Message ID | 20250424062154.2999219-1-dario.binacchi@amarulasolutions.com |
---|---|
Headers | show |
Series |
|
Related | show |
On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > This version keeps the version v9 patches that can be merged and > removes the patches that will need to be modified in case Peng's > PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > The idea is to speed up the merging of the patches in the series > that have already been reviewed and are not dependent on the > introduction of the assigned-clocks-sscs property, and postpone > the patches for spread spectrum to a future series once it becomes > clear what needs to be done. > > [...] Applied, thanks! [01/19] dt-bindings: clock: imx8mm: add VIDEO_PLL clocks commit: 20e5d201b5d8f830e702d7d183f6b1b246b78d8a [02/19] clk: imx8mm: rename video_pll1 to video_pll commit: 26a33196b5b68cf199b6c4283a254aa92d2aaf4b [03/19] dt-bindings: clock: imx8mp: add VIDEO_PLL clocks commit: 2d50415e2457c6f6621c2faa3b01b11150fb9c67 [04/19] clk: imx8mp: rename video_pll1 to video_pll commit: 21bb969f608cefd8d847cf6eb50a193d9f1fbb87 [05/19] dt-bindings: clock: imx8m-anatop: add oscillators and PLLs commit: 2ba124053687c933031a6dc5b2e16ceaca250934 [10/19] clk: imx: add hw API imx_anatop_get_clk_hw commit: 17e3c1a272d97e49b4f3fbfe1f1b889e120d2be8 [11/19] clk: imx: add support for i.MX8MM anatop clock driver commit: 3cbc38cf42ca42d2dc9a93c949e0381ff919df71 [12/19] clk: imx: add support for i.MX8MN anatop clock driver commit: 80badb1d7264e83b512475898e7459f464a009c9 [13/19] clk: imx: add support for i.MX8MP anatop clock driver commit: 4c82bbe8b5437c7f16b2891ce33210c0f1410597 [14/19] clk: imx8mp: rename ccm_base to base commit: 1a77907dbbecfbe5e6a1aec28afd49a1dc184b7a [16/19] dt-bindings: clock: imx8m-clock: add PLLs commit: 6a55647af3334f1d935ece67de4a838a864b53fc Best regards,
Hello Shawn, On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: > > > On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > > This version keeps the version v9 patches that can be merged and > > removes the patches that will need to be modified in case Peng's > > PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > > The idea is to speed up the merging of the patches in the series > > that have already been reviewed and are not dependent on the > > introduction of the assigned-clocks-sscs property, and postpone > > the patches for spread spectrum to a future series once it becomes > > clear what needs to be done. > > > > [...] > > Applied, thanks! > > [01/19] dt-bindings: clock: imx8mm: add VIDEO_PLL clocks > commit: 20e5d201b5d8f830e702d7d183f6b1b246b78d8a > [02/19] clk: imx8mm: rename video_pll1 to video_pll > commit: 26a33196b5b68cf199b6c4283a254aa92d2aaf4b > [03/19] dt-bindings: clock: imx8mp: add VIDEO_PLL clocks > commit: 2d50415e2457c6f6621c2faa3b01b11150fb9c67 > [04/19] clk: imx8mp: rename video_pll1 to video_pll > commit: 21bb969f608cefd8d847cf6eb50a193d9f1fbb87 > [05/19] dt-bindings: clock: imx8m-anatop: add oscillators and PLLs > commit: 2ba124053687c933031a6dc5b2e16ceaca250934 > [10/19] clk: imx: add hw API imx_anatop_get_clk_hw > commit: 17e3c1a272d97e49b4f3fbfe1f1b889e120d2be8 > [11/19] clk: imx: add support for i.MX8MM anatop clock driver > commit: 3cbc38cf42ca42d2dc9a93c949e0381ff919df71 > [12/19] clk: imx: add support for i.MX8MN anatop clock driver > commit: 80badb1d7264e83b512475898e7459f464a009c9 > [13/19] clk: imx: add support for i.MX8MP anatop clock driver > commit: 4c82bbe8b5437c7f16b2891ce33210c0f1410597 > [14/19] clk: imx8mp: rename ccm_base to base > commit: 1a77907dbbecfbe5e6a1aec28afd49a1dc184b7a > [16/19] dt-bindings: clock: imx8m-clock: add PLLs > commit: 6a55647af3334f1d935ece67de4a838a864b53fc > Please check the remaining patches, as they are required for correctly building the ones merged by Abel. The kernel test robot has already reported build errors. Thanks and regards, Dario > Best regards, > -- > Abel Vesa <abel.vesa@linaro.org>
