Message ID | 1400090591-5064-1-git-send-email-t.figa@samsung.com |
---|---|
State | New |
Headers | show |
On 05/15/14 03:03, Tomasz Figa wrote: Hi Mike, I've talked to Tomasz about current samsung related clock stuff. Since they are mostly having dependency on samsung tree now not clock core stuff, so would be better if it could be sent to upstream via samsung tree. And as you know, updating arch/arm/ and clock stuff are usually required for adding new SoC or supporting CCF newly... How do you think? Basically I need your agreement for it. Thanks, Kukjin > The following changes since commit 2916f9a2c6d9200b4c840a613cd1fa1dad04240f: > > ARM: S3C24XX: remove SAMSUNG_CLOCK remnants after ccf conversion (2014-05-13 08:00:46 +0900) > > are available in the git repository at: > > git://git.kernel.org/pub/scm/linux/kernel/git/tfiga/samsung-clk.git for_3.16/exynos5260 > > for you to fetch changes up to d39e55e06371c1ba9d11f4a17a56a0f925d12415: > > clk/exynos5260: add clock file for exynos5260 (2014-05-14 19:16:55 +0200) > > ---------------------------------------------------------------- > > Samsung clock Exynos5260 support for v3.16. > > This pull request contains patches preparing Samsung Common Clock Framework > helpers to support Exynos5260 by adding support for multiple clock providers > and then adding clock driver for Exynos5260. > > ---------------------------------------------------------------- > Pankaj Dubey (1): > clk/samsung: add support for pll2550xx > > Rahul Sharma (4): > clk/samsung: add support for multiple clock providers > clk/samsung: add support for pll2650xx > clk/exynos5260: add macros and documentation for exynos5260 > clk/exynos5260: add clock file for exynos5260 > > .../devicetree/bindings/clock/exynos5260-clock.txt | 190 ++ > drivers/clk/samsung/Makefile | 1 + > drivers/clk/samsung/clk-exynos4.c | 47 +- > drivers/clk/samsung/clk-exynos5250.c | 25 +- > drivers/clk/samsung/clk-exynos5260.c | 1980 ++++++++++++++++++++ > drivers/clk/samsung/clk-exynos5260.h | 459 +++++ > drivers/clk/samsung/clk-exynos5420.c | 24 +- > drivers/clk/samsung/clk-exynos5440.c | 18 +- > drivers/clk/samsung/clk-pll.c | 223 ++- > drivers/clk/samsung/clk-pll.h | 2 + > drivers/clk/samsung/clk-s3c2410.c | 51 +- > drivers/clk/samsung/clk-s3c2412.c | 29 +- > drivers/clk/samsung/clk-s3c2443.c | 46 +- > drivers/clk/samsung/clk-s3c64xx.c | 44 +- > drivers/clk/samsung/clk.c | 114 +- > drivers/clk/samsung/clk.h | 72 +- > include/dt-bindings/clock/exynos5260-clk.h | 469 +++++ > 17 files changed, 3589 insertions(+), 205 deletions(-) > create mode 100644 Documentation/devicetree/bindings/clock/exynos5260-clock.txt > create mode 100644 drivers/clk/samsung/clk-exynos5260.c > create mode 100644 drivers/clk/samsung/clk-exynos5260.h > create mode 100644 include/dt-bindings/clock/exynos5260-clk.h > --
Quoting Kukjin Kim (2014-05-14 12:59:22) > On 05/15/14 03:03, Tomasz Figa wrote: > > Hi Mike, > > I've talked to Tomasz about current samsung related clock stuff. Since > they are mostly having dependency on samsung tree now not clock core > stuff, so would be better if it could be sent to upstream via samsung > tree. And as you know, updating arch/arm/ and clock stuff are usually > required for adding new SoC or supporting CCF newly... The Samsung clk pull requests only touch two arch/arm Kconfig files and one dtsi file. That's not a lot of arch/arm churn. Is there a strong reason that this needs to go through the samsung/arm-soc trees? Otherwise it should continue to go through the clk tree. > > How do you think? Basically I need your agreement for it. Based on the above pull requests I do not see the need for changing how code gets merged. Regards, Mike > > Thanks, > Kukjin > > > > The following changes since commit 2916f9a2c6d9200b4c840a613cd1fa1dad04240f: > > > > ARM: S3C24XX: remove SAMSUNG_CLOCK remnants after ccf conversion (2014-05-13 08:00:46 +0900) > > > > are available in the git repository at: > > > > git://git.kernel.org/pub/scm/linux/kernel/git/tfiga/samsung-clk.git for_3.16/exynos5260 > > > > for you to fetch changes up to d39e55e06371c1ba9d11f4a17a56a0f925d12415: > > > > clk/exynos5260: add clock file for exynos5260 (2014-05-14 19:16:55 +0200) > > > > ---------------------------------------------------------------- > > > > Samsung clock Exynos5260 support for v3.16. > > > > This pull request contains patches preparing Samsung Common Clock Framework > > helpers to support Exynos5260 by adding support for multiple clock providers > > and then adding clock driver for