2017-03-09 20:36:54

by Julia Lawall

[permalink] [raw]
Subject: outreachy

Hello,

I discussed the issue of outreachy patches for bcm with Greg, and we are
not convinced that not having the patches CCd to you is such a good idea.
While we don't want to spam you with noise, some of the applicants are
starting to make more significant changes that it could be useful for you
to be aware of.

Could we try a compromise where you are not CCd on whitespace patches, but
you are CCd on patches that actually modify the code?

thanks,
julia


2017-03-09 20:52:34

by Scott Branden

[permalink] [raw]
Subject: Re: outreachy

Hi Julia,

On 17-03-09 12:36 PM, Julia Lawall wrote:
> Hello,
>
> I discussed the issue of outreachy patches for bcm with Greg, and we are
> not convinced that not having the patches CCd to you is such a good idea.
> While we don't want to spam you with noise, some of the applicants are
> starting to make more significant changes that it could be useful for you
> to be aware of.
>
> Could we try a compromise where you are not CCd on whitespace patches, but
> you are CCd on patches that actually modify the code?
All I'm asking is you work through your outreachy patches internal first
to get rid of the most basic mistakes and email traffic it is geerating.
Once that learning process is through then they can be sent out like
any other patches to the kernel mailing lists and maintainers.
>
> thanks,
> julia
>

Thanks,
Scott

2017-03-09 20:56:59

by Stephen Warren

[permalink] [raw]
Subject: Re: outreachy

On 03/09/2017 01:51 PM, Scott Branden wrote:
> Hi Julia,
>
> On 17-03-09 12:36 PM, Julia Lawall wrote:
>> Hello,
>>
>> I discussed the issue of outreachy patches for bcm with Greg, and we are
>> not convinced that not having the patches CCd to you is such a good idea.
>> While we don't want to spam you with noise, some of the applicants are
>> starting to make more significant changes that it could be useful for you
>> to be aware of.
>>
>> Could we try a compromise where you are not CCd on whitespace patches,
>> but
>> you are CCd on patches that actually modify the code?
>
> All I'm asking is you work through your outreachy patches internal first
> to get rid of the most basic mistakes and email traffic it is geerating.
> Once that learning process is through then they can be sent out like
> any other patches to the kernel mailing lists and maintainers.

+1 from me too; I find these patches rather high volume and had to add a
filter to keep them out of my primary inbox. I don't know what process
is in place, but I would suggest:

1) Senders send everything to the outreachy list, where they are
reviewed for basic issues, like learning to use git send-email, learning
checkpatch, etc. In this case, only send the patch to the outreachy
mailing list and nowhere else.

2) Once a patch has passed review there, then send the patch to the
regular kernel mailing list just like any other patch; follow the output
of get_maintainers.pl.

We have something like (1) inside NVIDIA for new contributors and
pre-upstreaming IP review. It helps all the newcomers, but without
requiring anyone involved in (2) to change behaviour.

The process I suggest is very much inline with the typically suggested
"asking questions" process: (1) read docs yourself (2) ask local
contacts for help, (3) start asking wider audiences for help.

2017-03-09 21:30:57

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: outreachy

On Thu, Mar 09, 2017 at 01:56:49PM -0700, Stephen Warren wrote:
> On 03/09/2017 01:51 PM, Scott Branden wrote:
> > Hi Julia,
> >
> > On 17-03-09 12:36 PM, Julia Lawall wrote:
> > > Hello,
> > >
> > > I discussed the issue of outreachy patches for bcm with Greg, and we are
> > > not convinced that not having the patches CCd to you is such a good idea.
> > > While we don't want to spam you with noise, some of the applicants are
> > > starting to make more significant changes that it could be useful for you
> > > to be aware of.
> > >
> > > Could we try a compromise where you are not CCd on whitespace patches,
> > > but
> > > you are CCd on patches that actually modify the code?
> >
> > All I'm asking is you work through your outreachy patches internal first
> > to get rid of the most basic mistakes and email traffic it is geerating.
> > Once that learning process is through then they can be sent out like
> > any other patches to the kernel mailing lists and maintainers.
>
> +1 from me too; I find these patches rather high volume and had to add a
> filter to keep them out of my primary inbox.

