Please remind 'WARNING: about Samsung merge window over'

Please remind 'WARNING: about Samsung merge window over'

[Kukjin Kim]

Please remind following which is from discussion about Samsung merge stuff
with Russell.

---

> > So, please send the majority of patches *before* the merge window and
> > try to avoid sending anything but fixes during the merge window itself.
> > The merge window is for code which has _already_ been merged and has
> > been in linux-next to be merged into Linus' tree.  It's not for new
> > code to be reviewed.
> 
> Next merge window, which will probably be the last I will handle for the
> S5P range, will have an earlier cut-off point. It was a big rush to get
> this lot reviewed and merged.

It also means that *NONE* of the stuff in your second pull request was
in linux-next for *any* time.  It also conflicted with my tree.  The
result of that is I'm expecting mainline to be dead for all ARM with:

arch/arm/Kconfig:883: can't open file "arch/arm/plat-s5pc1xx/Kconfig"

because of a mis-merge - and since that conflict was not picked up by
linux-next, it could not have been previously published.

For the next merge window, I will avoid reading email for the duration
of the window - which means those who are late for the merge window will
simply not have their patches and trees pulled.  This means we _will_
have time to resolve conflicts and find problems _before_ the merge
window

---

I have no more plan to merge s5p stuff for this merge window, but reviewed
here until now.

Thanks.

Best regards,
Kgene.
--
Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
SW Solution Development Team, Samsung Electronics Co., Ltd.

Please remind 'WARNING: about Samsung merge window over'

[Kyungmin Park]

Absurd !!!

Did you mentioned that samsung merge window over before?
why it's reminder mail? it's unilateral notification at this time.

In my feelings. only include the your side (LSI) patches and don't
include others (mobile). since it's not fit your taste.
Can you explain the difference between "S5PV310 CPU support" and
"S5PV210 cpufreq support"?
Did you get any ACKs for V310 support? Why do you exclude the v210
cpufreq support even though it's acked from others?

On Mon, Aug 2, 2010 at 2:52 PM, Kukjin Kim <kgene.kim@samsung.com> wrote:
> Please remind following which is from discussion about Samsung merge stuff
> with Russell.
>
> ---
>
>> > So, please send the majority of patches *before* the merge window and
>> > try to avoid sending anything but fixes during the merge window itself.
>> > The merge window is for code which has _already_ been merged and has
>> > been in linux-next to be merged into Linus' tree. ?It's not for new
>> > code to be reviewed.
>>
>> Next merge window, which will probably be the last I will handle for the
>> S5P range, will have an earlier cut-off point. It was a big rush to get
>> this lot reviewed and merged.
>
> It also means that *NONE* of the stuff in your second pull request was
> in linux-next for *any* time. ?It also conflicted with my tree. ?The
> result of that is I'm expecting mainline to be dead for all ARM with:
>
> arch/arm/Kconfig:883: can't open file "arch/arm/plat-s5pc1xx/Kconfig"
>
> because of a mis-merge - and since that conflict was not picked up by
> linux-next, it could not have been previously published.
>
> For the next merge window, I will avoid reading email for the duration
> of the window - which means those who are late for the merge window will
> simply not have their patches and trees pulled. ?This means we _will_
> have time to resolve conflicts and find problems _before_ the merge
> window
>
> ---
>
> I have no more plan to merge s5p stuff for this merge window, but reviewed
> here until now.
>
> Thanks.
>
> Best regards,
> Kgene.
> --
> Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
> SW Solution Development Team, Samsung Electronics Co., Ltd.
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>

Please remind 'WARNING: about Samsung merge window over'

[Grant Likely]

On Mon, Aug 2, 2010 at 12:24 AM, Kyungmin Park <kmpark@infradead.org> wrote:
> Absurd !!!
>
> Did you mentioned that samsung merge window over before?
> why it's reminder mail? it's unilateral notification at this time.

Kgene doesn't have to.  As has been widely discussed, Linus is getting
stricter about the merge window process.  The window is for merging
what has already been collected and tested by maintainers.  It is not
for picking up new stuff.  Maintainers picking up new stuff late risk
Linus not merging their tree *at all*.

g.

