From patchwork Tue Jul 25 20:34:44 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Doug Anderson X-Patchwork-Id: 706780 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 A7B44C001DE for ; Tue, 25 Jul 2023 20:38:04 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229877AbjGYUiC (ORCPT ); Tue, 25 Jul 2023 16:38:02 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:35422 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231308AbjGYUhq (ORCPT ); Tue, 25 Jul 2023 16:37:46 -0400 Received: from mail-pg1-x530.google.com (mail-pg1-x530.google.com [IPv6:2607:f8b0:4864:20::530]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 8A5E42D69 for ; Tue, 25 Jul 2023 13:37:22 -0700 (PDT) Received: by mail-pg1-x530.google.com with SMTP id 41be03b00d2f7-55adfa61199so4222502a12.2 for ; Tue, 25 Jul 2023 13:37:22 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=chromium.org; s=google; t=1690317442; x=1690922242; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=xwFg3zsWqs3v6aVHVxyVzqQhxIAjeJPLZHSa2wy+9Ws=; b=NiXGokZlrOVDzOQ2ltZp46BU4axptzmEuTs5CHWfOeJJepB8/LANCzCNvgd1Xt+64k g2CW15MAoAP/XZnUB6Ho0rRb+jyuN/yizpeqXU7vQKgEsA/6FOg/2z6haNgZ0rDwEmvb 0QD5h8/VBU6R7cukfQ0ip1prI0nC/2O0rpv2g= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1690317442; x=1690922242; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=xwFg3zsWqs3v6aVHVxyVzqQhxIAjeJPLZHSa2wy+9Ws=; b=bk8J7Oe/DXNIAoZqaNca53UK5ypYq3fLgMwPf2uIPYy6qJs4NMlM+My5gkLZqSQa0a soYbuYCYd3nIQ4cszzo2zm2u4/d+TnAVlchHKSG/Jwgbr91N6UeR8G0oK+CK323WqxVu GFFhlxl5Q6mHdGPE21PkM3TEZ3s0XqRAPkZ1M/aegcD3/Htyavx9+o+v7zz3Zno3m9Xo EDlkto3UO5b7PLkDMCgdmm5tzbnY2Vb1rLuwuA2moy/vxBkfN8ccsS+VAG7F/5GskrTt PFM8TCWBgk4AVI4BsLJMrxow/lEdh3bJwJ0LQA84GJzwln36jOTDIXG1FBudgE5dB1oy p7KQ== X-Gm-Message-State: ABy/qLZBUq31lSZPbQEPJgZ6feEpzXbK/9Wg6Vm/DuKrGCv1EAjmlipz HlhgrXLuh5wUevynXnr8UwxRdQ== X-Google-Smtp-Source: APBJJlHdQjvaPQjzafgjxD6Hh0+jIBP031nwcf3BkzInMUar0z4GEhlrH8PgUSsjP88Dg5AcKCIMPA== X-Received: by 2002:a17:90a:a406:b0:268:7eb:f599 with SMTP id y6-20020a17090aa40600b0026807ebf599mr232634pjp.30.1690317437028; Tue, 25 Jul 2023 13:37:17 -0700 (PDT) Received: from tictac2.mtv.corp.google.com ([2620:15c:9d:2:c363:4681:f5b8:301]) by smtp.gmail.com with ESMTPSA id bg1-20020a17090b0d8100b002676e961261sm1396951pjb.1.2023.07.25.13.37.15 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 25 Jul 2023 13:37:16 -0700 (PDT) From: Douglas Anderson To: Jiri Kosina , Benjamin Tissoires , Bjorn Andersson , Konrad Dybcio , Rob Herring , Frank Rowand , Krzysztof Kozlowski , Conor Dooley , Neil Armstrong , Sam Ravnborg , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann Cc: cros-qcom-dts-watchers@chromium.org, Chris Morgan , linux-input@vger.kernel.org, hsinyi@google.com, linux-arm-msm@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org, Dmitry Torokhov , devicetree@vger.kernel.org, Daniel Vetter , yangcong5@huaqin.corp-partner.google.com, Douglas Anderson Subject: [PATCH v3 09/10] HID: i2c-hid: Do panel follower work on the system_wq Date: Tue, 25 Jul 2023 13:34:44 -0700 Message-ID: <20230725133443.v3.9.I962bb462ede779005341c49320740ed95810021d@changeid> X-Mailer: git-send-email 2.41.0.487.g6d72f3e995-goog In-Reply-To: <20230725203545.2260506-1-dianders@chromium.org> References: <20230725203545.2260506-1-dianders@chromium.org> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-input@vger.kernel.org Turning on an i2c-hid device can be a slow process. This is why i2c-hid devices use PROBE_PREFER_ASYNCHRONOUS. Unfortunately, when