Hah! That's the joy of being a maintainer of a driver in staging. Even
if you filter out outreachy, you are going to get a lot of "basic
mistakes" and other type patches cc:ed to you.

I strongly suggest, that if you all don't like this type of stuff,
either:
- work to get the code out of staging as soon as possible (i.e.
send me coding style fixes for everything right now, and then
fix up the rest of the stuff.)
- take yourself off the maintainer list for this code.

It's your choice, outreachy right now is a lot of patches, but again,
it's not going to keep you from getting the "basic" stuff sent to you
in ways that is totally wrong.

thanks,

greg k-h

2017-03-09 22:15:46

by Florian Fainelli

[permalink] [raw]
Subject: Re: outreachy

On 03/09/2017 01:20 PM, Greg KH wrote:
> On Thu, Mar 09, 2017 at 01:56:49PM -0700, Stephen Warren wrote:
>> On 03/09/2017 01:51 PM, Scott Branden wrote:
>>> Hi Julia,
>>>
>>> On 17-03-09 12:36 PM, Julia Lawall wrote:
>>>> Hello,
>>>>
>>>> I discussed the issue of outreachy patches for bcm with Greg, and we are
>>>> not convinced that not having the patches CCd to you is such a good idea.
>>>> While we don't want to spam you with noise, some of the applicants are
>>>> starting to make more significant changes that it could be useful for you
>>>> to be aware of.
>>>>
>>>> Could we try a compromise where you are not CCd on whitespace patches,
>>>> but
>>>> you are CCd on patches that actually modify the code?
>>>
>>> All I'm asking is you work through your outreachy patches internal first
>>> to get rid of the most basic mistakes and email traffic it is geerating.
>>> Once that learning process is through then they can be sent out like
>>> any other patches to the kernel mailing lists and maintainers.
>>
>> +1 from me too; I find these patches rather high volume and had to add a
>> filter to keep them out of my primary inbox.
>
> Hah! That's the joy of being a maintainer of a driver in staging. Even
> if you filter out outreachy, you are going to get a lot of "basic
> mistakes" and other type patches cc:ed to you.
>
> I strongly suggest, that if you all don't like this type of stuff,
> either:
> - work to get the code out of staging as soon as possible (i.e.
> send me coding style fixes for everything right now, and then
> fix up the rest of the stuff.)
> - take yourself off the maintainer list for this code.

Keep in mind that most people on this CC list are getting these patches
because of the bcm283* regular expression, and maybe that's what needs
fixing here in the first place.

Incidentally, Stephen did send a patch to get him removed from the
MAINTAINERS entry for Raspberry Pi stuff, so I guess, problem solved for
him. We still have a ton of people from Broadcom who are going to
receive these emails, with mild interest in staging patches.

>
> It's your choice, outreachy right now is a lot of patches, but again,
> it's not going to keep you from getting the "basic" stuff sent to you
> in ways that is totally wrong.

That is absolutely true, but the thing is that we really got a big spike
of patch submissions lately, and that was totally not accepted. I am not
asking for a "heads-up" email telling people that they are going to
receive more traffic than usual (because that would be too much over
head), but if there was an internal review first on the outreachy
mailing-list and second a proper submission which is going to pass your
acceptance criteria, we would be de facto reducing the amount of emails
that we received.

