GSOC Project: Merge code from two Realtek Wifi Drivers

[Sugam Srivastava]

To whomever it may concern,
I am very interested in the GSOC project "Merge code from two Realtek
Wifi Drivers". I have been reading all the available resources and
trying to get a deeper insight from the code. I wanted to ask whether
there would be some type of pre-proposal screening or we have to
directly start working on the proposal.

Regards,
Sugam Srivastava

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Martin Husemann]

Post by Sugam Srivastava
To whomever it may concern,
I am very interested in the GSOC project "Merge code from two Realtek
Wifi Drivers". I have been reading all the available resources and
trying to get a deeper insight from the code. I wanted to ask whether
there would be some type of pre-proposal screening or we have to
directly start working on the proposal.

Both drivers have changed signifcantly on the "wifi" branch/topic, which is
probably not going to be merged before GSoC.

It is important to do the work for this project based on that branch, we
have to think about proper infrastructure for that (it might require you
to work with mercurial, instead of having mostly free choice as for our
other GSoC projects).

The wiki has a few tutorials how to deal with developement on the wifi
topic.

Martin

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Dave Huang]

Post by Mouse
I don't much care what the content is produced by, as long as it
conforms to an open spec.
I don't know whether you're an outlier. But I too am not interested in
being part of Google's product and would require substantial incentive,
shall we say, in order to be interested in trying to winkle content out
of a .docx or anything similarly obfuscated.

But .docx is Office Open XML, an ISO standard. Probably the most popular
open source package that reads/writes it is LibreOffice, available in
pkgsrc. So there shouldn't be any problem with .docx.

That said, both the .docx and the Google Doc are just text with some
headings/bold, and would look fine as a plain text document.

[Sugam Srivastava]

To whomever it may concern,
I have completed all the tutorials and have started working on my
proposal for the project. I have a couple of questions regarding the
project:
1. In which file would be put the merged code for the two drivers
2. I also have trouble deciding the deliverables for the evaluation. I
am thinking of dividing code into 2 parts and making each of them the
deliverables for the evaluations.
Regards,
Sugam.

Post by Martin Husemann

Post by Sugam Srivastava
To whomever it may concern,
I am very interested in the GSOC project "Merge code from two Realtek
Wifi Drivers". I have been reading all the available resources and
trying to get a deeper insight from the code. I wanted to ask whether
there would be some type of pre-proposal screening or we have to
directly start working on the proposal.

Both drivers have changed signifcantly on the "wifi" branch/topic, which is
probably not going to be merged before GSoC.
It is important to do the work for this project based on that branch, we
have to think about proper infrastructure for that (it might require you
to work with mercurial, instead of having mostly free choice as for our
other GSoC projects).
The wiki has a few tutorials how to deal with developement on the wifi
topic.
Martin

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Christos Zoulas]

Post by Sugam Srivastava
To whomever it may concern,
I have completed all the tutorials and have started working on my
proposal for the project. I have a couple of questions regarding the
1. In which file would be put the merged code for the two drivers

We can decide that later. I guess we should choose the name that represents
the interface family better.

Post by Sugam Srivastava
2. I also have trouble deciding the deliverables for the evaluation. I
am thinking of dividing code into 2 parts and making each of them the
deliverables for the evaluations.

Well, you can do this incrementally by making the drivers share a common
set of functions and increase the amount of sharing until the drivers are
merged.

christos


--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Sugam Srivastava]

Hi,
I have completed my draft proposal to my best, could you please have a
look at it and provide feedback.
Link: https://docs.google.com/document/d/1K1BFeZUstc4vjvpbGWo7UUm6SkSik6QxMdm4bxHMmtQ/edit?usp=sharing
Regards,
Sugam

Post by Martin Husemann

Post by Sugam Srivastava
To whomever it may concern,
I have completed all the tutorials and have started working on my
proposal for the project. I have a couple of questions regarding the
1. In which file would be put the merged code for the two drivers

- sys/dev/ic/if_rt.... (where ... could be some chipset number typical
for this realtek family). There would be .c and .h files.
- sys/dev/pci/if_rtwn.c the pci specific parts of the driver, using the
code from dev/ic as a "library"
- sys/dev/usb/if_urtwn.c the usb specific parts

Post by Sugam Srivastava
2. I also have trouble deciding the deliverables for the evaluation. I
am thinking of dividing code into 2 parts and making each of them the
deliverables for the evaluations.

In this case I don't think this makes much sense, as you have to understand
the common parts first, so you can not finish one driver and then do the
other.
Martin

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Greg Troxel]

Post by Sugam Srivastava
Link: https://docs.google.com/document/d/1K1BFeZUstc4vjvpbGWo7UUm6SkSik6QxMdm4bxHMmtQ/edit?usp=sharing

I'm not involved in GSOC mentoring (long discussion omitted), but in
NetBSD culture you should be aware that several things are not greeted
by wild enthusiasm:

proprietary formats like .docx (maybe I am blurring you with someone
else)

links to centralized services run by advertising companies

Were it me -- which it is definitely not!! -- I would want to see plain
text included in the email, or maybe pdf if diagrams were really
necessary, and the content produced by Free Software.