On Thu, May 08, 2025 at 10:32:10AM +0200, Dario Binacchi wrote: > Hello Shawn, > > On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: > > > > > > On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > > > This version keeps the version v9 patches that can be merged and > > > removes the patches that will need to be modified in case Peng's > > > PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > > > The idea is to speed up the merging of the patches in the series > > > that have already been reviewed and are not dependent on the > > > introduction of the assigned-clocks-sscs property, and postpone > > > the patches for spread spectrum to a future series once it becomes > > > clear what needs to be done. > > > > > > [...] > > > > Applied, thanks! > > > > [01/19] dt-bindings: clock: imx8mm: add VIDEO_PLL clocks > > commit: 20e5d201b5d8f830e702d7d183f6b1b246b78d8a > > [02/19] clk: imx8mm: rename video_pll1 to video_pll > > commit: 26a33196b5b68cf199b6c4283a254aa92d2aaf4b > > [03/19] dt-bindings: clock: imx8mp: add VIDEO_PLL clocks > > commit: 2d50415e2457c6f6621c2faa3b01b11150fb9c67 > > [04/19] clk: imx8mp: rename video_pll1 to video_pll > > commit: 21bb969f608cefd8d847cf6eb50a193d9f1fbb87 > > [05/19] dt-bindings: clock: imx8m-anatop: add oscillators and PLLs > > commit: 2ba124053687c933031a6dc5b2e16ceaca250934 > > [10/19] clk: imx: add hw API imx_anatop_get_clk_hw > > commit: 17e3c1a272d97e49b4f3fbfe1f1b889e120d2be8 > > [11/19] clk: imx: add support for i.MX8MM anatop clock driver > > commit: 3cbc38cf42ca42d2dc9a93c949e0381ff919df71 > > [12/19] clk: imx: add support for i.MX8MN anatop clock driver > > commit: 80badb1d7264e83b512475898e7459f464a009c9 > > [13/19] clk: imx: add support for i.MX8MP anatop clock driver > > commit: 4c82bbe8b5437c7f16b2891ce33210c0f1410597 > > [14/19] clk: imx8mp: rename ccm_base to base > > commit: 1a77907dbbecfbe5e6a1aec28afd49a1dc184b7a > > [16/19] dt-bindings: clock: imx8m-clock: add PLLs > > commit: 6a55647af3334f1d935ece67de4a838a864b53fc > > > > Please check the remaining patches, as they are required for correctly > building the > ones merged by Abel. The kernel test robot has already reported build errors. I assume the remaining patches are DTS ones? If so, I do not see how clock drivers would require DTS change to build correctly. Do you have a pointer to the reported build errors? My understanding about the build dependency is the opposite, i.e. the DTS changes require clock defines merged by Abel to build? In that case, I suggest Abel pick up the whole series with my ack on DTS changes. Acked-by: Shawn Guo <shawnguo@kernel.org> Alternatively, I can pick up the remaining patches after clock changes land on mainline (the next -rc1). Shawn To unsubscribe from this group and stop receiving emails from it, send an email to linux-amarula+unsubscribe@amarulasolutions.com.
On Fri, May 9, 2025 at 2:02 PM Shawn Guo <shawnguo2@yeah.net> wrote: > > On Thu, May 08, 2025 at 10:32:10AM +0200, Dario Binacchi wrote: > > Hello Shawn, > > > > On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: > > > > > > > > > On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > > > > This version keeps the version v9 patches that can be merged and > > > > removes the patches that will need to be modified in case Peng's > > > > PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > > > > The idea is to speed up the merging of the patches in the series > > > > that have already been reviewed and are not dependent on the > > > > introduction of the assigned-clocks-sscs property, and postpone > > > > the patches for spread spectrum to a future series once it becomes > > > > clear what needs to be done. > > > > > > > > [...] > > > > > > Applied, thanks! > > > > > > [01/19] dt-bindings: clock: imx8mm: add VIDEO_PLL clocks > > > commit: 20e5d201b5d8f830e702d7d183f6b1b246b78d8a > > > [02/19] clk: imx8mm: rename video_pll1 to video_pll > > > commit: 26a33196b5b68cf199b6c4283a254aa92d2aaf4b > > > [03/19] dt-bindings: clock: imx8mp: add VIDEO_PLL clocks > > > commit: 2d50415e2457c6f6621c2faa3b01b11150fb9c67 > > > [04/19] clk: imx8mp: rename video_pll1 to video_pll > > > commit: 21bb969f608cefd8d847cf6eb50a193d9f1fbb87 > > > [05/19] dt-bindings: clock: imx8m-anatop: add oscillators and PLLs > > > commit: 2ba124053687c933031a6dc5b2e16ceaca250934 > > > [10/19] clk: imx: add hw API imx_anatop_get_clk_hw > > > commit: 17e3c1a272d97e49b4f3fbfe1f1b889e120d2be8 > > > [11/19] clk: imx: add support for i.MX8MM anatop clock driver > > > commit: 3cbc38cf42ca42d2dc9a93c949e0381ff919df71 > > > [12/19] clk: imx: add support for i.MX8MN anatop clock driver > > > commit: 80badb1d7264e83b512475898e7459f464a009c9 > > > [13/19] clk: imx: add support for i.MX8MP anatop clock driver > > > commit: 4c82bbe8b5437c7f16b2891ce33210c0f1410597 > > > [14/19] clk: imx8mp: rename ccm_base to base > > > commit: 1a77907dbbecfbe5e6a1aec28afd49a1dc184b7a > > > [16/19] dt-bindings: clock: imx8m-clock: add PLLs > > > commit: 6a55647af3334f1d935ece67de4a838a864b53fc > > > > > > > Please