linux/drivers/interconnect, branch v5.10

interconnect: fix memory trashing in of_count_icc_providers()

2020-11-20T14:01:35Z

of_count_icc_providers() function uses for_each_available_child_of_node() helper to recursively check all the available nodes. This helper already properly handles child nodes' reference count, so there is no need to do it explicitly. Remove the excessive call to of_node_put(). This fixes memory trashing when CONFIG_OF_DYNAMIC is enabled (for example arm/multi_v7_defconfig). Fixes: b1d681d8d324 ("interconnect: Add sync state support") Signed-off-by: Marek Szyprowski Link: https://lore.kernel.org/r/20201119103746.32564-1-m.szyprowski@samsung.com Signed-off-by: Georgi Djakov

interconnect: qcom: qcs404: Remove GPU and display RPM IDs

2020-11-20T13:52:05Z

The following errors are noticed during boot on a QCS404 board: [ 2.926647] qcom_icc_rpm_smd_send mas 6 error -6 [ 2.934573] qcom_icc_rpm_smd_send mas 8 error -6 These errors show when we try to configure the GPU and display nodes. Since these particular nodes aren't supported on RPM and are purely local, we should just change their mas_rpm_id to -1 to avoid any requests being sent for these master IDs. Reviewed-by: Mike Tipton Reviewed-by: Bjorn Andersson Link: https://lore.kernel.org/r/20201118111044.26056-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: msm8916: Remove rpm-ids from non-RPM nodes

2020-11-20T13:51:51Z

Some nodes are incorrectly marked as RPM-controlled (they have RPM master and slave ids assigned), but are actually controlled by the application CPU instead. The RPM complains when we send requests for resources that it can't control. Let's fix this by replacing the IDs, with the default "-1" in which case no requests are sent. Reviewed-by: Mike Tipton Link: https://lore.kernel.org/r/20201112105140.10092-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: msm8974: Don't boost the NoC rate during boot

2020-11-17T22:21:47Z

It has been reported that on Fairphone 2 (msm8974-based), increasing the clock rate for some of the NoCs during boot may lead to hangs. Let's restore the original behavior and not touch the clock rate of any of the NoCs to fix the regression. Reported-by: Luca Weiss Tested-by: Luca Weiss Fixes: b1d681d8d324 ("interconnect: Add sync state support") Link: https://lore.kernel.org/r/20201109124512.10776-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: msm8974: Prevent integer overflow in rate

2020-11-17T22:18:05Z

When sync_state support got introduced recently, by default we try to set the NoCs to run initially at maximum rate. But as these values are aggregated, we may end with a really big clock rate value, which is then converted from "u64" to "long" during the clock rate rounding. But on 32bit platforms this may result an overflow. Fix it by making sure that the rate is within range. Reported-by: Luca Weiss Reviewed-by: Brian Masney Link: https://lore.kernel.org/r/20201106144847.7726-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: use icc_sync state for sm8[12]50

2020-10-27T14:01:22Z

In addition to the rest of Qcom interconnect drivers use icc_sync_state for SM8150/SM8250 interconnect drivers to notify the interconnect framework when all consumers are probed and there is no need to keep the bandwidth set to maximum anymore. Also move the BCM initialization before creating the nodes to set the max bandwidth in hardware for the initialization/probing stage. Signed-off-by: Dmitry Baryshkov Fixes: 7d3b0b0d8184 ("interconnect: qcom: Use icc_sync_state") Link: https://lore.kernel.org/r/20201027133418.976687-1-dmitry.baryshkov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: Ensure that the floor bandwidth value is enforced

2020-10-22T10:26:26Z

Take into account the initial bandwidth from the framework and update the internal sum and max values before committing if needed. This will ensure that the floor bandwidth values are enforced until the providers get into sync state. Fixes: 7d3b0b0d8184 ("interconnect: qcom: Use icc_sync_state") Tested-by: Akash Asthana Link: https://lore.kernel.org/r/20201021155938.9223-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: sc7180: Init BCMs before creating the nodes

2020-10-15T06:24:00Z

Currently if we use sync_state, by default the bandwidth is maxed out, but in order to set this in hardware, the BCMs (Bus Clock Managers) need to be initialized first. Move the BCM initialization before creating the nodes to fix this. Fixes: 7d3b0b0d8184 ("interconnect: qcom: Use icc_sync_state") Acked-by: Saravana Kannan Link: https://lore.kernel.org/r/20201013135913.29059-3-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: sdm845: Init BCMs before creating the nodes

2020-10-15T06:24:00Z

Currently if we use sync_state, by default the bandwidth is maxed out, but in order to set this in hardware, the BCMs (Bus Clock Managers) need to be initialized first. Move the BCM initialization before creating the nodes to fix this. Fixes: 7d3b0b0d8184 ("interconnect: qcom: Use icc_sync_state") Acked-by: Saravana Kannan Link: https://lore.kernel.org/r/20201013135913.29059-2-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov

interconnect: Aggregate before setting initial bandwidth

2020-10-15T06:24:00Z

When setting the initial bandwidth, make sure to call the aggregate() function (if such is implemented for the current provider), to handle cases when data needs to be aggregated first. Fixes: b1d681d8d324 ("interconnect: Add sync state support") Acked-by: Saravana Kannan Link: https://lore.kernel.org/r/20201013135913.29059-1-georgi.djakov@linaro.org Signed-off-by: Georgi Djakov