Dominik Brodowski wrote:
> On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
>>Linux Kernel Mailing List wrote:
>>>commit 42935656914b813c99f91cbac421fe677a6f34ab
>>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
>>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
>>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
>>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
>>Why was this not CC'd to the IDE maintainer, and linux-ide?
> For it is trivial, PCMCIA-related and my time is very limited these days.
That's pathetic. You couldn't even CC linux-kernel on your answer. And
this is not even the first or second time you've been asked to CC a
maintainer.
If you don't have time for a simple CC to the relevant maintainer,
allowing for peer review, then find some other project to contribute to.
We don't need more closed development assholes.
This is open source, keyword OPEN.
Jeff
On Thu, 2006-03-02 at 06:17 -0500, Jeff Garzik wrote:
> Dominik Brodowski wrote:
> > On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
> >>Linux Kernel Mailing List wrote:
> >>>commit 42935656914b813c99f91cbac421fe677a6f34ab
> >>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
> >>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
> >>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
> >>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
>
> >>Why was this not CC'd to the IDE maintainer, and linux-ide?
>
> > For it is trivial, PCMCIA-related and my time is very limited these days.
>
> That's pathetic. You couldn't even CC linux-kernel on your answer. And
> this is not even the first or second time you've been asked to CC a
> maintainer.
I personally don't consider that maintainers have a right to demand
CC's. Sure it's polite and good to CC them, but that's not the same as
having the right to demand this.
However I do think that the "top level" mailinglists deserve one. Eg if
you're unsure, linux-kernel. if you know it's ide, linux-ide etc etc.
On Thu, Mar 02 2006, Arjan van de Ven wrote:
> On Thu, 2006-03-02 at 06:17 -0500, Jeff Garzik wrote:
> > Dominik Brodowski wrote:
> > > On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
> > >>Linux Kernel Mailing List wrote:
> > >>>commit 42935656914b813c99f91cbac421fe677a6f34ab
> > >>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
> > >>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
> > >>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
> > >>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
> >
> > >>Why was this not CC'd to the IDE maintainer, and linux-ide?
> >
> > > For it is trivial, PCMCIA-related and my time is very limited these days.
> >
> > That's pathetic. You couldn't even CC linux-kernel on your answer. And
> > this is not even the first or second time you've been asked to CC a
> > maintainer.
>
> I personally don't consider that maintainers have a right to demand
> CC's. Sure it's polite and good to CC them, but that's not the same as
> having the right to demand this.
How do you expect the patch to be picked up, if you don't cc the
maintainer? Looking up the maintainer is trivial. We can't always rely
on akpm forwarding patches, seems a lot saner to put the onus on the
submitter to make sure it gets to the right place.
Lack of time is really not an excuse. Looking up and cc'ing the
maintainer is certainly the least time consuming part of
producing/testing a patch.
--
Jens Axboe
Arjan van de Ven wrote:
> I personally don't consider that maintainers have a right to demand
> CC's. Sure it's polite and good to CC them, but that's not the same as
> having the right to demand this.
>
> However I do think that the "top level" mailinglists deserve one. Eg if
> you're unsure, linux-kernel. if you know it's ide, linux-ide etc etc.
Agreed. I should have said "maintainer or list". Brodo CC'd neither.
Jeff
On Thu, 2006-03-02 at 12:42 +0100, Jens Axboe wrote:
> On Thu, Mar 02 2006, Arjan van de Ven wrote:
> > On Thu, 2006-03-02 at 06:17 -0500, Jeff Garzik wrote:
> > > Dominik Brodowski wrote:
> > > > On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
> > > >>Linux Kernel Mailing List wrote:
> > > >>>commit 42935656914b813c99f91cbac421fe677a6f34ab
> > > >>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
> > > >>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
> > > >>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
> > > >>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
> > >
> > > >>Why was this not CC'd to the IDE maintainer, and linux-ide?
> > >
> > > > For it is trivial, PCMCIA-related and my time is very limited these days.
> > >
> > > That's pathetic. You couldn't even CC linux-kernel on your answer. And
> > > this is not even the first or second time you've been asked to CC a
> > > maintainer.
