Some buses need additional manual adjustments atop the usual
calculations. Fill in the missing coefficients.
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
---
drivers/interconnect/qcom/qcm2290.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/interconnect/qcom/qcm2290.c b/drivers/interconnect/qcom/qcm2290.c
index 3c3b24264a5b..457e5713ae43 100644
--- a/drivers/interconnect/qcom/qcm2290.c
+++ b/drivers/interconnect/qcom/qcm2290.c
@@ -1198,6 +1198,7 @@ static const struct qcom_icc_desc qcm2290_bimc = {
.regmap_cfg = &qcm2290_bimc_regmap_config,
/* M_REG_BASE() in vendor msm_bus_bimc_adhoc driver */
.qos_offset = 0x8000,
+ .ab_percent = 153,
};
static struct qcom_icc_node * const qcm2290_cnoc_nodes[] = {
@@ -1324,6 +1325,7 @@ static const struct qcom_icc_desc qcm2290_mmnrt_virt = {
.bus_clk_desc = &mmaxi_0_clk,
.regmap_cfg = &qcm2290_snoc_regmap_config,
.qos_offset = 0x15000,
+ .ab_percent = 142,
};
static struct qcom_icc_node * const qcm2290_mmrt_virt_nodes[] = {
@@ -1339,6 +1341,7 @@ static const struct qcom_icc_desc qcm2290_mmrt_virt = {
.bus_clk_desc = &mmaxi_1_clk,
.regmap_cfg = &qcm2290_snoc_regmap_config,
.qos_offset = 0x15000,
+ .ab_percent = 139,
};
static const struct of_device_id qcm2290_noc_of_match[] = {
--
2.41.0
On Wed, Jul 26, 2023 at 06:25:44PM +0200, Konrad Dybcio wrote: > Some buses need additional manual adjustments atop the usual > calculations. Fill in the missing coefficients. > > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> What about the funny util-fact/vrail-comp on the mas-apps-proc node downstream? Thanks, Stephan > --- > drivers/interconnect/qcom/qcm2290.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/interconnect/qcom/qcm2290.c b/drivers/interconnect/qcom/qcm2290.c > index 3c3b24264a5b..457e5713ae43 100644 > --- a/drivers/interconnect/qcom/qcm2290.c > +++ b/drivers/interconnect/qcom/qcm2290.c > @@ -1198,6 +1198,7 @@ static const struct qcom_icc_desc qcm2290_bimc = { > .regmap_cfg = &qcm2290_bimc_regmap_config, > /* M_REG_BASE() in vendor msm_bus_bimc_adhoc driver */ > .qos_offset = 0x8000, > + .ab_percent = 153, > }; > > static struct qcom_icc_node * const qcm2290_cnoc_nodes[] = { > @@ -1324,6 +1325,7 @@ static const struct qcom_icc_desc qcm2290_mmnrt_virt = { > .bus_clk_desc = &mmaxi_0_clk, > .regmap_cfg = &qcm2290_snoc_regmap_config, > .qos_offset = 0x15000, > + .ab_percent = 142, > }; > > static struct qcom_icc_node * const qcm2290_mmrt_virt_nodes[] = { > @@ -1339,6 +1341,7 @@ static const struct qcom_icc_desc qcm2290_mmrt_virt = { > .bus_clk_desc = &mmaxi_1_clk, > .regmap_cfg = &qcm2290_snoc_regmap_config, > .qos_offset = 0x15000, > + .ab_percent = 139, > }; > > static const struct of_device_id qcm2290_noc_of_match[] = { > > -- > 2.41.0 >
On 26.07.2023 19:18, Stephan Gerhold wrote: > On Wed, Jul 26, 2023 at 06:25:44PM +0200, Konrad Dybcio wrote: >> Some buses need additional manual adjustments atop the usual >> calculations. Fill in the missing coefficients. >> >> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > > What about the funny util-fact/vrail-comp on the mas-apps-proc node > downstream? Can't see it neither on msm-5.4 (with the icc API) nor in the 4.19 (msmbus) device tree. Konrad
On Wed, Jul 26, 2023 at 07:20:27PM +0200, Konrad Dybcio wrote: > On 26.07.2023 19:18, Stephan Gerhold wrote: > > On Wed, Jul 26, 2023 at 06:25:44PM +0200, Konrad Dybcio wrote: > >> Some buses need additional manual adjustments atop the usual > >> calculations. Fill in the missing coefficients. > >> > >> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > > > > What about the funny util-fact/vrail-comp on the mas-apps-proc node > > downstream? > Can't see it neither on msm-5.4 (with the icc API) nor in the 4.19 (msmbus) > device tree. > Not sure where to get up-to-date device trees nowadays. The AOSP repo I was looking at has this commit where it was added: https://android.googlesource.com/kernel/msm-extra/devicetree/+/02f8c342b23c20a5cf967df649814be37a08227c%5E%21/#F0 Stephan
On 26.07.2023 19:26, Stephan Gerhold wrote: > On Wed, Jul 26, 2023 at 07:20:27PM +0200, Konrad Dybcio wrote: >> On 26.07.2023 19:18, Stephan Gerhold wrote: >>> On Wed, Jul 26, 2023 at 06:25:44PM +0200, Konrad Dybcio wrote: >>>> Some buses need additional manual adjustments atop the usual >>>> calculations. Fill in the missing coefficients. >>>> >>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> >>> >>> What about the funny util-fact/vrail-comp on the mas-apps-proc node >>> downstream? >> Can't see it neither on msm-5.4 (with the icc API) nor in the 4.19 (msmbus) >> device tree. >> > > Not sure where to get up-to-date device trees nowadays. The AOSP repo > I was looking at has this commit where it was added: > https://android.googlesource.com/kernel/msm-extra/devicetree/+/02f8c342b23c20a5cf967df649814be37a08227c%5E%21/#F0 Oh right, take a look at this one: https://git.codelinaro.org/clo/la/kernel/msm-4.14/-/commit/201df022706e100cef8d28983c6a7b883fcaec5a I guess I'll need to update the icc driver then. Konrad
On 26.07.2023 19:46, Konrad Dybcio wrote: > On 26.07.2023 19:26, Stephan Gerhold wrote: >> On Wed, Jul 26, 2023 at 07:20:27PM +0200, Konrad Dybcio wrote: >>> On 26.07.2023 19:18, Stephan Gerhold wrote: >>>> On Wed, Jul 26, 2023 at 06:25:44PM +0200, Konrad Dybcio wrote: >>>>> Some buses need additional manual adjustments atop the usual >>>>> calculations. Fill in the missing coefficients. >>>>> >>>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> >>>> >>>> What about the funny util-fact/vrail-comp on the mas-apps-proc node >>>> downstream? >>> Can't see it neither on msm-5.4 (with the icc API) nor in the 4.19 (msmbus) >>> device tree. >>> >> >> Not sure where to get up-to-date device trees nowadays. The AOSP repo >> I was looking at has this commit where it was added: >> https://android.googlesource.com/kernel/msm-extra/devicetree/+/02f8c342b23c20a5cf967df649814be37a08227c%5E%21/#F0 > Oh right, take a look at this one: > > https://git.codelinaro.org/clo/la/kernel/msm-4.14/-/commit/201df022706e100cef8d28983c6a7b883fcaec5a > > I guess I'll need to update the icc driver then. Moreover, this having vrail-comp = 96, means I'll have to go with the wrecked downstream way of *(100/(percent)) instead of *(percent/100)... I also noticed that sm6125 makes very heavy use of per-node clocks.. Konrad
© 2016 - 2025 Red Hat, Inc.