The outreachy list obviously has people like you and Julia who are
willing to help and provide feedback, so I really don't see what's the
problem in setting up a two tier review here, it does not change
anything for you, but it does change a lot for us.
--
Florian

2017-03-10 06:01:19

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: outreachy

On Thu, Mar 09, 2017 at 02:15:21PM -0800, Florian Fainelli wrote:
> On 03/09/2017 01:20 PM, Greg KH wrote:
> > On Thu, Mar 09, 2017 at 01:56:49PM -0700, Stephen Warren wrote:
> >> On 03/09/2017 01:51 PM, Scott Branden wrote:
> >>> Hi Julia,
> >>>
> >>> On 17-03-09 12:36 PM, Julia Lawall wrote:
> >>>> Hello,
> >>>>
> >>>> I discussed the issue of outreachy patches for bcm with Greg, and we are
> >>>> not convinced that not having the patches CCd to you is such a good idea.
> >>>> While we don't want to spam you with noise, some of the applicants are
> >>>> starting to make more significant changes that it could be useful for you
> >>>> to be aware of.
> >>>>
> >>>> Could we try a compromise where you are not CCd on whitespace patches,
> >>>> but
> >>>> you are CCd on patches that actually modify the code?
> >>>
> >>> All I'm asking is you work through your outreachy patches internal first
> >>> to get rid of the most basic mistakes and email traffic it is geerating.
> >>> Once that learning process is through then they can be sent out like
> >>> any other patches to the kernel mailing lists and maintainers.
> >>
> >> +1 from me too; I find these patches rather high volume and had to add a
> >> filter to keep them out of my primary inbox.
> >
> > Hah! That's the joy of being a maintainer of a driver in staging. Even
> > if you filter out outreachy, you are going to get a lot of "basic
> > mistakes" and other type patches cc:ed to you.
> >
> > I strongly suggest, that if you all don't like this type of stuff,
> > either:
> > - work to get the code out of staging as soon as possible (i.e.
> > send me coding style fixes for everything right now, and then
> > fix up the rest of the stuff.)
> > - take yourself off the maintainer list for this code.
>
> Keep in mind that most people on this CC list are getting these patches
> because of the bcm283* regular expression, and maybe that's what needs
> fixing here in the first place.

Yes, I suggest the someone fixes that if they do not wish to get these
types of emails. Having a regex like that in MAINTAINERS is very crazy,
from another thread I don't think it's really doing what you all want it
to do (meaning it's hitting a lot more files than expected.)

> > It's your choice, outreachy right now is a lot of patches, but again,
> > it's not going to keep you from getting the "basic" stuff sent to you
> > in ways that is totally wrong.
>
> That is absolutely true, but the thing is that we really got a big spike
> of patch submissions lately, and that was totally not accepted. I am not
> asking for a "heads-up" email telling people that they are going to
> receive more traffic than usual (because that would be too much over
> head), but if there was an internal review first on the outreachy
> mailing-list and second a proper submission which is going to pass your
> acceptance criteria, we would be de facto reducing the amount of emails
> that we received.
>
> The outreachy list obviously has people like you and Julia who are
> willing to help and provide feedback, so I really don't see what's the
> problem in setting up a two tier review here, it does not change
> anything for you, but it does change a lot for us.

Again, even if outreachy isn't happening, you are still going to be
getting these types of patches from all of the "normal" people that send
staging cleanup patches. So it's not going to buy you all that much of
a reprieve.

So please send in a MAINTAINERS patch if you don't wish to get these
kinds of patches. Or, again, just spend a day and send me cleanup
patches to keep anyone else from needing to send in basic checkpatch
fixes for this code.

thanks,

greg k-h

2017-03-17 15:25:29

by Pavel Machek

[permalink] [raw]
Subject: Re: outreachy

Hi!

>
> Hah! That's the joy of being a maintainer of a driver in staging. Even
> if you filter out outreachy, you are going to get a lot of "basic
> mistakes" and other type patches cc:ed to you.
>
> I strongly suggest, that if you all don't like this type of stuff,
> either:
> - work to get the code out of staging as soon as possible (i.e.
> send me coding style fixes for everything right now, and then
> fix up the rest of the stuff.)
> - take yourself off the maintainer list for this code.
>
> It's your choice, outreachy right now is a lot of patches, but again,
> it's not going to keep you from getting the "basic" stuff sent to you
> in ways that is totally wrong.

Could we get these trivial patches off the lkml? Yes, lkml already has
a lot of traffic, but no, this is not useful :-(.

Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html


Attachments:
(No filename) (973.00 B)
signature.asc (181.00 B)
Digital signature
Download all attachments

2017-03-17 16:56:58

by Julia Lawall

[permalink] [raw]
Subject: Re: outreachy



On Fri, 17 Mar 2017, Pavel Machek wrote:

> Hi!
>
> >
> > Hah! That's the joy of being a maintainer of a driver in staging. Even
> > if you filter out outreachy, you are going to get a lot of "basic
> > mistakes" and other type patches cc:ed to you.
> >
> > I strongly suggest, that if you all don't like this type of stuff,
> > either:
> > - work to get the code out of staging as soon as possible (i.e.
> > send me coding style fixes for everything right now, and then
> > fix up the rest of the stuff.)
> > - take yourself off the maintainer list for this code.
> >
> > It's your choice, outreachy right now is a lot of patches, but again,
> > it's not going to keep you from getting the "basic" stuff sent to you
> > in ways that is totally wrong.
>
> Could we get these trivial patches off the lkml? Yes, lkml already has
> a lot of traffic, but no, this is not useful :-(.

The outreachy instructions say to use the -nol argument to
get_maintainers, which would prevent them from being sent to any mailing
list. However others thought that all patches should be sent to mailing
lists, and so I haven't enforced anything for people who have omitted
-nol. However I have tried to remove bcm maintainers from CC lists on
replies and reminded people not to send you patches,

julia


>
> Pavel
> --
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
>

2017-03-19 07:38:02

by Michael Zoran

[permalink] [raw]
Subject: Re: outreachy/moving a driver out of staging

On Thu, 2017-03-09 at 22:20 +0100, Greg KH wrote:
> On Thu, Mar 09, 2017 at 01:56:49PM -0700, Stephen Warren wrote:
> > On 03/09/2017 01:51 PM, Scott Branden wrote:
> > > Hi Julia,
> > >
> > > On 17-03-09 12:36 PM, Julia Lawall wrote:
> > > > Hello,
> > > >
> > > > I discussed the issue of outreachy patches for bcm with Greg,
> > > > and we are
> > > > not convinced that not having the patches CCd to you is such a
> > > > good idea.
> > > > While we don't want to spam you with noise, some of the
> > > > applicants are
> > > > starting to make more significant changes that it could be
> > > > useful for you
> > > > to be aware of.
> > > >
> > > > Could we try a compromise where you are not CCd on whitespace
> > > > patches,
> > > > but
> > > > you are CCd on patches that actually modify the code?
> > >
> > > All I'm asking is you work through your outreachy patches
> > > internal first
> > > to get rid of the most basic mistakes and email traffic it is
> > > geerating.
> > >  Once that learning process is through then they can be sent out
> > > like
> > > any other patches to the kernel mailing lists and maintainers.
> >
> > +1 from me too; I find these patches rather high volume and had to
> > add a
> > filter to keep them out of my primary inbox.
>
> Hah!  That's the joy of being a maintainer of a driver in
> staging.  Even
> if you filter out outreachy, you are going to get a lot of "basic
> mistakes" and other type patches cc:ed to you.
>
> I strongly suggest, that if you all don't like this type of stuff,
> either:
> - work to get the code out of staging as soon as possible (i.e.
>   send me coding style fixes for everything right now, and then
>   fix up the rest of the stuff.)
> - take yourself off the maintainer list for this code.
>
> It's your choice, outreachy right now is a lot of patches, but again,
> it's not going to keep you from getting the "basic" stuff sent to you
> in ways that is totally wrong.
>
> thanks,
>
> greg k-h
>

Hi Greg,

I just noticed this e-mail. What exactly is the requirement to get a
driver or subsystem out of staging?

I can image a day when vc04_services or VideoCore gets moved out of
staging at some point. What exactly would it take to make something
like that happen?


2017-03-19 09:16:03

by Russell King (Oracle)

[permalink] [raw]
Subject: Re: outreachy/moving a driver out of staging

On Sun, Mar 19, 2017 at 12:37:47AM -0700, Michael Zoran wrote:
> I just noticed this e-mail. What exactly is the requirement to get a
> driver or subsystem out of staging?