> >
> > I personally don't consider that maintainers have a right to demand
> > CC's. Sure it's polite and good to CC them, but that's not the same as
> > having the right to demand this.
>
> How do you expect the patch to be picked up, if you don't cc the
> maintainer?
in this case brodo IS the ide-cs maintainer arguably ;)
but if you do a kernel wide change (say add a parameter to a function)
being forced to look up 5000 different maintainers is nonsense.
sure sending it to lkml and some bigger lists is good for that, but
still.
On Thu, 2006-03-02 at 12:42 +0100, Jens Axboe wrote:
> On Thu, Mar 02 2006, Arjan van de Ven wrote:
> > On Thu, 2006-03-02 at 06:17 -0500, Jeff Garzik wrote:
> > > Dominik Brodowski wrote:
> > > > On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
> > > >>Linux Kernel Mailing List wrote:
> > > >>>commit 42935656914b813c99f91cbac421fe677a6f34ab
> > > >>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
> > > >>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
> > > >>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
> > > >>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
> > >
> > > >>Why was this not CC'd to the IDE maintainer, and linux-ide?
> > >
> > > > For it is trivial, PCMCIA-related and my time is very limited these days.
> > >
> > > That's pathetic. You couldn't even CC linux-kernel on your answer. And
> > > this is not even the first or second time you've been asked to CC a
> > > maintainer.
> >
> > I personally don't consider that maintainers have a right to demand
> > CC's. Sure it's polite and good to CC them, but that's not the same as
> > having the right to demand this.
>
> How do you expect the patch to be picked up, if you don't cc the
> maintainer? Looking up the maintainer is trivial. We can't always rely
> on akpm forwarding patches, seems a lot saner to put the onus on the
> submitter to make sure it gets to the right place.
ok so this was adding a PCMCIA ID to a PCMCIA IDE driver. The patch was
mailed first to the pcmcia mailing list.
This is one of those drivers that hits multiple maintainers, arguable
Dominik is the primary maintainer of this driver. The patch doesn't do
ANYTHING structural to the driver, all it does it adds a device ID.
there is therefore zero IDE related change in it.
Sure I can it being nice to CC linux-ide ANYWAY, but, to be honest,
while I see that is important for changes to the driver that change the
structure of it and how it interacts with the IDE layer, I fail to see
the hard required reason for that for just adding a *PCMCIA* ID.
I think Jeff is a bit overreacting in this case.
Arjan van de Ven wrote:
> On Thu, 2006-03-02 at 12:42 +0100, Jens Axboe wrote:
>
>>On Thu, Mar 02 2006, Arjan van de Ven wrote:
>>
>>>On Thu, 2006-03-02 at 06:17 -0500, Jeff Garzik wrote:
>>>
>>>>Dominik Brodowski wrote:
>>>>
>>>>>On Wed, Mar 01, 2006 at 06:36:17PM -0500, Jeff Garzik wrote:
>>>>>
>>>>>>Linux Kernel Mailing List wrote:
>>>>>>
>>>>>>>commit 42935656914b813c99f91cbac421fe677a6f34ab
>>>>>>>tree d37a0d20998f4d87a4bd014300f707c3852ef5f9
>>>>>>>parent 82d56e6d2e616bee0e712330bad06b634f007a46
>>>>>>>author David Brownell <[email protected]> Wed, 25 Jan 2006 22:36:32 -0800
>>>>>>>committer Dominik Brodowski <[email protected]> Wed, 01 Mar 2006
>>>>
>>>>>>Why was this not CC'd to the IDE maintainer, and linux-ide?
>>>>
>>>>>For it is trivial, PCMCIA-related and my time is very limited these days.
>>>>
>>>>That's pathetic. You couldn't even CC linux-kernel on your answer. And
>>>>this is not even the first or second time you've been asked to CC a
>>>>maintainer.
>>>
>>>I personally don't consider that maintainers have a right to demand
>>>CC's. Sure it's polite and good to CC them, but that's not the same as
>>>having the right to demand this.
>>
>>How do you expect the patch to be picked up, if you don't cc the
>>maintainer? Looking up the maintainer is trivial. We can't always rely
>>on akpm forwarding patches, seems a lot saner to put the onus on the
>>submitter to make sure it gets to the right place.