It could be that others are happy about this and I am the outlier, so
take this with a grain of salt, if at all.

[Mouse]

Post by Mouse
I don't much care what the content is produced by, as long as it
conforms to an open spec.
I don't know whether you're an outlier. But I too am not interested
in being part of Google's product and would require substantial
incentive, shall we say, in order to be interested in trying to
winkle content out of a .docx or anything similarly obfuscated.

But docx is Office Open XML, an ISO standard.

But is it an open standard, or is it pay-to-play? In my (admittedly
very limited) experience, ISO specs tend to be the latter. (Also, even
if it is an open spec, it still calls for nontrivial effort for me to
convert it to anything useful, bringing me back to what I said about
how the more work it takes me the less I like it.)

/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML ***@rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Martin Husemann]

Post by Greg Troxel

Post by Sugam Srivastava
Link: https://docs.google.com/document/d/1K1BFeZUstc4vjvpbGWo7UUm6SkSik6QxMdm4bxHMmtQ/edit?usp=sharing

I'm not involved in GSOC mentoring (long discussion omitted), but in
NetBSD culture you should be aware that several things are not greeted
proprietary formats like .docx (maybe I am blurring you with someone
else)
links to centralized services run by advertising companies

Hey Greg,

I understand well where you are coming from (and I usually do not open
any links mailed on our [or other] mailing lists), but I'd like to
clarify some things in defense of the students here:

- It is Google Summer of Code - and of course they encourage students
to use the Google tools for early sharing of drafts.
- It only needed plain Firefox on my NetBSD machine to view the draft
on google docs - no downloading of proprietary or whatever
files

If the draft at this stage needed any of the formatting sugar is another
thing, but students often try to show they invested effort in the draft
by beautifying it like this.

Sugam, I'll answer separately for the content of the draft.

Martin

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Martin Husemann]

Post by Sugam Srivastava
To whomever it may concern,
I have completed all the tutorials and have started working on my
proposal for the project. I have a couple of questions regarding the
1. In which file would be put the merged code for the two drivers

We typically split it up like:

- sys/dev/ic/if_rt.... (where ... could be some chipset number typical
for this realtek family). There would be .c and .h files.
- sys/dev/pci/if_rtwn.c the pci specific parts of the driver, using the
code from dev/ic as a "library"
- sys/dev/usb/if_urtwn.c the usb specific parts

Post by Sugam Srivastava
2. I also have trouble deciding the deliverables for the evaluation. I
am thinking of dividing code into 2 parts and making each of them the
deliverables for the evaluations.

In this case I don't think this makes much sense, as you have to understand
the common parts first, so you can not finish one driver and then do the
other.


Martin

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Robert Elz]

Date: Sun, 11 Apr 2021 17:51:14 -0400
From: Greg Troxel <***@lexort.com>
Message-ID: <***@s1.lexort.com>

| links to centralized services run by advertising companies

| It could be that others are happy about this and I am the outlier, so
| take this with a grain of salt, if at all.

Similar - my general position is that if someone wants me to read something
they must send it to me in a format I can handle.

It isn't worth my time trying to go fetch someone else's files from who
knows where (advertising or not) or try to deal with bizarre formats.

kre

ps: I am also not a GSOC mentor, and wouldn't be mentoring that project
even if I were, so this is, similarly, just a general comment.



--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Mouse]

Post by Greg Troxel

Post by Sugam Srivastava
Link: https://docs.google.com/document/d/1K1BFeZUstc4vjvpbGWo7UUm6SkSik6QxMdm4bxHMmtQ/edit?usp=sharing

I'm not involved in GSOC mentoring (long discussion omitted), but in
NetBSD culture you should be aware that several things are not
proprietary formats like .docx (maybe I am blurring you with
someone else)
links to centralized services run by advertising companies
Were it me -- which it is definitely not!! -- I would want to see
plain text included in the email, or maybe pdf if diagrams were
really necessary, and the content produced by Free Software.

I don't much care what the content is produced by, as long as it
conforms to an open spec. (Indeed, I just recently ran into an
interoperability problem due to ignoring REQUIREDs in the spec...by a
piece of open-source software.)

Actually, that's not quite true. The more work I have to do to get
something usable (which usually means plain text) out of it, the less I
like it. But if it doesn't conform to an open spec, there's pretty
much no hope of my getting anything useful out of it.

As for "Free Software"...that, I usually mildly dislike, because most
of the uses of the term in my experience are referring to GPLed
software, which in my opinion is nothing short of hypocritical.

Post by Greg Troxel
It could be that others are happy about this and I am the outlier, so
take this with a grain of salt, if at all.

I don't know whether you're an outlier. But I too am not interested in
being part of Google's product and would require substantial incentive,
shall we say, in order to be interested in trying to winkle content out
of a .docx or anything similarly obfuscated. (I didn't see the mail
you quoted the docs.google.com link above from. But I did see a mail
on tech-kern saying it was some proposal with links to docs.google.com
and www.icloud.com, with an attachment with a .docx name; my reaction
was an eyeroll and a delete.)

/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML ***@rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de