>
> In my feelings. only include the your side (LSI) patches and don't
> include others (mobile). since it's not fit your taste.
> Can you explain the difference between "S5PV310 CPU support" and
> "S5PV210 cpufreq support"?
> Did you get any ACKs for V310 support? Why do you exclude the v210
> cpufreq support even though it's acked from others?
>
> On Mon, Aug 2, 2010 at 2:52 PM, Kukjin Kim <kgene.kim@samsung.com> wrote:
>> Please remind following which is from discussion about Samsung merge stuff
>> with Russell.
>>
>> ---
>>
>>> > So, please send the majority of patches *before* the merge window and
>>> > try to avoid sending anything but fixes during the merge window itself.
>>> > The merge window is for code which has _already_ been merged and has
>>> > been in linux-next to be merged into Linus' tree. ?It's not for new
>>> > code to be reviewed.
>>>
>>> Next merge window, which will probably be the last I will handle for the
>>> S5P range, will have an earlier cut-off point. It was a big rush to get
>>> this lot reviewed and merged.
>>
>> It also means that *NONE* of the stuff in your second pull request was
>> in linux-next for *any* time. ?It also conflicted with my tree. ?The
>> result of that is I'm expecting mainline to be dead for all ARM with:
>>
>> arch/arm/Kconfig:883: can't open file "arch/arm/plat-s5pc1xx/Kconfig"
>>
>> because of a mis-merge - and since that conflict was not picked up by
>> linux-next, it could not have been previously published.
>>
>> For the next merge window, I will avoid reading email for the duration
>> of the window - which means those who are late for the merge window will
>> simply not have their patches and trees pulled. ?This means we _will_
>> have time to resolve conflicts and find problems _before_ the merge
>> window
>>
>> ---
>>
>> I have no more plan to merge s5p stuff for this merge window, but reviewed
>> here until now.
>>
>> Thanks.
>>
>> Best regards,
>> Kgene.
>> --
>> Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
>> SW Solution Development Team, Samsung Electronics Co., Ltd.
>>
>>
>> _______________________________________________
>> linux-arm-kernel mailing list
>> linux-arm-kernel at lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>



-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.

Please remind 'WARNING: about Samsung merge window over'

[Russell King - ARM Linux]

On Mon, Aug 02, 2010 at 03:24:13PM +0900, Kyungmin Park wrote:
> Absurd !!!
> 
> Did you mentioned that samsung merge window over before?
> why it's reminder mail? it's unilateral notification at this time.

It's standard Linux kernel development procedure.

The merge window is not for new code - the merge window is for code which
has already been merged by maintainers to submit it to Linus in an orderly
manner and fixup any conflicts which may occur.

New code which hasn't been reviewed by the beginning of the merge window
and which hasn't been in linux-next for a _reasonable_ amount of time
should wait until the following merge window unless there's a _very_ good
reason to make an exception.

Please remind 'WARNING: about Samsung merge window over'

[Kyungmin Park]

On Mon, Aug 2, 2010 at 4:31 PM, Russell King - ARM Linux
<linux@arm.linux.org.uk> wrote:
> On Mon, Aug 02, 2010 at 03:24:13PM +0900, Kyungmin Park wrote:
>> Absurd !!!
>>
>> Did you mentioned that samsung merge window over before?
>> why it's reminder mail? it's unilateral notification at this time.
>
> It's standard Linux kernel development procedure.
>
> The merge window is not for new code - the merge window is for code which
> has already been merged by maintainers to submit it to Linus in an orderly
> manner and fixup any conflicts which may occur.
>
> New code which hasn't been reviewed by the beginning of the merge window
> and which hasn't been in linux-next for a _reasonable_ amount of time
> should wait until the following merge window unless there's a _very_ good
> reason to make an exception.

I don't complain the merge window procedure. Even though follow the
normal procedure, send the patches and reviewed before the merge
windows. it's not included at this merge window. Our team tries to
include some features but will wait next merge window. Frankly I'm not
sure our implementation can be merged since he focus on their the
latest chip but I want to cover the existing chips even though it's
not released outside.

As the maintainer, he merged his features freely but others can't get
the chance if the implementation is not fit his taste or company
policy.

Look at the example.
[PATCH v2 0/4] ARM: S5P: Support gpio interrupts
http://marc.info/?l=linux-arm-kernel&m=127840208508625&w=2

We need the gpio interrupt support but he refused it since it used too
many irq. Actually his board don't use this features.
"It is because there are too many gpio interrupts and having support
of all of them is unnecessary as realistically only few of them maybe
used."
And current kernel don't support the sparse irq feature. then it's
reasonable to merge it first and revise it later.

Another why FIMC support is missing? we modified it as his requested
and send it properly.
[PATCH v3 3/8] ARM: Samsung: Add platform definitions and helpers for
http://marc.info/?l=linux-arm-kernel&m=127990218813931&w=2
[PATCH v3 1/8] ARM: Samsung: Add register definitions for Samsung S5P
http://marc.info/?l=linux-arm-kernel&m=127990218613922&w=2

I want to listen other maintainers opinions.

Thank you,
Kyungmin Park

Please remind 'WARNING: about Samsung merge window over'

[Russell King - ARM Linux]

On Mon, Aug 02, 2010 at 05:35:37PM +0900, Kyungmin Park wrote:
> As the maintainer, he merged his features freely but others can't get
> the chance if the implementation is not fit his taste or company
> policy.
> 
> Look at the example.
> [PATCH v2 0/4] ARM: S5P: Support gpio interrupts
> http://marc.info/?l=linux-arm-kernel&m=127840208508625&w=2
> 
> We need the gpio interrupt support but he refused it since it used too
> many irq. Actually his board don't use this features.
> "It is because there are too many gpio interrupts and having support
> of all of them is unnecessary as realistically only few of them maybe
> used."