>
>
>
> ok so this was adding a PCMCIA ID to a PCMCIA IDE driver. The patch was
> mailed first to the pcmcia mailing list.
>
> This is one of those drivers that hits multiple maintainers, arguable
> Dominik is the primary maintainer of this driver. The patch doesn't do
> ANYTHING structural to the driver, all it does it adds a device ID.
> there is therefore zero IDE related change in it.
>
> Sure I can it being nice to CC linux-ide ANYWAY, but, to be honest,
> while I see that is important for changes to the driver that change the
> structure of it and how it interacts with the IDE layer, I fail to see
> the hard required reason for that for just adding a *PCMCIA* ID.
>
> I think Jeff is a bit overreacting in this case.
About a quarter of the time when non-netdev maintainers add IDs, through
the magic of merges, we've wound up with duplicate IDs in the driver.
I've snipped several duplicate IDs from tulip and other net drivers over
the years.
Further, in the past Brodo has _already_ been asked to CC relevant
maintainers and lists -- or at least LKML -- with his patches. He has
established a pattern of lacking time to add CC's to his emails; it
wasn't just this incident.
Where is the peer review?
Jeff
On Thu, Mar 02, 2006 at 07:15:03AM -0500, Jeff Garzik wrote:
> Arjan van de Ven wrote:
> >Sure I can it being nice to CC linux-ide ANYWAY, but, to be honest,
> >while I see that is important for changes to the driver that change the
> >structure of it and how it interacts with the IDE layer, I fail to see
> >the hard required reason for that for just adding a *PCMCIA* ID.
> >
> >I think Jeff is a bit overreacting in this case.
>
> About a quarter of the time when non-netdev maintainers add IDs, through
> the magic of merges, we've wound up with duplicate IDs in the driver.
> I've snipped several duplicate IDs from tulip and other net drivers over
> the years.
>
> Further, in the past Brodo has _already_ been asked to CC relevant
> maintainers and lists -- or at least LKML -- with his patches. He has
> established a pattern of lacking time to add CC's to his emails; it
> wasn't just this incident.
>
> Where is the peer review?
I think it's fairly safe and obvious to say that Dominik is the peer
review for these tables - he _is_ the PCMCIA maintainer, he _is_
arguably the maintainer for the ide-cs driver, he _is_ the person
who invented these tables, he _is_ the one taking patches from people
to add IDs, he _is_ the one reviewing such patches.
If you want to know what's going on in PCMCIA land, subscribe to
linux-pcmcia. In the same way that if you want to know what's going
in in IDE land, you subscribe to linux-ide, or PCI land linux-pci,
SCSI land linux-scsi, network land netdev.
Using your argument (which seems to be demanding that any patch to
any IDE driver no matter how trivial must be on linux-ide) that a patch
to a PCI network device driver must be copied to linux-pci and netdev
even though it may not touch the PCI specific code.
What about a cardbus network card? Should any patch no matter how
trivial be sent to netdev, linux-pci and linux-pcmcia - all those three
mailing lists are within the "sphere of influence" of any patch to that
driver.
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
> About a quarter of the time when non-netdev maintainers add IDs, through
> the magic of merges, we've wound up with duplicate IDs in the driver.
> I've snipped several duplicate IDs from tulip and other net drivers over
> the years.
sure. But in this case Dominik IS the maintainer
> Further, in the past Brodo has _already_ been asked to CC relevant
> maintainers and lists -- or at least LKML -- with his patches.
he mailed the relevant list, linux-pcmcia ... whats wrong?
Maybe you are right about the pattern, but this time it's not that...
maybe you should apologize and flame Dominik the next time instead ;)
Russell King wrote:
> I think it's fairly safe and obvious to say that Dominik is the peer
> review for these tables - he _is_ the PCMCIA maintainer, he _is_
> arguably the maintainer for the ide-cs driver, he _is_ the person
> who invented these tables, he _is_ the one taking patches from people
> to add IDs, he _is_ the one reviewing such patches.
