Message ID | 20210404164914.712946-1-bjorn.andersson@linaro.org |
---|---|
State | New |
Headers | show |
Series | [GIT,PULL] Qualcomm ARM64 DT updates for 5.13 | expand |
From: Arnd Bergmann <arnd@arndb.de> On Sun, 4 Apr 2021 11:49:14 -0500, Bjorn Andersson wrote: > The following changes since commit a38fd8748464831584a19438cbb3082b5a2dab15: > > Linux 5.12-rc2 (2021-03-05 17:33:41 -0800) > > are available in the Git repository at: > > https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux.git tags/qcom-arm64-for-5.13 > > [...] I noticed a couple of new warnings from 'make dtbs_check W=1': qcom/sc7180.dtsi:1204.21-1220.6: Warning (avoid_unnecessary_addr_size): /soc@0/geniqup@ac0000/i2c@a8c000: unnecessary #address-cells/#size-cells without "ranges" or child "reg" property qcom/sc7180.dtsi:965.21-981.6: Warning (avoid_unnecessary_addr_size): /soc@0/geniqup@8c0000/i2c@890000: unnecessary #address-cells/#size-cells without "ranges" or child "reg" property qcom/sdm845.dtsi:3912.23-4045.5: Warning (simple_bus_reg): /soc@0/camss@a00000: simple-bus unit address format error, expected "acb3000" qcom/sdm845.dtsi:4041.10-4044.6: Warning (graph_child_address): /soc@0/camss@a00000/ports: graph node has single child node 'port@0', #address-cells/#size-cells are not necessary qcom/sdm845.dtsi:4102.32-4129.5: Warning (simple_bus_reg): /soc@0/dsi-opp-table: missing or empty reg/ranges property qcom/msm8916-samsung-a5u-eur.dt.yaml: spmi@200f000: reg: [[33615872, 4096], [37748736, 4194304], [46137344, 4194304], [58720256, 2097152], [33595392, 8448]] is too long qcom/sc7180-trogdor-lazor-r0.dt.yaml: gmu@506a000: compatible:0: 'qcom,adreno-gmu-618.0' is not one of ['qcom,adreno-gmu-630.2'] qcom/sc7180-trogdor-lazor-r1-kb.dt.yaml: memory@80900000: 'device_type' is a required property qcom/sdm850-lenovo-yoga-c630.dt.yaml: memory@97b00000: 'device_type' is a required property Maybe see if you can address these in a follow-up, to avoid regressions. Merged into arm/dt, thanks! merge commit: 61bac46eed14ad73cd41c5b7075350fef78ca4be Arnd
On Thu 08 Apr 10:24 CDT 2021, Arnd Bergmann wrote: > From: Arnd Bergmann <arnd@arndb.de> > > On Sun, 4 Apr 2021 11:49:14 -0500, Bjorn Andersson wrote: > > The following changes since commit a38fd8748464831584a19438cbb3082b5a2dab15: > > > > Linux 5.12-rc2 (2021-03-05 17:33:41 -0800) > > > > are available in the Git repository at: > > > > https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux.git tags/qcom-arm64-for-5.13 > > > > [...] > > I noticed a couple of new warnings from 'make dtbs_check W=1': > > qcom/sc7180.dtsi:1204.21-1220.6: Warning (avoid_unnecessary_addr_size): /soc@0/geniqup@ac0000/i2c@a8c000: unnecessary #address-cells/#size-cells without "ranges" or child "reg" property > qcom/sc7180.dtsi:965.21-981.6: Warning (avoid_unnecessary_addr_size): /soc@0/geniqup@8c0000/i2c@890000: unnecessary #address-cells/#size-cells without "ranges" or child "reg" property qcom/sdm845.dtsi:3912.23-4045.5: Warning (simple_bus_reg): /soc@0/camss@a00000: simple-bus unit address format error, expected "acb3000" > qcom/sdm845.dtsi:4041.10-4044.6: Warning (graph_child_address): /soc@0/camss@a00000/ports: graph node has single child node 'port@0', #address-cells/#size-cells are not necessary > qcom/sdm845.dtsi:4102.32-4129.5: Warning (simple_bus_reg): /soc@0/dsi-opp-table: missing or empty reg/ranges property > qcom/msm8916-samsung-a5u-eur.dt.yaml: spmi@200f000: reg: [[33615872, 4096], [37748736, 4194304], [46137344, 4194304], [58720256, 2097152], [33595392, 8448]] is too long > qcom/sc7180-trogdor-lazor-r0.dt.yaml: gmu@506a000: compatible:0: 'qcom,adreno-gmu-618.0' is not one of ['qcom,adreno-gmu-630.2'] > qcom/sc7180-trogdor-lazor-r1-kb.dt.yaml: memory@80900000: 'device_type' is a required property > qcom/sdm850-lenovo-yoga-c630.dt.yaml: memory@97b00000: 'device_type' is a required property > I saw your other email regarding these and per Rob's request we've ramped up our effort on cleaning these out. I've however still not figured out a good way to incorporate the dtbs_check in my patch handling workflow. > Maybe see if you can address these in a follow-up, to avoid regressions. > I'm about to send you a another set of pull requests with some more goodies that was lingering on the list. I will take a look to see if I can follow up on that with some fixes for above warnings - and take another look at incorporating dtbs_check in my workflow. > Merged into arm/dt, thanks! > Thank you! Bjorn > merge commit: 61bac46eed14ad73cd41c5b7075350fef78ca4be > > Arnd
On Thu, Apr 8, 2021 at 6:06 PM Bjorn Andersson <bjorn.andersson@linaro.org> wrote: > On Thu 08 Apr 10:24 CDT 2021, Arnd Bergmann wrote: > > Maybe see if you can address these in a follow-up, to avoid regressions. > > > > I'm about to send you a another set of pull requests with some more > goodies that was lingering on the list. I will take a look to see if I > can follow up on that with some fixes for above warnings - and take > another look at incorporating dtbs_check in my workflow. Ok, sounds good. As I mentioned, I'm also just learning how to use dtbs_check properly. Unfortunately, it takes ages to run on a non-parallel build, and using 'make -j32' as I normally do means the output gets reordered with every run. It's probably ok if you figure out how to only run it for the files you actually change in the Qualcomm tree. Arnd