check the remaining patches, as they are required for correctly > > building the > > ones merged by Abel. The kernel test robot has already reported build errors. > > I assume the remaining patches are DTS ones? Yes > If so, I do not see how > clock drivers would require DTS change to build correctly. Do you have > a pointer to the reported build errors? https://lore.kernel.org/oe-kbuild-all/202505090537.ss8RbFln-lkp@intel.com/ DTC compilation fails. I think it's because of the patch [16/19] dt-bindings: clock: imx8m-clock: add PLLs which was merged without the corresponding DTS changes. Thanks and regards, Dario > > My understanding about the build dependency is the opposite, i.e. the > DTS changes require clock defines merged by Abel to build? In that > case, I suggest Abel pick up the whole series with my ack on DTS changes. > > Acked-by: Shawn Guo <shawnguo@kernel.org> > > Alternatively, I can pick up the remaining patches after clock changes > land on mainline (the next -rc1). > > Shawn >
On 09/05/2025 14:43, Dario Binacchi wrote: >>> >>> Please check the remaining patches, as they are required for correctly >>> building the >>> ones merged by Abel. The kernel test robot has already reported build errors. >> >> I assume the remaining patches are DTS ones? > Yes >> If so, I do not see how >> clock drivers would require DTS change to build correctly. Do you have >> a pointer to the reported build errors? > > https://lore.kernel.org/oe-kbuild-all/202505090537.ss8RbFln-lkp@intel.com/ > > DTC compilation fails. I think it's because of the patch > [16/19] dt-bindings: clock: imx8m-clock: add PLLs > which was merged without the corresponding DTS changes. NAK, why did you create patchset forcing DTS to be put into the driver subsystem? It is even beyond ABI break. Best regards, Krzysztof To unsubscribe from this group and stop receiving emails from it, send an email to linux-amarula+unsubscribe@amarulasolutions.com.
Hello Abel, On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: > > > On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > > This version keeps the version v9 patches that can be merged and > > removes the patches that will need to be modified in case Peng's > > PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > > The idea is to speed up the merging of the patches in the series > > that have already been reviewed and are not dependent on the > > introduction of the assigned-clocks-sscs property, and postpone > > the patches for spread spectrum to a future series once it becomes > > clear what needs to be done. > > > > [...] > > Applied, thanks! I was surprised to see that the series has been removed from linux-next. It’s been 8 months since the first version dated September 28, 2024. The most critical phase was version 3 - https://lore.kernel.org/all/20241106090549.3684963-1-dario.binacchi@amarulasolutions.com/ - where two key issues emerged: 1 The CCM design is flawed because "in the current design, CCM is taken as the producer of CLK_IMX8M_VIDEO_PLL, not the consumer." 2 A driver for anatop needs to be implemented because "using clocks to replace fsl,ssc-clocks is possible under CCM mode, but you need to develop the fsl,imx8mm-anatop clock driver." These development guidelines, agreed upon with Krzysztof and Peng, enabled a coherent implementation of both the DT bindings and the code. The following versions, from v4 to v8, were necessary to review and refine those implementations, bringing us to January 2025. At that point, Peng opened a separate pull request - https://github.com/devicetree-org/dt-schema/pull/154 - for the definition of general-purpose DT bindings for spread spectrum handling, which ended up invalidating mine. While waiting for his pull request to be accepted, I submitted version 9, trying to at least get the patches for the anatop driver merged, eventually reaching version 12. This final version was merged, but then a few days ago it was dropped. As it stands now: - We still don’t have proper spread spectrum handling - Peng’s pull request has been stalled since February 20 - We don’t have a driver for anatop - The CCM design remains flawed - Not even the first 4 patches of the series were merged — these were simply a replication for i.MX8MM and i.MX8MP of patch bedcf9d1dcf88 ("clk: imx: rename video\_pll1 to video\_pll"), which was already merged some time ago. Could you please let me know if you're still interested in this series? If so, could you suggest how to resolve the issues that led