>
> If you want to know what's going on in PCMCIA land, subscribe to
> linux-pcmcia. In the same way that if you want to know what's going
> in in IDE land, you subscribe to linux-ide, or PCI land linux-pci,
> SCSI land linux-scsi, network land netdev.
>
> Using your argument (which seems to be demanding that any patch to
> any IDE driver no matter how trivial must be on linux-ide) that a patch
> to a PCI network device driver must be copied to linux-pci and netdev
> even though it may not touch the PCI specific code.
IDE driver -> IDE reviewers
network driver -> network reviewers
The bus associated with the driver is only a tiny detail. Many drivers
(IDE!) are multi-bus, even.
Linus occasionally complains about stuff hiding on non-LKML lists...
Even a CC to LKML would have been sufficient here. That was not done.
Jeff
Arjan van de Ven wrote:
>>About a quarter of the time when non-netdev maintainers add IDs, through
>>the magic of merges, we've wound up with duplicate IDs in the driver.
>>I've snipped several duplicate IDs from tulip and other net drivers over
>>the years.
>
>
> sure. But in this case Dominik IS the maintainer
Bus is far less relevant. Should sbus and SoC bus patches skip
subsystem peer review? Of course not. PCMCIA is not a special case here.
>>Further, in the past Brodo has _already_ been asked to CC relevant
>>maintainers and lists -- or at least LKML -- with his patches.
>
>
> he mailed the relevant list, linux-pcmcia ... whats wrong?
linux-pcmcia is not the place to talk about IDE and net driver patches.
Jeff
On Thu, 2006-03-02 at 07:38 -0500, Jeff Garzik wrote:
> Russell King wrote:
> > I think it's fairly safe and obvious to say that Dominik is the peer
> > review for these tables - he _is_ the PCMCIA maintainer, he _is_
> > arguably the maintainer for the ide-cs driver, he _is_ the person
> > who invented these tables, he _is_ the one taking patches from people
> > to add IDs, he _is_ the one reviewing such patches.
> >
> > If you want to know what's going on in PCMCIA land, subscribe to
> > linux-pcmcia. In the same way that if you want to know what's going
> > in in IDE land, you subscribe to linux-ide, or PCI land linux-pci,
> > SCSI land linux-scsi, network land netdev.
> >
> > Using your argument (which seems to be demanding that any patch to
> > any IDE driver no matter how trivial must be on linux-ide) that a patch
> > to a PCI network device driver must be copied to linux-pci and netdev
> > even though it may not touch the PCI specific code.
>
> IDE driver -> IDE reviewers
>
> network driver -> network reviewers
I would turn this around
IDE impacting change -> IDE reviewers
network driver impacting change -> network reviewers
the reason for the review is because of the impact, not because a file
happens to live in a certain directory.
Jeff,
On Thu, Mar 02, 2006 at 07:15:03AM -0500, Jeff Garzik wrote:
> Further, in the past Brodo has _already_ been asked to CC relevant
> maintainers and lists -- or at least LKML -- with his patches. He has
> established a pattern of lacking time to add CC's to his emails; it
> wasn't just this incident.
I've done so in the past, after you first told me of your wish, so it is no
pattern. Also I'll try to do so again in future -- I just missed it this
time. Please accept my apology for this oversight, it wasn't done on purpose.
> Where is the peer review?
Patches are sent to linux-pcmcia for review. As noted above, before and in
private communication, I'll try to remember to CC all others who might have
an interest in the patches.
Dominik
Dominik Brodowski wrote:
> Jeff,
>
> On Thu, Mar 02, 2006 at 07:15:03AM -0500, Jeff Garzik wrote:
>
>>Further, in the past Brodo has _already_ been asked to CC relevant
>>maintainers and lists -- or at least LKML -- with his patches. He has
>>established a pattern of lacking time to add CC's to his emails; it
>>wasn't just this incident.
>
>
> I've done so in the past, after you first told me of your wish, so it is no
> pattern. Also I'll try to do so again in future -- I just missed it this
> time. Please accept my apology for this oversight, it wasn't done on purpose.
Fair enough. I apologize for blowing up at you. It was the "lack of
time" comment that led me to where we are now.
Jeff