lwn.git/drivers/interconnect/qcom/osm-l3.c, branch docs-6.5

interconnect: qcom: osm-l3: drop unuserd header inclusion

2023-03-20T14:42:27+00:00

The commit 4529992c9474 ("interconnect: qcom: osm-l3: Use platform-independent node ids") made osm-l3 driver use platform-independent IDs, removing the need to include platform headers. Fixes: 4529992c9474 ("interconnect: qcom: osm-l3: Use platform-independent node ids") Signed-off-by: Dmitry Baryshkov Reviewed-by: Bjorn Andersson Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230103031159.1060075-1-dmitry.baryshkov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: fix registration race

2023-03-07T20:19:05+00:00

The current interconnect provider registration interface is inherently racy as nodes are not added until the after adding the provider. This can specifically cause racing DT lookups to fail: of_icc_xlate_onecell: invalid index 0 cpu cpu0: error -EINVAL: error finding src node cpu cpu0: dev_pm_opp_of_find_icc_paths: Unable to get path0: -22 qcom-cpufreq-hw: probe of 18591000.cpufreq failed with error -22 Switch to using the new API where the provider is not registered until after it has been fully initialised. Fixes: 5bc9900addaf ("interconnect: qcom: Add OSM L3 interconnect provider support") Cc: stable@vger.kernel.org # 5.7 Reviewed-by: Konrad Dybcio Signed-off-by: Johan Hovold Link: https://lore.kernel.org/r/20230306075651.2449-6-johan+linaro@kernel.org Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: fix icc_onecell_data allocation

2023-03-06T14:48:23+00:00

This is a struct with a trailing zero-length array of icc_node pointers but it's allocated as if it were a single array of icc_nodes instead. Fortunately this overallocates memory rather then allocating less memory than required. Fix by replacing devm_kcalloc() with devm_kzalloc() and struct_size() macro. Fixes: 5bc9900addaf ("interconnect: qcom: Add OSM L3 interconnect provider support") Signed-off-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20230105002221.1416479-2-dmitry.baryshkov@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: Simplify osm_l3_set()

2022-11-14T13:05:30+00:00

The aggregation over votes for all nodes in the provider will always only find the bandwidth votes for the destination side of the path. Further more, the average kBps value will always be 0. Simplify the logic by directly looking at the destination node's peak bandwidth request. Signed-off-by: Bjorn Andersson Tested-by: Steev Klimaszewski Reviewed-by: Sibi Sankar Link: https://lore.kernel.org/r/20221111032515.3460-5-quic_bjorande@quicinc.com Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: Add per-core EPSS L3 support

2022-11-14T13:05:30+00:00

The EPSS instance in e.g. SM8350 and SC8280XP has per-core L3 voting enabled. In this configuration, the "shared" vote is done using the REG_L3_VOTE register instead of PERF_STATE. Rename epss_l3 to clarify that it's affecting the PERF_STATE register and add a new L3_VOTE description. Given platform lineage it's assumed that the L3_VOTE-based case will be the predominant one, so use this for a new generic qcom,epss-l3 compatible. While adding the EPSS generic, also add qcom,osm-l3. Signed-off-by: Bjorn Andersson Tested-by: Steev Klimaszewski Reviewed-by: Sibi Sankar Link: https://lore.kernel.org/r/20221111032515.3460-4-quic_bjorande@quicinc.com Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: Squash common descriptors

2022-11-14T13:05:30+00:00

Each platform defines their own OSM L3 descriptor, but in practice there's only two: one for OSM and one for EPSS. Remove the duplicated definitions. Signed-off-by: Bjorn Andersson Tested-by: Steev Klimaszewski Reviewed-by: Sibi Sankar Link: https://lore.kernel.org/r/20221111032515.3460-3-quic_bjorande@quicinc.com Signed-off-by: Georgi Djakov

interconnect: qcom: osm-l3: Use platform-independent node ids

2022-11-14T13:05:30+00:00

The identifiers used for nodes needs to be unique in the running system, but defining them per platform results in a lot of duplicated definitions and prevents us from using generic compatibles. As these identifiers are not exposed outside the kernel, change to use driver-local numbers, picked completely at random. Signed-off-by: Bjorn Andersson Tested-by: Steev Klimaszewski Reviewed-by: Sibi Sankar Link: https://lore.kernel.org/r/20221111032515.3460-2-quic_bjorande@quicinc.com Signed-off-by: Georgi Djakov

interconnect: osm-l3: Ignore return value of icc_provider_del() in .remove()

2022-08-16T13:38:24+00:00

icc_provider_del() already emits an error message on failure. In this case letting .remove() return the corresponding error code results in another error message and the device is removed anyhow. (See platform_remove().) So ignore the return value of icc_provider_del() and return 0 unconditionally. This is a preparation for making platform remove callbacks return void. Signed-off-by: Uwe Kleine-König Link: https://lore.kernel.org/r/20220718121409.171773-6-u.kleine-koenig@pengutronix.de Signed-off-by: Georgi Djakov

interconnect: qcom: constify icc_node pointers

2022-04-23T12:11:48+00:00

Pointers to struct qcom_icc_node (and similar structures) are not modified, so they can be made const for safety. The contents of struct qcom_icc_node must stay non-const. Signed-off-by: Krzysztof Kozlowski Link: https://lore.kernel.org/r/20220412102623.227607-2-krzysztof.kozlowski@linaro.org Signed-off-by: Georgi Djakov

interconnect: qcom: Add EPSS L3 support on SC7280

2021-11-22T14:57:22+00:00

Add Epoch Subsystem (EPSS) L3 interconnect provider support on SC7280 SoCs. Signed-off-by: Odelu Kukatla Reviewed-by: Stephen Boyd Reviewed-by: Bjorn Andersson Link: https://lore.kernel.org/r/1634812857-10676-3-git-send-email-okukatla@codeaurora.org Signed-off-by: Georgi Djakov