From patchwork Tue Feb 28 13:43:57 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Konrad Dybcio X-Patchwork-Id: 657923 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 41965C7EE2E for ; Tue, 28 Feb 2023 13:44:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229582AbjB1No1 (ORCPT ); Tue, 28 Feb 2023 08:44:27 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:42976 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229571AbjB1No0 (ORCPT ); Tue, 28 Feb 2023 08:44:26 -0500 Received: from mail-lj1-x234.google.com (mail-lj1-x234.google.com [IPv6:2a00:1450:4864:20::234]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 5EAAC2CFD5 for ; Tue, 28 Feb 2023 05:44:24 -0800 (PST) Received: by mail-lj1-x234.google.com with SMTP id h9so10290551ljq.2 for ; Tue, 28 Feb 2023 05:44:24 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1677591862; h=cc:to:content-transfer-encoding:mime-version:message-id:date :subject:from:from:to:cc:subject:date:message-id:reply-to; bh=livAJ6GGfJ6Yo5cEGXwtOwLMX6hnHXxiMTTXO2jFdVs=; b=z3jaSsasOAIikO5Hh7k77PFLaPQashXi4aKfFyM4Cbw3FBvpr6HWNFecxRy4ohUBpB sMHKf2vqg/esSUqC2oS1xZGNcjEvxK60M6+pX49wmKP4J78+LLvHIMxQZyTMACPRkbQS y2lT8HFWBeHeXqXmIR2drZaqxYSY0tO2VY/QUFo6yCiCqYMQqUgDILsWRdYWQDnFoloO NEH5wQq1UDDVn0RWEWlPPICFYiZ/iXFwUfGYS7620HdwhGq8B/djGaT4AsjXlb/YYi4x syDiG5WixlrCjqJy04omFRDZexKdBD8KFjakTvEudtSJ4x2b/PwPQIeXwtUUYBcaavpF dlmw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; t=1677591862; h=cc:to:content-transfer-encoding:mime-version:message-id:date :subject:from:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=livAJ6GGfJ6Yo5cEGXwtOwLMX6hnHXxiMTTXO2jFdVs=; b=QI+CsMjxvWBZ08ncboA4IkWDAsuId0nagcJCp+ZEl1iChxQYSD9dS0y39fq4CSOTpF Wdh/hdEoicZoH7lvBjAIu9XsF2uszjFkZObnLI5G72+pfblndKh4hDYfxd2KQ81jlXR8 YE9H2YFdkq0ZARiNMicLGE62M2+mR6QDvKSvCb82Fdo1iL98gDK6OnRtAu4f5HgjaY13 O5bhI08yu+ry4xfl1EP4r7BcPXtJByAXCPIdF8n+FHx0AeRj57mc01ZHFunUWRSEJ68P NotGkMUqzbB7aaPxpREpvAPsDq6K7STa2rn/bJaej9lYHiiyNnYU5ZiRX0rX7XNmpoa1 RWgw== X-Gm-Message-State: AO0yUKUp4D5J1MdpVH9XnkohVOHDUrcnyIX2z1E1IuBS3sURX9lh0pas Dmo8BACZNnY2YStY4OacnrfbBb3dERGUxmZ0 X-Google-Smtp-Source: AK7set8hFpxLuUgdrVruez1s9H5DZshtbOi559zF7oSN13kteXR59EYavpga8mQRu+ec6DOwmJvNhw== X-Received: by 2002:a2e:a544:0:b0:295:9517:b98f with SMTP id e4-20020a2ea544000000b002959517b98fmr3697534ljn.15.1677591862530; Tue, 28 Feb 2023 05:44:22 -0800 (PST) Received: from [192.168.1.101] (abym99.neoplus.adsl.tpnet.pl. [83.9.32.99]) by smtp.gmail.com with ESMTPSA id j20-20020a2e3c14000000b00295a583a20bsm1203975lja.74.2023.02.28.05.44.21 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 28 Feb 2023 05:44:22 -0800 (PST) From: Konrad Dybcio Subject: [PATCH v6 0/9] The great interconnecification fixation Date: Tue, 28 Feb 2023 14:43:57 +0100 Message-Id: <20230228-topic-qos-v6-0-3c37a349656f@linaro.org> MIME-Version: 1.0 X-B4-Tracking: v=1; b=H4sIAB0F/mMC/x2N0QqDMAxFf0XyvIBUHdVfGXtoY5wBaV2ziSD+u 8HHczmXc4ByEVYYqgMKb6KSk8HzUQHNIX0YZTQGV7umds7jL69C+M2KXdt0FMj70fdgfgzKGEt INNsj/ZfFxrXwJPsdeL3P8wKa5SGocAAAAA== To: Andy Gross , Bjorn Andersson , Georgi Djakov , Dmitry Baryshkov Cc: Bryan O'Donoghue , linux-arm-msm@vger.kernel.org, linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org, Konrad Dybcio X-Mailer: b4 0.12.1 X-Developer-Signature: v=1; a=ed25519-sha256; t=1677591861; l=6806; i=konrad.dybcio@linaro.org; s=20230215; h=from:subject:message-id; bh=wExNDR7+0rE75/CJ66KXmYxMQAnJVJqPhLy0GyiD2C8=; b=f04Z83CivUuXBPWMxjSQTWwGgy9jZWJW/oBqwv4yaraRbUAVVgntM8s91HEThkOwONKaAAvy1n41 MzdJwL49C4CVMPTGdEbExG2nLVFgUihk1smI6sRmwoNRcPaou4rc X-Developer-Key: i=konrad.dybcio@linaro.org; a=ed25519; pk=iclgkYvtl2w05SSXO5EjjSYlhFKsJ+5OSZBjOkQuEms= Precedence: bulk List-ID: X-Mailing-List: linux-arm-msm@vger.kernel.org Hi! v5 -> v6 changelog: - Completely rewrite the commit message of [1/9], I realized that there was actually no issue with the present upstream setups and the only drivers suffering from ghost votes were.. my own OOT drivers.. As a consequence of that, all fixes tags were dropped and the patch has been kept, since it was deemed useful for newer SoCs that don't distinguish ap_owned nodes. - Change the number of allowed bus_clocks from (0-2 in the previous revision, 0-inf in the current upstream state) to {0, 2}. Scaling is only possible with a pair of wake-sleep clocks, but some providers don't do scaling at all (see 8996 A0NoC, 660 GNoC). Drop the cheeky -1 / 0 / >0 checks from the previous revision. [7/9] - bus_clocks are now forced to be named "bus", "bus_a", as there is no need for variance here - we don't do scaling on non-SMD RPM bus clocks. [7/9] - The interface clocks are now only turned on when the associated bus is running at a non-zero frequency [6/9] instead of being always on and leaking power Tested on MSM8996 Kagura, SM6375 PDX225 (OOT), MSM8998 Maple (OOT) Link to v5: https://lore.kernel.org/linux-arm-msm/20230217-topic-icc-fixes-v5-v5-0-c9a550f9fdb9@linaro.org/ v4 -> v5 changelog: - Previously the "Always set QoS params on QNoC" contained part of what should have been included in "make QoS INVALID default".. (very bad) Fix it! - Drop negative offset and keep_alive, they will be resubmitted with new icc driver submissions - use b4 this time.. hopefully the series gets to everybody now Link to v4: https://lore.kernel.org/linux-arm-msm/20230214143720.2416762-1-konrad.dybcio@linaro.org/ v3 -> v4 changelog: - Drop "Always set QoS params on QNoC", it only causes issues.. this can be investigated another day, as it's not necessary for operation - Drop "Add a way to always set QoS registers", same as /\ - Add a way (and use it) to have no bus_clocks (the ones we set rate on), as at least msm8996 has a bus (A0NoC) that doesn't have any and does all the scaling through RPM requests - Promote 8996 icc to core_initcall - Introduce keep_alive (see patch [11/12]) (important!, will be used by at least 6375) - Allow negative QoS offsets in preparation for introducing 8998 icc [12/12] Link to v3: https://lore.kernel.org/linux-arm-msm/20230116132152.405535-1-konrad.dybcio@linaro.org/ v2 -> v3 changelog: - Drop "Don't set QoS params before non-zero bw is requested" - Rebase on next - [1/9] ("..make QoS INVALID default.."): remove unused define for MODE_INVALID_VAL - Pick up tags v1 -> v2 changelog: - reorder "make QoS INVALID default", makes more sense to have it before "Always set QoS params on QNoC" - Limit ap_owned-independent QoS setting to QNoC only - Add new patches for handling the 8996-and-friends clocks situation and optional BIMC regardless-of-ap_owned QoS programming [1] https://lore.kernel.org/linux-arm-msm/14e06574-f95e-8960-0243-8c95a1c294e9@linaro.org/T/#m056692bea71d4c272968d5e07afbd9eb07a88123 [2] https://lore.kernel.org/linux-arm-msm/20230110132202.956619-1-konrad.dybcio@linaro.org/ This series grew quite a bit bigger than the previous [1] attempt, so I decided to also add a cover letter. Link to v2: [2] It addresses a few things that were not quite right: - Setting QoS params before a "real" (non-zero) bandwidth request makes little sense (since there's no data supposed to flow through the bus, why would the QoS matter) and (at least newer) downstream prevents that from happening. Do the same in Patch 1. - QNoC type buses expect to always have their QoS registers set as long as there's a non-INVALID QoS mode set; ap_owned is not really a thing on these anymore, Patch 3 handles that. - The recent MSM8996 boot fix was done quickly and not quite properly, leading to possibly setting the aggregate bus rate on "normal" hardware interface clocks; this series handles that by limiting the number of bus_clocks to 2 (which is the maximum that makes sense, anyway) and handling the rest as "intf_clocks", which are required to access the hardware at the other end. Patches 5-8 take care of that and Patch 10 reverts the _optional moniker in clk_get_ to make sure we always have the bus scaling clocks, as they're well, kind of important ;) - Similarly to QNoC, BIMC on "newer" (which can be loosely approximated by "new enough" == "has only BIMC and QNoC hosts") SoCs expects to always receive QoS programming, whereas BIMC on "older" SoCs cries like a wild boar and crashes the platform when trying to do so unconditionally. Patch 9 adds a way to take care of that for newer SoCs (like SM6375) - QoS mode INVALID was assumed by developers before to be the default ("I didn't specify any QoS settings, so the driver can't assume I did.. right? right!?" - wrong, partial struct initialization led to 0 being set and 0 corresponded to QoS mode FIXED). Make it so, as that's the logical choice. This allows the "Always set QoS params on QNoC" patch to work without setting tons of what-should- -obviously-be-the-default values everywhere, as well as fixes older drivers that set ap_owned = true but left the QoS mode field unset. Patch 2 cleans that up. - Some nodes are physically connected over more than one channel (usually DDR or other high-throughput paths). Patch 4 allows that to be reflected in calculations. This will be required for at least MSM8998 and SM6375 (which will be submitted soon after this lands) Signed-off-by: Konrad Dybcio --- Konrad Dybcio (9): interconnect: qcom: rpm: make QoS INVALID default interconnect: qcom: rpm: Add support for specifying channel num interconnect: qcom: Sort kerneldoc entries interconnect: qcom: rpm: Rename icc desc clocks to bus_blocks interconnect: qcom: rpm: Rename icc provider num_clocks to num_bus_clocks interconnect: qcom: rpm: Handle interface clocks interconnect: qcom: icc-rpm: Enforce 2 or 0 bus clocks interconnect: qcom: rpm: Don't use clk_get_optional for bus clocks anymore interconnect: qcom: msm8996: Promote to core_initcall drivers/interconnect/qcom/icc-rpm.c | 91 +++++++++++++++++++++++++------------ drivers/interconnect/qcom/icc-rpm.h | 38 +++++++++++----- drivers/interconnect/qcom/msm8996.c | 35 ++++++++------ drivers/interconnect/qcom/sdm660.c | 17 +++---- 4 files changed, 116 insertions(+), 65 deletions(-) --- base-commit: 058f4df42121baadbb8a980c06011e912784dbd2 change-id: 20230228-topic-qos-5435cac88d89 Best regards,