| Message ID | 20230413131705.3073911-1-brgl@bgdev.pl |
|---|---|
| Headers | show |
| Series | arm64: qcom: fix the reboot reason handling on sa8775p | expand |
On 13/04/2023 15:17, Bartosz Golaszewski wrote: > From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> > > This module is used by the Qualcomm sa8775p platform for passing the > reboot reason to the bootloader. Enable building it in the arm64 > defconfig as a module. > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Best regards, Krzysztof
On Sun, Apr 16, 2023, at 17:16, Krzysztof Kozlowski wrote: > On 13/04/2023 15:17, Bartosz Golaszewski wrote: >> From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> >> >> This module is used by the Qualcomm sa8775p platform for passing the >> reboot reason to the bootloader. Enable building it in the arm64 >> defconfig as a module. >> >> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> > > Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Applied, thanks! Arnd
On Thu, Apr 13, 2023, at 15:17, Bartosz Golaszewski wrote: > From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> > > SA8775P uses nvmem to pass the reboot reason magic value to the bootloader. > Remove the reboot modes from the PON node and introduce an SDAM node passed > to the nvmem-reboot-mode driver. While at it: convert the bindings for > nvmem-reboot-mode to YAML and enable it for arm64 in defconfig. > > Bartosz Golaszewski (2): > arm64: defconfig: enable building the nvmem-reboot-mode module > dt-bindings: power: reset: convert nvmem-reboot-mode bindings to YAML > > Parikshit Pareek (2): > arm64: dts: qcom: sa8775p: pmic: remove the PON modes > arm64: dts: qcom: sa8775p: pmic: add the sdam_0 node It looks like it's too late for the dts patches in v6.4, but I was picking up defconfig patches from the list, and that one seems useful regardless of the rest, so I added it to the soc/defconfig branch. Arnd
On 4/13/2023 9:42 PM, Krzysztof Kozlowski wrote: > On 13/04/2023 15:17, Bartosz Golaszewski wrote: >> From: Parikshit Pareek <quic_ppareek@quicinc.com> >> >> Remove the power on reasons with reboot from the pmm8654au_0_pon. >> Instead, the PoN reaons should be part of different sdam_0 mode, to > > typo: reasons > >> be interoduced. > > introduced > > Anyway it does not say why. Are these power reasons not correct? > Hi Krzysztof, Since sm8350 the PMIC PON peripheral was split into PON_HLOS and PON_PBS to avoid security concerns with HLOS APPS being able to trigger a PMIC WARM_RESET unilaterally. When the split occurred, the spare registers ended up in PON_PBS, not PON_HLOS. Thus at that time, we moved to using an SDAM register for Linux “reboot reason” configuration. And bootloader also SDAM register to get these reboot region data to get into bootloader/edl, so to have this working we need to use SDAM. >> >> Signed-off-by: Parikshit Pareek <quic_ppareek@quicinc.com> >> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> >> --- > > Best regards, > Krzysztof > -Shazad
On 18/04/2023 06:39, Shazad Hussain wrote: > > > On 4/13/2023 9:42 PM, Krzysztof Kozlowski wrote: >> On 13/04/2023 15:17, Bartosz Golaszewski wrote: >>> From: Parikshit Pareek <quic_ppareek@quicinc.com> >>> >>> Remove the power on reasons with reboot from the pmm8654au_0_pon. >>> Instead, the PoN reaons should be part of different sdam_0 mode, to >> >> typo: reasons >> >>> be interoduced. >> >> introduced >> >> Anyway it does not say why. Are these power reasons not correct? >> > > Hi Krzysztof, > Since sm8350 the PMIC PON peripheral was split into PON_HLOS and PON_PBS > to avoid security concerns with HLOS APPS being able to trigger a PMIC > WARM_RESET unilaterally. When the split occurred, the spare registers > ended up in PON_PBS, not PON_HLOS. Thus at that time, we moved to using > an SDAM register for Linux “reboot reason” configuration. And bootloader > also SDAM register to get these reboot region data to get into > bootloader/edl, so to have this working we need to use SDAM. The explanation of their correctness should be in commit msg. Best regards, Krzysztof