you to drop it? Thanks and regards, Dario > > [01/19] dt-bindings: clock: imx8mm: add VIDEO_PLL clocks > commit: 20e5d201b5d8f830e702d7d183f6b1b246b78d8a > [02/19] clk: imx8mm: rename video_pll1 to video_pll > commit: 26a33196b5b68cf199b6c4283a254aa92d2aaf4b > [03/19] dt-bindings: clock: imx8mp: add VIDEO_PLL clocks > commit: 2d50415e2457c6f6621c2faa3b01b11150fb9c67 > [04/19] clk: imx8mp: rename video_pll1 to video_pll > commit: 21bb969f608cefd8d847cf6eb50a193d9f1fbb87 > [05/19] dt-bindings: clock: imx8m-anatop: add oscillators and PLLs > commit: 2ba124053687c933031a6dc5b2e16ceaca250934 > [10/19] clk: imx: add hw API imx_anatop_get_clk_hw > commit: 17e3c1a272d97e49b4f3fbfe1f1b889e120d2be8 > [11/19] clk: imx: add support for i.MX8MM anatop clock driver > commit: 3cbc38cf42ca42d2dc9a93c949e0381ff919df71 > [12/19] clk: imx: add support for i.MX8MN anatop clock driver > commit: 80badb1d7264e83b512475898e7459f464a009c9 > [13/19] clk: imx: add support for i.MX8MP anatop clock driver > commit: 4c82bbe8b5437c7f16b2891ce33210c0f1410597 > [14/19] clk: imx8mp: rename ccm_base to base > commit: 1a77907dbbecfbe5e6a1aec28afd49a1dc184b7a > [16/19] dt-bindings: clock: imx8m-clock: add PLLs > commit: 6a55647af3334f1d935ece67de4a838a864b53fc > > Best regards, > -- > Abel Vesa <abel.vesa@linaro.org>
On 23/05/2025 17:19, Dario Binacchi wrote: > Hello Abel, > > On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: >> >> >> On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: >>> This version keeps the version v9 patches that can be merged and >>> removes the patches that will need to be modified in case Peng's >>> PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. >>> The idea is to speed up the merging of the patches in the series >>> that have already been reviewed and are not dependent on the >>> introduction of the assigned-clocks-sscs property, and postpone >>> the patches for spread spectrum to a future series once it becomes >>> clear what needs to be done. >>> >>> [...] >> >> Applied, thanks! > > I was surprised to see that the series has been removed from linux-next. Did you miss entire email thread explaining why? I think you never answered to several emails in this thread... and we - including myself - sent them a lot. > > It’s been 8 months since the first version dated September 28, 2024. > The most critical phase was version 3 - > https://lore.kernel.org/all/20241106090549.3684963-1-dario.binacchi@amarulasolutions.com/ > - > where two key issues emerged: > > 1 The CCM design is flawed because "in the current design, CCM is > taken as the producer of CLK_IMX8M_VIDEO_PLL, not the consumer." > > 2 A driver for anatop needs to be implemented because "using clocks > to replace fsl,ssc-clocks is possible under CCM mode, but you need > to develop the fsl,imx8mm-anatop clock driver." > > These development guidelines, agreed upon with Krzysztof and Peng, > enabled a coherent implementation of both the DT bindings and the > code. The following versions, from v4 to v8, were necessary to > review and refine those implementations, bringing us to January 2025. > > At that point, Peng opened a separate pull request - > https://github.com/devicetree-org/dt-schema/pull/154 - > for the definition of general-purpose DT bindings for spread spectrum > handling, which ended up invalidating mine. > > While waiting for his pull request to be accepted, I submitted version 9, > trying to at least get the patches for the anatop driver merged, > eventually reaching version 12. > > This final version was merged, but then a few days ago it was dropped. And explained why. There were bug reports which you completely ignored. > > As it stands now: > > - We still don’t have proper spread spectrum handling > - Peng’s pull request has been stalled since February 20 > - We don’t have a driver for anatop > - The CCM design remains flawed > - Not even the first 4 patches of the series were merged — these were > simply a replication for i.MX8MM and i.MX8MP of patch > bedcf9d1dcf88 ("clk: imx: rename video\_pll1 to video\_pll"), which > was already merged some time ago. > > Could you please let me know if you're still interested in this series? > If so, could you suggest how to resolve the issues that led you to drop it? You got several replies what is wrong. Can you respond to these instead of coming now surprised? Best regards, Krzysztof To unsubscribe from this group and stop receiving emails from it, send an email to linux-amarula+unsubscribe@amarulasolutions.com.