This is why each driver in staging is supposed to have a TODO file
listing each point that needs to be addressed, and when each point
is addressed, the TODO file should be updated. The TODO file tells
people what the remaining faults are with the driver.

I see that there's a todo file here:

drivers/staging/vc04_services/interface/vchi/TODO

and the very first thing I looked at was:

- Figure out an alternative to the dmac_map_area() hack.

$ grep -r dmac_map_area drivers/staging/vc04_services/interface/vchi
drivers/staging/vc04_services/interface/vchi/TODO: - Figure out an alternative
to the dmac_map_area() hack.

So that one looks like it's resolved, so the TODO file is out of date.
The first step, therefore, is to get the TODO files updated with the
work that has been done, so it's possible to know what work remains.

I notice also that drivers/staging/vc04_services/interface/vchiq_arm
does not have a TODO file, so that needs reviewing and a TODO file
generated.

--
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.

2017-03-19 10:59:20

by Michael Zoran

[permalink] [raw]
Subject: Re: outreachy/moving a driver out of staging

On Sun, 2017-03-19 at 09:15 +0000, Russell King - ARM Linux wrote:
> On Sun, Mar 19, 2017 at 12:37:47AM -0700, Michael Zoran wrote:
> > I just noticed this e-mail.  What exactly is the requirement to get
> > a
> > driver or subsystem out of staging?
>
> This is why each driver in staging is supposed to have a TODO file
> listing each point that needs to be addressed, and when each point
> is addressed, the TODO file should be updated.  The TODO file tells
> people what the remaining faults are with the driver.
>
> I see that there's a todo file here:
>
> drivers/staging/vc04_services/interface/vchi/TODO
>
> and the very first thing I looked at was:
>
>   - Figure out an alternative to the dmac_map_area() hack.
>
> $ grep -r dmac_map_area drivers/staging/vc04_services/interface/vchi
> drivers/staging/vc04_services/interface/vchi/TODO:  - Figure out an
> alternative
> to the dmac_map_area() hack.
>
> So that one looks like it's resolved, so the TODO file is out of
> date.
> The first step, therefore, is to get the TODO files updated with the
> work that has been done, so it's possible to know what work remains.
>
> I notice also that drivers/staging/vc04_services/interface/vchiq_arm
> does not have a TODO file, so that needs reviewing and a TODO file
> generated.
>

I've actually been submitting changes to the TODO files as I've been
going. Just these days I'm not sure which tree things are being
applied to anymore.

So I'm assuming once all the TODO items are addressed, and everything
is checkpatch.pl clean it can be moved out of staging? I would assume
an existing subsystem would need to agree to accept it into their topic
branch correct?

2017-03-20 10:28:07

by Pavel Machek

[permalink] [raw]
Subject: Re: outreachy

Hi!

> > > Hah! That's the joy of being a maintainer of a driver in staging. Even
> > > if you filter out outreachy, you are going to get a lot of "basic
> > > mistakes" and other type patches cc:ed to you.
> > >
> > > I strongly suggest, that if you all don't like this type of stuff,
> > > either:
> > > - work to get the code out of staging as soon as possible (i.e.
> > > send me coding style fixes for everything right now, and then
> > > fix up the rest of the stuff.)
> > > - take yourself off the maintainer list for this code.
> > >
> > > It's your choice, outreachy right now is a lot of patches, but again,
> > > it's not going to keep you from getting the "basic" stuff sent to you
> > > in ways that is totally wrong.
> >
> > Could we get these trivial patches off the lkml? Yes, lkml already has
> > a lot of traffic, but no, this is not useful :-(.
>
> The outreachy instructions say to use the -nol argument to
> get_maintainers, which would prevent them from being sent to any mailing
> list. However others thought that all patches should be sent to mailing
> lists, and so I haven't enforced anything for people who have omitted
> -nol. However I have tried to remove bcm maintainers from CC lists on
> replies and reminded people not to send you patches,

Wonderful :-(.