On 18.04.2023 06:39, Shazad Hussain wrote: > > > On 4/13/2023 9:42 PM, Krzysztof Kozlowski wrote: >> On 13/04/2023 15:17, Bartosz Golaszewski wrote: >>> From: Parikshit Pareek <quic_ppareek@quicinc.com> >>> >>> Remove the power on reasons with reboot from the pmm8654au_0_pon. >>> Instead, the PoN reaons should be part of different sdam_0 mode, to >> >> typo: reasons >> >>> be interoduced. >> >> introduced >> >> Anyway it does not say why. Are these power reasons not correct? >> > > Hi Krzysztof, > Since sm8350 the PMIC PON peripheral was split into PON_HLOS and PON_PBS > to avoid security concerns with HLOS APPS being able to trigger a PMIC > WARM_RESET unilaterally. When the split occurred, the spare registers > ended up in PON_PBS, not PON_HLOS. Thus at that time, we moved to using > an SDAM register for Linux “reboot reason” configuration. And bootloader > also SDAM register to get these reboot region data to get into > bootloader/edl, so to have this working we need to use SDAM. > Does that imply all PMICs following the PMK8350 scheme (separate HLOS and PBS) should direct reboot mode writes to SDAM? Konrad >>> >>> Signed-off-by: Parikshit Pareek <quic_ppareek@quicinc.com> >>> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> >>> --- >> >> Best regards, >> Krzysztof >> > > -Shazad
On 4/18/2023 3:13 PM, Konrad Dybcio wrote: > > > On 18.04.2023 06:39, Shazad Hussain wrote: >> >> >> On 4/13/2023 9:42 PM, Krzysztof Kozlowski wrote: >>> On 13/04/2023 15:17, Bartosz Golaszewski wrote: >>>> From: Parikshit Pareek <quic_ppareek@quicinc.com> >>>> >>>> Remove the power on reasons with reboot from the pmm8654au_0_pon. >>>> Instead, the PoN reaons should be part of different sdam_0 mode, to >>> >>> typo: reasons >>> >>>> be interoduced. >>> >>> introduced >>> >>> Anyway it does not say why. Are these power reasons not correct? >>> >> >> Hi Krzysztof, >> Since sm8350 the PMIC PON peripheral was split into PON_HLOS and PON_PBS >> to avoid security concerns with HLOS APPS being able to trigger a PMIC >> WARM_RESET unilaterally. When the split occurred, the spare registers >> ended up in PON_PBS, not PON_HLOS. Thus at that time, we moved to using >> an SDAM register for Linux “reboot reason” configuration. And bootloader >> also SDAM register to get these reboot region data to get into >> bootloader/edl, so to have this working we need to use SDAM. >> > Does that imply all PMICs following the PMK8350 scheme (separate HLOS and > PBS) should direct reboot mode writes to SDAM? > > Konrad Yes, that's what the expectation is with bootloader using SDAM as well. >>>> >>>> Signed-off-by: Parikshit Pareek <quic_ppareek@quicinc.com> >>>> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> >>>> --- >>> >>> Best regards, >>> Krzysztof >>> >> >> -Shazad -Shazad
From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> SA8775P uses nvmem to pass the reboot reason magic value to the bootloader. Remove the reboot modes from the PON node and introduce an SDAM node passed to the nvmem-reboot-mode driver. While at it: convert the bindings for nvmem-reboot-mode to YAML and enable it for arm64 in defconfig. Bartosz Golaszewski (2): arm64: defconfig: enable building the nvmem-reboot-mode module dt-bindings: power: reset: convert nvmem-reboot-mode bindings to YAML Parikshit Pareek (2): arm64: dts: qcom: sa8775p: pmic: remove the PON modes arm64: dts: qcom: sa8775p: pmic: add the sdam_0 node .../power/reset/nvmem-reboot-mode.txt | 26 ---------- .../power/reset/nvmem-reboot-mode.yaml | 52 +++++++++++++++++++ arch/arm64/boot/dts/qcom/sa8775p-pmics.dtsi | 23 +++++++- arch/arm64/configs/defconfig | 1 + 4 files changed, 74 insertions(+), 28 deletions(-) delete mode 100644 Documentation/devicetree/bindings/power/reset/nvmem-reboot-mode.txt create mode 100644 Documentation/devicetree/bindings/power/reset/nvmem-reboot-mode.yaml