Exynos5260. > > > > ---------------------------------------------------------------- > > Pankaj Dubey (1): > > clk/samsung: add support for pll2550xx > > > > Rahul Sharma (4): > > clk/samsung: add support for multiple clock providers > > clk/samsung: add support for pll2650xx > > clk/exynos5260: add macros and documentation for exynos5260 > > clk/exynos5260: add clock file for exynos5260 > > > > .../devicetree/bindings/clock/exynos5260-clock.txt | 190 ++ > > drivers/clk/samsung/Makefile | 1 + > > drivers/clk/samsung/clk-exynos4.c | 47 +- > > drivers/clk/samsung/clk-exynos5250.c | 25 +- > > drivers/clk/samsung/clk-exynos5260.c | 1980 ++++++++++++++++++++ > > drivers/clk/samsung/clk-exynos5260.h | 459 +++++ > > drivers/clk/samsung/clk-exynos5420.c | 24 +- > > drivers/clk/samsung/clk-exynos5440.c | 18 +- > > drivers/clk/samsung/clk-pll.c | 223 ++- > > drivers/clk/samsung/clk-pll.h | 2 + > > drivers/clk/samsung/clk-s3c2410.c | 51 +- > > drivers/clk/samsung/clk-s3c2412.c | 29 +- > > drivers/clk/samsung/clk-s3c2443.c | 46 +- > > drivers/clk/samsung/clk-s3c64xx.c | 44 +- > > drivers/clk/samsung/clk.c | 114 +- > > drivers/clk/samsung/clk.h | 72 +- > > include/dt-bindings/clock/exynos5260-clk.h | 469 +++++ > > 17 files changed, 3589 insertions(+), 205 deletions(-) > > create mode 100644 Documentation/devicetree/bindings/clock/exynos5260-clock.txt > > create mode 100644 drivers/clk/samsung/clk-exynos5260.c > > create mode 100644 drivers/clk/samsung/clk-exynos5260.h > > create mode 100644 include/dt-bindings/clock/exynos5260-clk.h > > --
Hi Mike, On 14.05.2014 22:13, Mike Turquette wrote: > Quoting Kukjin Kim (2014-05-14 12:59:22) >> On 05/15/14 03:03, Tomasz Figa wrote: >> >> Hi Mike, >> >> I've talked to Tomasz about current samsung related clock stuff. Since >> they are mostly having dependency on samsung tree now not clock core >> stuff, so would be better if it could be sent to upstream via samsung >> tree. And as you know, updating arch/arm/ and clock stuff are usually >> required for adding new SoC or supporting CCF newly... > > The Samsung clk pull requests only touch two arch/arm Kconfig files and > one dtsi file. That's not a lot of arch/arm churn. Is there a strong > reason that this needs to go through the samsung/arm-soc trees? > Otherwise it should continue to go through the clk tree. Obviously they are patches for Samsung clock drivers. ;) The issue here is that there is a number of patches already merged in Samsung tree on which the patches discussed here depend. > >> >> How do you think? Basically I need your agreement for it. > > Based on the above pull requests I do not see the need for changing how > code gets merged. As long as there are no dependencies on arch code and series being applied do not touch arch code, this is perfectly fine. Unfortunately this is rarely the case, at least for Samsung platforms and at least for now. After we finish with arch clean-up and move all code to appropriate subsystems, it should become easier, though. Best regards, Tomasz
Quoting Tomasz Figa (2014-05-14 13:20:14) > Hi Mike, > > On 14.05.2014 22:13, Mike Turquette wrote: > > Quoting Kukjin Kim (2014-05-14 12:59:22) > >> On 05/15/14 03:03, Tomasz Figa wrote: > >> > >> Hi Mike, > >> > >> I've talked to Tomasz about current samsung related clock stuff. Since > >> they are mostly having dependency on samsung tree now not clock core > >> stuff, so would be better if it could be sent to upstream via samsung > >> tree. And as you know, updating arch/arm/ and clock stuff are usually > >> required for adding new SoC or supporting CCF newly... > > > > The Samsung clk pull requests only touch two arch/arm Kconfig files and > > one dtsi file. That's not a lot of arch/arm churn. Is there a strong > > reason that this needs to go through the samsung/arm-soc trees? > > Otherwise it should continue to go through the clk tree. > > Obviously they are patches for Samsung clock drivers. ;) > > The issue here is that there is a number of patches already merged in > Samsung tree on which the patches discussed here depend. OK, I think I misread the original email. I thought you were asking for future pull requests to go through the samsung tree, but you only mean the ones in this thread. No problem there. Acked-by: Mike Turquette <mturquette@linaro.org> Regards, Mike > > > > >> > >> How do you think? Basically I need your agreement for it. > > > > Based on the above pull requests I do not see the need for changing how > > code gets merged. > > As long as there are no dependencies on arch code and series being > applied do not touch arch code, this is perfectly fine. Unfortunately > this is rarely the case, at least for Samsung platforms and at least for > now. After we finish with arch clean-up and move all code to appropriate > subsystems, it should become easier, though. > > Best regards, > Tomasz