Can we at least make those people put the word "outreachy" in the
subject so the emails are easier to delete?
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html


Attachments:
(No filename) (1.53 kB)
signature.asc (181.00 B)
Digital signature
Download all attachments

2017-03-20 10:31:55

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: outreachy

On Mon, Mar 20, 2017 at 11:20:32AM +0100, Pavel Machek wrote:
> Hi!
>
> > > > Hah! That's the joy of being a maintainer of a driver in staging. Even
> > > > if you filter out outreachy, you are going to get a lot of "basic
> > > > mistakes" and other type patches cc:ed to you.
> > > >
> > > > I strongly suggest, that if you all don't like this type of stuff,
> > > > either:
> > > > - work to get the code out of staging as soon as possible (i.e.
> > > > send me coding style fixes for everything right now, and then
> > > > fix up the rest of the stuff.)
> > > > - take yourself off the maintainer list for this code.
> > > >
> > > > It's your choice, outreachy right now is a lot of patches, but again,
> > > > it's not going to keep you from getting the "basic" stuff sent to you
> > > > in ways that is totally wrong.
> > >
> > > Could we get these trivial patches off the lkml? Yes, lkml already has
> > > a lot of traffic, but no, this is not useful :-(.
> >
> > The outreachy instructions say to use the -nol argument to
> > get_maintainers, which would prevent them from being sent to any mailing
> > list. However others thought that all patches should be sent to mailing
> > lists, and so I haven't enforced anything for people who have omitted
> > -nol. However I have tried to remove bcm maintainers from CC lists on
> > replies and reminded people not to send you patches,
>
> Wonderful :-(.
>
> Can we at least make those people put the word "outreachy" in the
> subject so the emails are easier to delete?

It's easy for you to filter away as-is if you want to to by just looking
at the cc: list for outreachy right now, don't make a new step for
people to jump through because you don't want to be bothered. You only
have 10 more days of it if you want to just ignore any patch until
then...

greg k-h

2017-03-20 16:38:45

by Pavel Machek

[permalink] [raw]
Subject: Re: outreachy

On Mon 2017-03-20 11:30:08, Greg KH wrote:
> On Mon, Mar 20, 2017 at 11:20:32AM +0100, Pavel Machek wrote:
> > Hi!
> >
> > > > > Hah! That's the joy of being a maintainer of a driver in staging. Even
> > > > > if you filter out outreachy, you are going to get a lot of "basic
> > > > > mistakes" and other type patches cc:ed to you.
> > > > >
> > > > > I strongly suggest, that if you all don't like this type of stuff,
> > > > > either:
> > > > > - work to get the code out of staging as soon as possible (i.e.
> > > > > send me coding style fixes for everything right now, and then
> > > > > fix up the rest of the stuff.)
> > > > > - take yourself off the maintainer list for this code.
> > > > >
> > > > > It's your choice, outreachy right now is a lot of patches, but again,
> > > > > it's not going to keep you from getting the "basic" stuff sent to you
> > > > > in ways that is totally wrong.
> > > >
> > > > Could we get these trivial patches off the lkml? Yes, lkml already has
> > > > a lot of traffic, but no, this is not useful :-(.
> > >
> > > The outreachy instructions say to use the -nol argument to
> > > get_maintainers, which would prevent them from being sent to any mailing
> > > list. However others thought that all patches should be sent to mailing
> > > lists, and so I haven't enforced anything for people who have omitted
> > > -nol. However I have tried to remove bcm maintainers from CC lists on
> > > replies and reminded people not to send you patches,
> >
> > Wonderful :-(.
> >
> > Can we at least make those people put the word "outreachy" in the
> > subject so the emails are easier to delete?
>
> It's easy for you to filter away as-is if you want to to by just looking
> at the cc: list for outreachy right now, don't make a new step for
> people to jump through because you don't want to be bothered. You only
> have 10 more days of it if you want to just ignore any patch until
> then...

10 more days... I can survive 10 more days. I was afraid we would be
stuck with it "forever".
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html


Attachments:
(No filename) (2.14 kB)
signature.asc (181.00 B)
Digital signature
Download all attachments