kernel/linux.git/drivers/interconnect, branch v6.6.132

interconnect: debugfs: initialize src_node and dst_node to empty strings

2026-01-30T09:27:35+00:00

[ Upstream commit 8cc27f5c6dd17dd090f3a696683f04336c162ff5 ] The debugfs_create_str() API assumes that the string pointer is either NULL or points to valid kmalloc() memory. Leaving the pointer uninitialized can cause problems. Initialize src_node and dst_node to empty strings before creating the debugfs entries to guarantee that reads and writes are safe. Fixes: 770c69f037c1 ("interconnect: Add debugfs test client") Signed-off-by: Georgi Djakov Reviewed-by: Kuan-Wei Chiu Tested-by: Kuan-Wei Chiu Link: https://lore.kernel.org/r/20260109122523.125843-1-djakov@kernel.org Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: debugfs: Fix incorrect error handling for NULL path

2026-01-11T14:21:18+00:00

[ Upstream commit 6bfe104fd0f94d0248af22c256ce725ee087157b ] The icc_commit_set() function, used by the debugfs interface, checks the validity of the global cur_path pointer using IS_ERR_OR_NULL(). However, in the specific case where cur_path is NULL, while IS_ERR_OR_NULL(NULL) correctly evaluates to true, the subsequent call to PTR_ERR(NULL) returns 0. This causes the function to return a success code (0) instead of an error, misleading the user into believing their bandwidth request was successfully committed when, in fact, no operation was performed. Fix this by adding an explicit check to return -EINVAL if cur_path is NULL. This prevents silent failures and ensures that an invalid operational sequence is immediately and clearly reported as an error. Fixes: 770c69f037c1 ("interconnect: Add debugfs test client") Signed-off-by: Kuan-Wei Chiu Link: https://lore.kernel.org/r/20251010151447.2289779-1-visitorckw@gmail.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: msm8996: add missing link to SLAVE_USB_HS

2026-01-11T14:21:18+00:00

[ Upstream commit 8cf9b43f6b4d90e19a9341edefdd46842d4adb55 ] >From the initial submission the interconnect driver missed the link from SNOC_PNOC to the USB 2 configuration space. Add missing link in order to let the platform configure and utilize this path. Fixes: 7add937f5222 ("interconnect: qcom: Add MSM8996 interconnect provider driver") Signed-off-by: Dmitry Baryshkov Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20251002-fix-msm8996-icc-v1-1-a36a05d1f869@oss.qualcomm.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: sc8180x: specify num_nodes

2025-08-15T10:08:44+00:00

[ Upstream commit 7e0b59496a02d25828612721e846ea4b717a97b9 ] Specify .num_nodes for several BCMs which missed this declaration. Fixes: 04548d4e2798 ("interconnect: qcom: sc8180x: Reformat node and bcm definitions") Signed-off-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20250704-rework-icc-v2-2-875fac996ef5@oss.qualcomm.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: sc8280xp: specify num_links for qnm_a1noc_cfg

2025-08-15T10:08:44+00:00

[ Upstream commit 02ee375506dceb7d32007821a2bff31504d64b99 ] The qnm_a1noc_cfg declaration didn't include .num_links definition, fix it. Fixes: f29dabda7917 ("interconnect: qcom: Add SC8280XP interconnect provider") Signed-off-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20250704-rework-icc-v2-1-875fac996ef5@oss.qualcomm.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: sc7280: Add missing num_links to xm_pcie3_1 node

2025-08-01T08:47:28+00:00

[ Upstream commit 886a94f008dd1a1702ee66dd035c266f70fd9e90 ] This allows adding interconnect paths for PCIe 1 in device tree later. Fixes: 46bdcac533cc ("interconnect: qcom: Add SC7280 interconnect provider driver") Signed-off-by: Xilin Wu Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20250613-sc7280-icc-pcie1-fix-v1-1-0b09813e3b09@radxa.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: icc-clk: Add missed num_nodes initialization

2024-10-04T14:29:40+00:00

[ Upstream commit c801ed86840ec38b2a9bcafeee3d7c9e14c743f3 ] With the new __counted_by annotation, the "num_nodes" struct member must be set before accessing the "nodes" array. This initialization was done in other places where a new struct icc_onecell_data is allocated, but this case in icc_clk_register() was missed. Set "num_nodes" after allocation. Fixes: dd4904f3b924 ("interconnect: qcom: Annotate struct icc_onecell_data with __counted_by") Signed-off-by: Kees Cook Reviewed-by: Gustavo A. R. Silva Link: https://lore.kernel.org/r/20240716214819.work.328-kees@kernel.org Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: qcm2290: Fix mas_snoc_bimc RPM master ID

2024-08-03T06:53:58+00:00

[ Upstream commit cd5ce4589081190281cc2537301edd4275fe55eb ] The value was wrong, resulting in misprogramming of the hardware. Fix it. Fixes: 1a14b1ac3935 ("interconnect: qcom: Add QCM2290 driver support") Reported-by: Stephan Gerhold Closes: https://lore.kernel.org/linux-arm-msm/ZgMs_xZVzWH5uK-v@gerhold.net/ Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20240618-topic-2290_icc_2-v1-1-64446888a133@linaro.org Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: qcom: qcm2290: Fix mas_snoc_bimc QoS port assignment

2024-06-12T09:12:18+00:00

[ Upstream commit 230d05b1179f6ce6f8dc8a2b99eba92799ac22d7 ] The value was wrong, resulting in misprogramming of the hardware. Fix it. Fixes: 1a14b1ac3935 ("interconnect: qcom: Add QCM2290 driver support") Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20240326-topic-rpm_icc_qos_cleanup-v1-2-357e736792be@linaro.org Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin

interconnect: Don't access req_list while it's being manipulated

2024-04-27T15:11:37+00:00

[ Upstream commit de1bf25b6d771abdb52d43546cf57ad775fb68a1 ] The icc_lock mutex was split into separate icc_lock and icc_bw_lock mutexes in [1] to avoid lockdep splats. However, this didn't adequately protect access to icc_node::req_list. The icc_set_bw() function will eventually iterate over req_list while only holding icc_bw_lock, but req_list can be modified while only holding icc_lock. This causes races between icc_set_bw(), of_icc_get(), and icc_put(). Example A: CPU0 CPU1 ---- ---- icc_set_bw(path_a) mutex_lock(&icc_bw_lock); icc_put(path_b) mutex_lock(&icc_lock); aggregate_requests() hlist_for_each_entry(r, ... hlist_del(... Example B: CPU0 CPU1 ---- ---- icc_set_bw(path_a) mutex_lock(&icc_bw_lock); path_b = of_icc_get() of_icc_get_by_index() mutex_lock(&icc_lock); path_find() path_init() aggregate_requests() hlist_for_each_entry(r, ... hlist_add_head(... Fix this by ensuring icc_bw_lock is always held before manipulating icc_node::req_list. The additional places icc_bw_lock is held don't perform any memory allocations, so we should still be safe from the original lockdep splats that motivated the separate locks. [1] commit af42269c3523 ("interconnect: Fix locking for runpm vs reclaim") Signed-off-by: Mike Tipton Fixes: af42269c3523 ("interconnect: Fix locking for runpm vs reclaim") Reviewed-by: Rob Clark Link: https://lore.kernel.org/r/20240305225652.22872-1-quic_mdtipton@quicinc.com Signed-off-by: Georgi Djakov Signed-off-by: Sasha Levin