On Tue, May 27, 2025 at 8:42 PM Krzysztof Kozlowski <krzk@kernel.org> wrote: > > On 23/05/2025 17:19, Dario Binacchi wrote: > > Hello Abel, > > > > On Mon, May 5, 2025 at 9:52 AM Abel Vesa <abel.vesa@linaro.org> wrote: > >> > >> > >> On Thu, 24 Apr 2025 08:21:30 +0200, Dario Binacchi wrote: > >>> This version keeps the version v9 patches that can be merged and > >>> removes the patches that will need to be modified in case Peng's > >>> PR https://github.com/devicetree-org/dt-schema/pull/154 is accepted. > >>> The idea is to speed up the merging of the patches in the series > >>> that have already been reviewed and are not dependent on the > >>> introduction of the assigned-clocks-sscs property, and postpone > >>> the patches for spread spectrum to a future series once it becomes > >>> clear what needs to be done. > >>> > >>> [...] > >> > >> Applied, thanks! > > > > I was surprised to see that the series has been removed from linux-next. > > Did you miss entire email thread explaining why? I think you never > answered to several emails in this thread... and we - including myself - > sent them a lot. I don't think so: The answers related to the errors encountered during DTC compilation: https://lore.kernel.org/oe-kbuild-all/CABGWkvp7n=Or-OqnLoOJsQQCHF+=8eQ9EV5=O+Qp4sQF49_DbA@mail.gmail.com/ https://lore.kernel.org/all/CABGWkvqfyH=dcuw6EDZaFVVebj8SZhJF0P944+mmzL5YK3-Pug@mail.gmail.com/ The patch to fix the regression reported by Mark Brown: https://lore.kernel.org/all/20250516134945.14692-1-dario.binacchi@amarulasolutions.com/ Also successfully tested by him. And when I didn’t reply, I was expecting the maintainers to handle it — as Peng Fan had requested in version 10: https://lore.kernel.org/all/20250314093503.GD12210@nxa18884-linux/ I may have made some mistakes, but don't tell me I never replied. After all, anyone can make mistakes: https://lore.kernel.org/all/d421a6e8-e72c-48d9-8806-09724723b5d8@kernel.org/ Best regards, Dario > > > > > It’s been 8 months since the first version dated September 28, 2024. > > The most critical phase was version 3 - > > https://lore.kernel.org/all/20241106090549.3684963-1-dario.binacchi@amarulasolutions.com/ > > - > > where two key issues emerged: > > > > 1 The CCM design is flawed because "in the current design, CCM is > > taken as the producer of CLK_IMX8M_VIDEO_PLL, not the consumer." > > > > 2 A driver for anatop needs to be implemented because "using clocks > > to replace fsl,ssc-clocks is possible under CCM mode, but you need > > to develop the fsl,imx8mm-anatop clock driver." > > > > These development guidelines, agreed upon with Krzysztof and Peng, > > enabled a coherent implementation of both the DT bindings and the > > code. The following versions, from v4 to v8, were necessary to > > review and refine those implementations, bringing us to January 2025. > > > > At that point, Peng opened a separate pull request - > > https://github.com/devicetree-org/dt-schema/pull/154 - > > for the definition of general-purpose DT bindings for spread spectrum > > handling, which ended up invalidating mine. > > > > While waiting for his pull request to be accepted, I submitted version 9, > > trying to at least get the patches for the anatop driver merged, > > eventually reaching version 12. > > > > This final version was merged, but then a few days ago it was dropped. > > And explained why. There were bug reports which you completely ignored. > > > > > As it stands now: > > > > - We still don’t have proper spread spectrum handling > > - Peng’s pull request has been stalled since February 20 > > - We don’t have a driver for anatop > > - The CCM design remains flawed > > - Not even the first 4 patches of the series were merged — these were > > simply a replication for i.MX8MM and i.MX8MP of patch > > bedcf9d1dcf88 ("clk: imx: rename video\_pll1 to video\_pll"), which > > was already merged some time ago. > > > > Could you please let me know if you're still interested in this series? > > If so, could you suggest how to resolve the issues that led you to drop it? > > > You got several replies what is wrong. Can you respond to these instead > of coming now surprised? > > Best regards, > Krzysztof