From patchwork Tue Sep 13 13:38:57 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Zanoni, Paulo R" X-Patchwork-Id: 9329145 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id D3C7A6077F for ; Tue, 13 Sep 2016 13:39:50 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id C2ECB2945B for ; Tue, 13 Sep 2016 13:39:50 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id B4A7829465; Tue, 13 Sep 2016 13:39:50 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.2 required=2.0 tests=BAYES_00, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 5F4C72945B for ; Tue, 13 Sep 2016 13:39:47 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 3FE1A6E112; Tue, 13 Sep 2016 13:39:46 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by gabe.freedesktop.org (Postfix) with ESMTPS id 82F766E112 for ; Tue, 13 Sep 2016 13:39:44 +0000 (UTC) Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by orsmga102.jf.intel.com with ESMTP; 13 Sep 2016 06:39:43 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos; i="5.30,328,1470726000"; d="scan'208"; a="1039225883" Received: from przanoni-mobl.amr.corp.intel.com ([10.254.180.219]) by fmsmga001.fm.intel.com with ESMTP; 13 Sep 2016 06:39:41 -0700 From: Paulo Zanoni To: intel-gfx@lists.freedesktop.org Date: Tue, 13 Sep 2016 10:38:57 -0300 Message-Id: <1473773937-19758-1-git-send-email-paulo.r.zanoni@intel.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1473715543.2435.67.camel@intel.com> References: <1473715543.2435.67.camel@intel.com> Cc: Paulo Zanoni , Stefan Richter , Dhinakaran Pandiyan , "stevenhoneyman @ gmail . com" Subject: [Intel-gfx] [PATCH] drm/i915/fbc: disable FBC on FIFO underruns X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP Ever since I started working on FBC I was already aware that FBC can really amplify the FIFO underrun symptoms. On systems where FIFO underruns were harmless error messages, enabling FBC would cause the underruns to give black screens. We recently tried to enable FBC on Haswell and got reports of a system that would hang after some hours of uptime, and the first bad commit was the one that enabled FBC. We also observed that this system had FIFO underrun error messages on its dmesg. Although we don't have any evidence that fixing the underruns would solve the bug and make FBC work properly on this machine, IMHO it's better if we minimize the amount of possible problems by just giving up FBC whenever we detect an underrun. v2: New version, different implementation and commit message. v3: Clarify the fact that we run from an IRQ handler (Chris). v4: Also add the underrun_detected check at can_choose() to avoid misleading dmesg messages (DK). v5: Fix Engrish, use READ_ONCE on the unlocked read (Chris). Cc: Stefan Richter Cc: Lyude Cc: stevenhoneyman@gmail.com Cc: Dhinakaran Pandiyan Cc: Chris Wilson Signed-off-by: Paulo Zanoni Reviewed-by: Chris Wilson Reviewed-by: Dhinakaran Pandiyan --- drivers/gpu/drm/i915/i915_drv.h | 3 ++ drivers/gpu/drm/i915/intel_drv.h | 1 + drivers/gpu/drm/i915/intel_fbc.c | 67 ++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_fifo_underrun.c | 2 + 4 files changed, 73 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 20b7743..5b1c241 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -971,6 +971,9 @@ struct intel_fbc { bool enabled; bool active; + bool underrun_detected; + struct work_struct underrun_work; + struct intel_fbc_state_cache { struct { unsigned int mode_flags; diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index abe7a4d..0d05712 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1508,6 +1508,7 @@ void intel_fbc_invalidate(struct drm_i915_private *dev_priv, void intel_fbc_flush(struct drm_i915_private *dev_priv, unsigned int frontbuffer_bits, enum fb_op_origin origin); void intel_fbc_cleanup_cfb(struct drm_i915_private *dev_priv); +void intel_fbc_handle_fifo_underrun_irq(struct