?p=loongson.git;a=blob;f=x11-base/xorg-server/files/xorg-server-1.6.0-loongson.patch;h=4351f28fd22df46132a0fa5f3fcda1efc89201c6;hb=HEAD
>
> Note there are two parts in the above patch, only the second part is necessary
> for X to work. Also note the line 39:
>
> 39 + 0x1fd00000);
>
> This address is Loongson specific. Since it is hardcoded, it won't be accepted
> by upstream because other MIPS machine may use a different address.
>
> In foreseeable future, upstream Xorg won't get support for Loongson, because
> there is any plan or whatsoever ever been made.
>

This patch is no needed, now loongson can use standard x server.

--
晏华

[Zhang Le]

Thanks for sharing the insights, but would you please talk about some details,
like:

1. the version of x server
2. version of kernel and accompanying kernel patch

[Zhang Le]

On 01:50 Sat 04 Jul , Zhang Le wrote:
> I want to explain a little bit about what I wrote about how "A healthy free
> software community (backed by a commercial company) works". I didn't really mean
> any specific company here. You can replace "a company" with "a group of
> companies".

I think I know where this misunderstanding came from.
And I apologize for this.

I should make it clear in the first place that I was not blaming anyone.
I just depicted a picture which may be realized some day in the future.

Maybe we can establish a foundation, just like Linux Foundation:
http://www.linuxfoundation.org/about

[Eric Fisher]

2009/7/10 Zhang Le <r0b...@gentoo.org>:

> Maybe we can establish a foundation, just like Linux Foundation:
> http://www.linuxfoundation.org/about
>

I like foundation.

[Wu Zhangjin]

Hi,

> Maybe we can establish a foundation, just like Linux Foundation:
> http://www.linuxfoundation.org/about

I think this is the most important point.

currently, the relation or contact among the existing groups(including
companies, individuals and communities and so on) around loongson is
incompact or simplex. there is a need to establish an/a
organization/platform to integrate all of these groups(not only the
high-level, but also the members) together, and then, all of the members
from different groups can work together, co-operate with each other to
cope with the problems existing or coming.

such an/a organization/platform can be a foundation or something
similar, and it should:

1. promote the communication among the members

provide the basic services, such as development platform/mailing
list/bbs/irc/wiki/blog, so the developers from different groups can
share information, collaborate with each other.

hold annual workshop, so people can present papers, demonstrate
loongson-based DIYs, share the experience of development on loongson and
son on.

2. absorb new members

for example, hold contest/training in high-schools and companies.

Regards,
Wu Zhangjin

[Zhang Le]

On 09:43 Fri 10 Jul , yanhua wrote:
>

> This patch is no needed, now loongson can use standard x server.

Hi, yanhua,

If you are sure that this patch is not needed on yeeloong, please tell me the
version of xorg-server and provide me the source of the siliconmotion driver.
I'd like to try it.

I just tried xorg-server-1.6.2 and xf86-video-siliconmotion-1.7.2 (from x.org).
It doesn't work.

However, xorg-server-1.6.2 and xf86-video-sis-0.10.1 works on 2F fuloong box.

And reason I believe should be xf86-video-siliconmotion uses libvgahw.so but
xf86-video-sis doesn't.

$ readelf -s /usr/lib/xorg/modules/drivers/siliconmotion_drv.so | grep -i vga
26: 0001bfc0 0 FUNC GLOBAL DEFAULT UND vgaHWSave
34: 0001bf80 0 FUNC GLOBAL DEFAULT UND vgaHWGetIndex
49: 0001bed0 0 FUNC GLOBAL DEFAULT UND vgaHWSetStdFuncs
51: 0001bec0 0 FUNC GLOBAL DEFAULT UND vgaHWddc1SetSpeedWeak
77: 0001bdc0 0 FUNC GLOBAL DEFAULT UND vgaHWRestore
95: 0001bd20 0 FUNC GLOBAL DEFAULT UND vgaHWMapMem
132: 0001bb90 0 FUNC GLOBAL DEFAULT UND vgaHWUnmapMem
144: 0001baf0 0 FUNC GLOBAL DEFAULT UND vgaHWCopyReg
158: 0001ba70 0 FUNC GLOBAL DEFAULT UND vgaHWSetMmioFuncs
163: 0001ba20 0 FUNC GLOBAL DEFAULT UND vgaHWGetHWRec
193: 0001b900 0 FUNC GLOBAL DEFAULT UND vgaHWGetIOBase
198: 0001b8d0 0 FUNC GLOBAL DEFAULT UND vgaHWInit
201: 00000000 0 NOTYPE GLOBAL DEFAULT UND vgaHWProtect
220: 00000000 0 NOTYPE GLOBAL DEFAULT UND resVgaShared
166: 00014378 1208 FUNC LOCAL DEFAULT 10 SMILynx_CrtcModeSet_vga
225: 00000000 0 NOTYPE GLOBAL DEFAULT UND vgaHWProtect
228: 0001b8d0 0 FUNC GLOBAL DEFAULT UND vgaHWInit
233: 0001b900 0 FUNC GLOBAL DEFAULT UND vgaHWGetIOBase
268: 0001ba20 0 FUNC GLOBAL DEFAULT UND vgaHWGetHWRec
273: 0001ba70 0 FUNC GLOBAL DEFAULT UND vgaHWSetMmioFuncs
288: 0001baf0 0 FUNC GLOBAL DEFAULT UND vgaHWCopyReg
301: 0001bb90 0 FUNC GLOBAL DEFAULT UND vgaHWUnmapMem
322: 00000000 0 NOTYPE GLOBAL DEFAULT UND resVgaShared
340: 0001bd20 0 FUNC GLOBAL DEFAULT UND vgaHWMapMem
359: 0001bdc0 0 FUNC GLOBAL DEFAULT UND vgaHWRestore
388: 0001bec0 0 FUNC GLOBAL DEFAULT UND vgaHWddc1SetSpeedWeak
390: 0001bed0 0 FUNC GLOBAL DEFAULT UND vgaHWSetStdFuncs
406: 0001bf80 0 FUNC GLOBAL DEFAULT UND vgaHWGetIndex
414: 0001bfc0 0 FUNC GLOBAL DEFAULT UND vgaHWSave

$ readelf -s /usr/lib/xorg/modules/drivers/sis_drv.so | grep -i vga
28: 00000000 0 NOTYPE GLOBAL DEFAULT UND resVgaMemShared
154: 00000000 0 NOTYPE GLOBAL DEFAULT UND resVgaIoShared
188: 00000000 0 NOTYPE GLOBAL DEFAULT UND resVgaShared

And I just checked the xf86-video-siliconmotion shipped with Loonux also uses
libvgahw.so.

Now I will try to patch xorg-server-1.6.2 and see if it works on yeeloong.

[Zhang Le]

On 02:00 Sat 11 Jul , Zhang Le wrote:
> Now I will try to patch xorg-server-1.6.2 and see if it works on yeeloong.

Patched xorg-server-1.6.2 works.

Now I will try vanilla xorg-server-1.6.2 again, if it still does not work, then
I am afraid there is only one explanation:

The patch is _still_ required.

Right?

[Zhang Le]

On 19:55 Sat 11 Jul , Zhang Le wrote:
> On 02:00 Sat 11 Jul , Zhang Le wrote:
> > Now I will try to patch xorg-server-1.6.2 and see if it works on yeeloong.
>
> Patched xorg-server-1.6.2 works.
>
> Now I will try vanilla xorg-server-1.6.2 again, if it still does not work,

And it _does not_ work, :(

> then I am afraid there is only one explanation:
> The patch is _still_ required.

--