On Thu, 8 Apr 2021 at 23:05, Arnd Bergmann <arnd@kernel.org> wrote: > > On Thu, Apr 8, 2021 at 6:06 PM Bjorn Andersson > <bjorn.andersson@linaro.org> wrote: > > On Thu 08 Apr 10:24 CDT 2021, Arnd Bergmann wrote: > > > Maybe see if you can address these in a follow-up, to avoid regressions. > > > > > > > I'm about to send you a another set of pull requests with some more > > goodies that was lingering on the list. I will take a look to see if I > > can follow up on that with some fixes for above warnings - and take > > another look at incorporating dtbs_check in my workflow. > > Ok, sounds good. As I mentioned, I'm also just learning how to use > dtbs_check properly. Unfortunately, it takes ages to run on a > non-parallel build, > and using 'make -j32' as I normally do means the output gets reordered > with every run. > > It's probably ok if you figure out how to only run it for the files you > actually change in the Qualcomm tree. Disabling all non-Qualcomm arches makes `make dtbs_check` a little bit less painful.
On Thu 08 Apr 16:53 CDT 2021, Dmitry Baryshkov wrote: > On Thu, 8 Apr 2021 at 23:05, Arnd Bergmann <arnd@kernel.org> wrote: > > > > On Thu, Apr 8, 2021 at 6:06 PM Bjorn Andersson > > <bjorn.andersson@linaro.org> wrote: > > > On Thu 08 Apr 10:24 CDT 2021, Arnd Bergmann wrote: > > > > Maybe see if you can address these in a follow-up, to avoid regressions. > > > > > > > > > > I'm about to send you a another set of pull requests with some more > > > goodies that was lingering on the list. I will take a look to see if I > > > can follow up on that with some fixes for above warnings - and take > > > another look at incorporating dtbs_check in my workflow. > > > > Ok, sounds good. As I mentioned, I'm also just learning how to use > > dtbs_check properly. Unfortunately, it takes ages to run on a > > non-parallel build, > > and using 'make -j32' as I normally do means the output gets reordered > > with every run. > > > > It's probably ok if you figure out how to only run it for the files you > > actually change in the Qualcomm tree. > > Disabling all non-Qualcomm arches makes `make dtbs_check` a little bit > less painful. > Yes, this is a trick I'm applying as well. It would however be nice if one didn't have to disable CONFIG_ARCH_* to achieve this, or even better if one could run dtbs_check on a single .dtb. In particular since many of the warnings are resolved by fixing the bindings, which causes dtbs_check to recheck all files. Regards, Bjorn
On Fri, 9 Apr 2021 at 02:05, Bjorn Andersson <bjorn.andersson@linaro.org> wrote: > > On Thu 08 Apr 16:53 CDT 2021, Dmitry Baryshkov wrote: > > > On Thu, 8 Apr 2021 at 23:05, Arnd Bergmann <arnd@kernel.org> wrote: > > > > > > On Thu, Apr 8, 2021 at 6:06 PM Bjorn Andersson > > > <bjorn.andersson@linaro.org> wrote: > > > > On Thu 08 Apr 10:24 CDT 2021, Arnd Bergmann wrote: > > > > > Maybe see if you can address these in a follow-up, to avoid regressions. > > > > > > > > > > > > > I'm about to send you a another set of pull requests with some more > > > > goodies that was lingering on the list. I will take a look to see if I > > > > can follow up on that with some fixes for above warnings - and take > > > > another look at incorporating dtbs_check in my workflow. > > > > > > Ok, sounds good. As I mentioned, I'm also just learning how to use > > > dtbs_check properly. Unfortunately, it takes ages to run on a > > > non-parallel build, > > > and using 'make -j32' as I normally do means the output gets reordered > > > with every run. > > > > > > It's probably ok if you figure out how to only run it for the files you > > > actually change in the Qualcomm tree. > > > > Disabling all non-Qualcomm arches makes `make dtbs_check` a little bit > > less painful. > > > > Yes, this is a trick I'm applying as well. > > It would however be nice if one didn't have to disable CONFIG_ARCH_* to > achieve this, or even better if one could run dtbs_check on a single > .dtb. In particular since many of the warnings are resolved by fixing > the bindings, which causes dtbs_check to recheck all files. As a quick hack: add the following rule to the top-level Makefile: %.dt.yaml: include/config/kernel.release scripts_dtc $(Q)$(MAKE) $(build)=Documentation/devicetree/bindings Documentation/devicetree/bindings/processed-schema.json $(Q)$(MAKE) $(build)=$(dtstree) $(dtstree)/$@ CHECK_DTBS=y Now you can run checks for a single device tree by running 'make ARCH=arm64 ...... sdm845-db845.dy.yaml' I don't know if it's worth including this into the upstream source tree.