Please remind 'WARNING: about Samsung merge window over'

[Kukjin Kim]

Russell King wrote:
> 
> On Mon, Aug 02, 2010 at 05:35:37PM +0900, Kyungmin Park wrote:
> > As the maintainer, he merged his features freely but others can't get
> > the chance if the implementation is not fit his taste or company
> > policy.
> >
> > Look at the example.
> > [PATCH v2 0/4] ARM: S5P: Support gpio interrupts
> > http://marc.info/?l=linux-arm-kernel&m=127840208508625&w=2
> >
> > We need the gpio interrupt support but he refused it since it used too
> > many irq. Actually his board don't use this features.
> > "It is because there are too many gpio interrupts and having support
> > of all of them is unnecessary as realistically only few of them maybe
> > used."
> 
> From my count, there's already 144 IRQs, and you'll be adding 27*8 = 216
> additional IRQs to that.  That's quite small compared to some platforms
> which have in the order of 512 or even 1024 IRQs.
> 
> > And current kernel don't support the sparse irq feature. then it's
> > reasonable to merge it first and revise it later.
> 
> sparse irq support has been queued for almost a month for the merge
> window which has just this morning opened.  This doesn't help you if
> you instantiate all your 360 interrupts though - just because you
> don't _use_ an interrupt which has been declared as existing doesn't
> reduce the size of the arrays.
> 
> > Another why FIMC support is missing? we modified it as his requested
> > and send it properly.
> > [PATCH v3 3/8] ARM: Samsung: Add platform definitions and helpers for
> > http://marc.info/?l=linux-arm-kernel&m=127990218813931&w=2
> 
> This looks like it's been missed.  People get busy and miss things on
> the mailing list, there's nothing special about that.
> 
Yeah...actually I missed reply for it...so I requested re-submit v4 patch to
Marek with some modifying and finished review it just now.

> > [PATCH v3 1/8] ARM: Samsung: Add register definitions for Samsung S5P
> > http://marc.info/?l=linux-arm-kernel&m=127990218613922&w=2
> 
> Kukjin Kim replied to this one with a point requiring an answer, but
> nothing came back.
> 
>    "Looks ok...however, I'm still thinking whether really need all these
>     definitions."
> 
> Seems to be a perfectly reasonable point to raise, and if there's no
> reply to justify them...

Thanks.

Best regards,
Kgene.
--
Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
SW Solution Development Team, Samsung Electronics Co., Ltd.

Please remind 'WARNING: about Samsung merge window over'

[Kyungmin Park]

On Tue, Aug 3, 2010 at 9:53 AM, Kukjin Kim <kgene.kim@samsung.com> wrote:
> Russell King wrote:
>>
>> On Mon, Aug 02, 2010 at 05:35:37PM +0900, Kyungmin Park wrote:
>> > As the maintainer, he merged his features freely but others can't get
>> > the chance if the implementation is not fit his taste or company
>> > policy.
>> >
>> > Look at the example.
>> > [PATCH v2 0/4] ARM: S5P: Support gpio interrupts
>> > http://marc.info/?l=linux-arm-kernel&m=127840208508625&w=2
>> >
>> > We need the gpio interrupt support but he refused it since it used too
>> > many irq. Actually his board don't use this features.
>> > "It is because there are too many gpio interrupts and having support
>> > of all of them is unnecessary as realistically only few of them maybe
>> > used."
>>
>> From my count, there's already 144 IRQs, and you'll be adding 27*8 = 216
>> additional IRQs to that. ?That's quite small compared to some platforms
>> which have in the order of 512 or even 1024 IRQs.

Can you comment this one?
and previous s5pc110 cpufreq also?

>>
>> > And current kernel don't support the sparse irq feature. then it's
>> > reasonable to merge it first and revise it later.
>>
>> sparse irq support has been queued for almost a month for the merge
>> window which has just this morning opened. ?This doesn't help you if
>> you instantiate all your 360 interrupts though - just because you
>> don't _use_ an interrupt which has been declared as existing doesn't
>> reduce the size of the arrays.
>>
>> > Another why FIMC support is missing? we modified it as his requested
>> > and send it properly.
>> > [PATCH v3 3/8] ARM: Samsung: Add platform definitions and helpers for
>> > http://marc.info/?l=linux-arm-kernel&m=127990218813931&w=2
>>
>> This looks like it's been missed. ?People get busy and miss things on
>> the mailing list, there's nothing special about that.
>>
> Yeah...actually I missed reply for it...so I requested re-submit v4 patch to
> Marek with some modifying and finished review it just now.
>
>> > [PATCH v3 1/8] ARM: Samsung: Add register definitions for Samsung S5P
>> > http://marc.info/?l=linux-arm-kernel&m=127990218613922&w=2
>>
>> Kukjin Kim replied to this one with a point requiring an answer, but
>> nothing came back.
>>
>> ? ?"Looks ok...however, I'm still thinking whether really need all these
>> ? ? definitions."
>>
>> Seems to be a perfectly reasonable point to raise, and if there's no
>> reply to justify them...
>
> Thanks.
>
> Best regards,
> Kgene.
> --
> Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
> SW Solution Development Team, Samsung Electronics Co., Ltd.
>
>