we're a panel follower the i2c-hid power up sequence now blocks the power on of the panel. Let's fix that by scheduling the work on the system_wq. Signed-off-by: Douglas Anderson --- (no changes since v2) Changes in v2: - ihid_core_panel_prepare_work() is now static. - Improve documentation for smp_wmb(). drivers/hid/i2c-hid/i2c-hid-core.c | 50 +++++++++++++++++++++++++++--- 1 file changed, 46 insertions(+), 4 deletions(-) diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c index fa6d1f624342..6940e74d8acb 100644 --- a/drivers/hid/i2c-hid/i2c-hid-core.c +++ b/drivers/hid/i2c-hid/i2c-hid-core.c @@ -110,7 +110,9 @@ struct i2c_hid { struct i2chid_ops *ops; struct drm_panel_follower panel_follower; + struct work_struct panel_follower_prepare_work; bool is_panel_follower; + bool prepare_work_finished; }; static const struct i2c_hid_quirks { @@ -1062,10 +1064,12 @@ static int i2c_hid_core_initial_power_up(struct i2c_hid *ihid) return ret; } -static int i2c_hid_core_panel_prepared(struct drm_panel_follower *follower) +static void ihid_core_panel_prepare_work(struct work_struct *work) { - struct i2c_hid *ihid = container_of(follower, struct i2c_hid, panel_follower); + struct i2c_hid *ihid = container_of(work, struct i2c_hid, + panel_follower_prepare_work); struct hid_device *hid = ihid->hid; + int ret; /* * hid->version is set on the first power up. If it's still zero then @@ -1073,15 +1077,52 @@ static int i2c_hid_core_panel_prepared(struct drm_panel_follower *follower) * steps. */ if (!hid->version) - return i2c_hid_core_initial_power_up(ihid); + ret = i2c_hid_core_initial_power_up(ihid); + else + ret = i2c_hid_core_resume(ihid); - return i2c_hid_core_resume(ihid); + if (ret) + dev_warn(&ihid->client->dev, "Power on failed: %d\n", ret); + else + WRITE_ONCE(ihid->prepare_work_finished, true); + + /* + * The work APIs provide a number of memory ordering guarantees + * including one that says that memory writes before schedule_work() + * are always visible to the work function, but they don't appear to + * guarantee that a write that happened in the work is visible after + * cancel_work_sync(). We'll add a write memory barrier here to match + * with i2c_hid_core_panel_unpreparing() to ensure that our write to + * prepare_work_finished is visible there. + */ + smp_wmb(); +} + +static int i2c_hid_core_panel_prepared(struct drm_panel_follower *follower) +{ + struct i2c_hid *ihid = container_of(follower, struct i2c_hid, panel_follower); + + /* + * Powering on a touchscreen can be a slow process. Queue the work to + * the system workqueue so we don't block the panel's power up. + */ + WRITE_ONCE(ihid->prepare_work_finished, false); + schedule_work(&ihid->panel_follower_prepare_work); + + return 0; } static int i2c_hid_core_panel_unpreparing(struct drm_panel_follower *follower) { struct i2c_hid *ihid = container_of(follower, struct i2c_hid, panel_follower); + cancel_work_sync(&ihid->panel_follower_prepare_work); + + /* Match with ihid_core_panel_prepare_work() */ + smp_rmb(); + if (!READ_ONCE(ihid->prepare_work_finished)) + return 0; + return i2c_hid_core_suspend(ihid); } @@ -1149,6 +1190,7 @@ int i2c_hid_core_probe(struct i2c_client *client, struct i2chid_ops *ops, init_waitqueue_head(&ihid->wait); mutex_init(&ihid->reset_lock); + INIT_WORK(&ihid->panel_follower_prepare_work, ihid_core_panel_prepare_work); /* we need to allocate the command buffer without knowing the maximum * size of the reports. Let's use HID_MIN_BUFFER_SIZE, then we do the