On 15.05.2014 00:07, Mike Turquette wrote: > Quoting Tomasz Figa (2014-05-14 13:20:14) >> Hi Mike, >> >> On 14.05.2014 22:13, Mike Turquette wrote: >>> Quoting Kukjin Kim (2014-05-14 12:59:22) >>>> On 05/15/14 03:03, Tomasz Figa wrote: >>>> >>>> Hi Mike, >>>> >>>> I've talked to Tomasz about current samsung related clock stuff. Since >>>> they are mostly having dependency on samsung tree now not clock core >>>> stuff, so would be better if it could be sent to upstream via samsung >>>> tree. And as you know, updating arch/arm/ and clock stuff are usually >>>> required for adding new SoC or supporting CCF newly... >>> >>> The Samsung clk pull requests only touch two arch/arm Kconfig files and >>> one dtsi file. That's not a lot of arch/arm churn. Is there a strong >>> reason that this needs to go through the samsung/arm-soc trees? >>> Otherwise it should continue to go through the clk tree. >> >> Obviously they are patches for Samsung clock drivers. ;) >> >> The issue here is that there is a number of patches already merged in >> Samsung tree on which the patches discussed here depend. > > OK, I think I misread the original email. I thought you were asking for > future pull requests to go through the samsung tree, but you only mean > the ones in this thread. No problem there. > > Acked-by: Mike Turquette <mturquette@linaro.org> Thanks. As you probably noticed with my pull request for 3.15, we've been getting quite a lot of burden due to dependencies between arch and clk patches lastly, but this is inevitable when we are moving things out of arch. The good news is that after patches from this pull request series, we will end up with just one, more or less active platform (s5pv210) that needs to be converted (and work already in progress). Best regards, Tomasz
On 05/15/14 07:16, Tomasz Figa wrote: > On 15.05.2014 00:07, Mike Turquette wrote: >> Quoting Tomasz Figa (2014-05-14 13:20:14) >>> Hi Mike, >>> >>> On 14.05.2014 22:13, Mike Turquette wrote: >>>> Quoting Kukjin Kim (2014-05-14 12:59:22) >>>>> On 05/15/14 03:03, Tomasz Figa wrote: >>>>> >>>>> Hi Mike, >>>>> >>>>> I've talked to Tomasz about current samsung related clock stuff. Since >>>>> they are mostly having dependency on samsung tree now not clock core >>>>> stuff, so would be better if it could be sent to upstream via samsung >>>>> tree. And as you know, updating arch/arm/ and clock stuff are usually >>>>> required for adding new SoC or supporting CCF newly... >>>> >>>> The Samsung clk pull requests only touch two arch/arm Kconfig files and >>>> one dtsi file. That's not a lot of arch/arm churn. Is there a strong >>>> reason that this needs to go through the samsung/arm-soc trees? >>>> Otherwise it should continue to go through the clk tree. >>> >>> Obviously they are patches for Samsung clock drivers. ;) >>> >>> The issue here is that there is a number of patches already merged in >>> Samsung tree on which the patches discussed here depend. >> >> OK, I think I misread the original email. I thought you were asking for >> future pull requests to go through the samsung tree, but you only mean >> the ones in this thread. No problem there. >> >> Acked-by: Mike Turquette<mturquette@linaro.org> > Mike, thanks for your ack on this whole pull-requests. I will pull 1 to 4 into samsung tree for 3.16. > Thanks. > > As you probably noticed with my pull request for 3.15, we've been > getting quite a lot of burden due to dependencies between arch and clk > patches lastly, but this is inevitable when we are moving things out of > arch. > > The good news is that after patches from this pull request series, we > will end up with just one, more or less active platform (s5pv210) that > needs to be converted (and work already in progress). > Sounds great. Thanks, Kukjin
On 05/16/14 06:11, Kukjin Kim wrote: [...] >>> OK, I think I misread the original email. I thought you were asking for >>> future pull requests to go through the samsung tree, but you only mean >>> the ones in this thread. No problem there. >>> >>> Acked-by: Mike Turquette<mturquette@linaro.org> >> > Mike, thanks for your ack on this whole pull-requests. > > I will pull 1 to 4 into samsung tree for 3.16. > Done, thanks. - Kukjin