drm_i915_private *dev_priv); /* intel_hdmi.c */ void intel_hdmi_init(struct drm_device *dev, i915_reg_t hdmi_reg, enum port port); diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 5dcb81a..10ca8cd 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -774,6 +774,14 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc) struct intel_fbc *fbc = &dev_priv->fbc; struct intel_fbc_state_cache *cache = &fbc->state_cache; + /* We don't need to use a state cache here since this information is + * global for all CRTC. + */ + if (fbc->underrun_detected) { + fbc->no_fbc_reason = "underrun detected"; + return false; + } + if (!cache->plane.visible) { fbc->no_fbc_reason = "primary plane not visible"; return false; @@ -859,6 +867,11 @@ static bool intel_fbc_can_choose(struct intel_crtc *crtc) return false; } + if (fbc->underrun_detected) { + fbc->no_fbc_reason = "underrun detected"; + return false; + } + if (fbc_on_pipe_a_only(dev_priv) && crtc->pipe != PIPE_A) { fbc->no_fbc_reason = "no enabled pipes can have FBC"; return false; @@ -1221,6 +1234,59 @@ void intel_fbc_global_disable(struct drm_i915_private *dev_priv) cancel_work_sync(&fbc->work.work); } +static void intel_fbc_underrun_work_fn(struct work_struct *work) +{ + struct drm_i915_private *dev_priv = + container_of(work, struct drm_i915_private, fbc.underrun_work); + struct intel_fbc *fbc = &dev_priv->fbc; + + mutex_lock(&fbc->lock); + + /* Maybe we were scheduled twice. */ + if (fbc->underrun_detected) + goto out; + + DRM_DEBUG_KMS("Disabling FBC due to FIFO underrun.\n"); + fbc->underrun_detected = true; + + intel_fbc_deactivate(dev_priv); +out: + mutex_unlock(&fbc->lock); +} + +/** + * intel_fbc_handle_fifo_underrun_irq - disable FBC when we get a FIFO underrun + * @dev_priv: i915 device instance + * + * Without FBC, most underruns are harmless and don't really cause too many + * problems, except for an annoying message on dmesg. With FBC, underruns can + * become black screens or even worse, especially when paired with bad + * watermarks. So in order for us to be on the safe side, completely disable FBC + * in case we ever detect a FIFO underrun on any pipe. An underrun on any pipe + * already suggests that watermarks may be bad, so try to be as safe as + * possible. + * + * This function is called from the IRQ handler. + */ +void intel_fbc_handle_fifo_underrun_irq(struct drm_i915_private *dev_priv) +{ + struct intel_fbc *fbc = &dev_priv->fbc; + + if (!fbc_supported(dev_priv)) + return; + + /* There's no guarantee that underrun_detected won't be set to true + * right after this check and before the work is scheduled, but that's + * not a problem since we'll check it again under the work function + * while FBC is locked. This check here is just to prevent us from + * unnecessarily scheduling the work, and it relies on the fact that we + * never switch underrun_detect back to false after it's true. */ + if (READ_ONCE(fbc->underrun_detected)) + return; + + schedule_work(&fbc->underrun_work); +} + /** * intel_fbc_init_pipe_state - initialize FBC's CRTC visibility tracking * @dev_priv: i915 device instance @@ -1292,6 +1358,7 @@ void intel_fbc_init(struct drm_i915_private *dev_priv) enum pipe pipe; INIT_WORK(&fbc->work.work, intel_fbc_work_fn); + INIT_WORK(&fbc->underrun_work, intel_fbc_underrun_work_fn); mutex_init(&fbc->lock); fbc->enabled = false; fbc->active = false; diff --git a/drivers/gpu/drm/i915/intel_fifo_underrun.c b/drivers/gpu/drm/i915/intel_fifo_underrun.c index 2aa7440..ebb4fed 100644 --- a/drivers/gpu/drm/i915/intel_fifo_underrun.c +++ b/drivers/gpu/drm/i915/intel_fifo_underrun.c @@ -372,6 +372,8 @@ void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv, if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false)) DRM_ERROR("CPU pipe %c FIFO underrun\n", pipe_name(pipe)); + + intel_fbc_handle_fifo_underrun_irq(dev_priv); } /**