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 gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id A77F1C25B75 for ; Wed, 15 May 2024 13:20:50 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 3F55610E83D; Wed, 15 May 2024 13:20:50 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="bC21dg5z"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.14]) by gabe.freedesktop.org (Postfix) with ESMTPS id 337EC10E83D for ; Wed, 15 May 2024 13:20:48 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1715779248; x=1747315248; h=date:from:to:cc:subject:message-id:reply-to:references: mime-version:in-reply-to; bh=6TTRpvnoqQhbZJbOzGLTMd1pOBaV21dfhic4HEL7aZw=; b=bC21dg5ztDNFYooEvOOscapag23MYU57ApOFxi7Up7y3CoNFmSWZk7wk iepk/sVXkIPNiNnEDeznpAjqmwRtztQS4SEFzc0rJA0aqXpuZBBC1mWVJ ubWP8YNuCNu+hxKjI0Nbyrh/QApALnF+OXoHXavQyojkqnu6n3IIcWMWc 5M8oWG0ekbybolFVtOw+6Qswh48PAgU4KnUgweBRt/9OCk5zTCvBt9RtM MW/lSuSslLBFlEFfK6HIaGM+VNtUv8A81ksGq5eMJcJ9cRuYXSiTSq3cZ +sj21pF4yZ4mTHWJvH3HisloNXmF8PGM6Vm0IGpRXdXlnl+monVvncf4p g==; X-CSE-ConnectionGUID: hjMfZ9W0RqCDyUExeNXVzg== X-CSE-MsgGUID: ftXmcPAOTHiP0QfEICmyaw== X-IronPort-AV: E=McAfee;i="6600,9927,11073"; a="15659548" X-IronPort-AV: E=Sophos;i="6.08,161,1712646000"; d="scan'208";a="15659548" Received: from fmviesa009.fm.intel.com ([10.60.135.149]) by orvoesa106.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 15 May 2024 06:20:47 -0700 X-CSE-ConnectionGUID: 8Y1JwjfKST2Lu3vrqkAAgw== X-CSE-MsgGUID: VKCOTLstSzebO4nWGSkSAg== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.08,161,1712646000"; d="scan'208";a="31073870" Received: from ideak-desk.fi.intel.com ([10.237.72.78]) by fmviesa009-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 15 May 2024 06:20:46 -0700 Date: Wed, 15 May 2024 16:20:49 +0300 From: Imre Deak To: Jani Nikula Cc: intel-gfx@lists.freedesktop.org Subject: Re: [PATCH 04/20] drm/i915/dp: Sanitize intel_dp_get_link_train_fallback_values() Message-ID: References: <20240514191418.2863344-1-imre.deak@intel.com> <20240514191418.2863344-5-imre.deak@intel.com> <87eda3cjj7.fsf@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <87eda3cjj7.fsf@intel.com> X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: imre.deak@intel.com Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" On Wed, May 15, 2024 at 04:05:32PM +0300, Jani Nikula wrote: > On Tue, 14 May 2024, Imre Deak wrote: > > Reduce the indentation in intel_dp_get_link_train_fallback_values() by > > adding separate helpers to reduce the link rate and lane count. Also > > simplify things by passing crtc_state to the function. > > > > This also prepares for later patches in the patchset adding a limitation > > on how the link params are reduced. > > Currently we have these rigid reductions of rate, lane and bpp. I don't > think we want to lock in to this policy, but rather make it more > flexible. So I think this patch is counter productive. > > My grand plan I never got around to was to create a list/array of > possible sets of link parameters that you could trivially sort, filter, > iterate, and remove parameters from. (And I don't mean a statically > written one, but a dynamically generated list.) > > In that scheme, you'd get the fallback values by advancing one element > in the data structure. And you know when you've exhausted all fallbacks > when you're at the end. None of this complicated "if we've finished x, > then try y, then z". Yes, having a different fallback order than the current one makes sense to me both BW utilization wise and because the standard describes it (as an alternative?). I don't think this patch makes it more difficult to impplement that, but the change does improve the readability of the current code. > And then you could sort the list based on bandwidth, and go from > uncompressed uhbr to compressed non-uhbr to whatever. > > > BR, > Jani. > > > > > > > Signed-off-by: Imre Deak > > --- > > .../drm/i915/display/intel_dp_link_training.c | 82 ++++++++++++------- > > 1 file changed, 51 insertions(+), 31 deletions(-) > > > > diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c > > index 1b4694b46cea7..1ea4aaf9592f1 100644 > > --- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c > > +++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c > > @@ -1109,11 +1109,37 @@ static bool intel_dp_can_link_train_fallback_for_edp(struct intel_dp *intel_dp, > > return true; > > } > > > > +static int reduce_link_rate(struct intel_dp *intel_dp, int current_rate) > > +{ > > + int rate_index; > > + int new_rate; > > + > > + rate_index = intel_dp_rate_index(intel_dp->common_rates, > > + intel_dp->num_common_rates, > > + current_rate); > > + > > + if (rate_index <= 0) > > + return -1; > > + > > + new_rate = intel_dp_common_rate(intel_dp, rate_index - 1); > > + > > + return new_rate; > > +} > > + > > +static int reduce_lane_count(struct intel_dp *intel_dp, int current_lane_count) > > +{ > > + if (current_lane_count > 1) > > + return current_lane_count >> 1; > > + > > + return -1; > > +} > > + > > static int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, > > - int link_rate, u8 lane_count) > > + const struct intel_crtc_state *crtc_state) > > { > > struct drm_i915_private *i915 = dp_to_i915(intel_dp); > > - int index; > > + int new_link_rate; > > + int new_lane_count; > > > > /* > > * TODO: Enable fallback on MST links once MST link compute can handle > > @@ -1131,36 +1157,32 @@ static int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, > > return 0; > > } > > > > - index = intel_dp_rate_index(intel_dp->common_rates, > > - intel_dp->num_common_rates, > > - link_rate); > > - if (index > 0) { > > - if (intel_dp_is_edp(intel_dp) && > > - !intel_dp_can_link_train_fallback_for_edp(intel_dp, > > - intel_dp_common_rate(intel_dp, index - 1), > > - lane_count)) { > > - drm_dbg_kms(&i915->drm, > > - "Retrying Link training for eDP with same parameters\n"); > > - return 0; > > - } > > - intel_dp->link_train.max_rate = intel_dp_common_rate(intel_dp, index - 1); > > - intel_dp->link_train.max_lane_count = lane_count; > > - } else if (lane_count > 1) { > > - if (intel_dp_is_edp(intel_dp) && > > - !intel_dp_can_link_train_fallback_for_edp(intel_dp, > > - intel_dp_max_common_rate(intel_dp), > > - lane_count >> 1)) { > > - drm_dbg_kms(&i915->drm, > > - "Retrying Link training for eDP with same parameters\n"); > > - return 0; > > - } > > - intel_dp->link_train.max_rate = intel_dp_max_common_rate(intel_dp); > > - intel_dp->link_train.max_lane_count = lane_count >> 1; > > - } else { > > + new_lane_count = crtc_state->lane_count; > > + new_link_rate = reduce_link_rate(intel_dp, crtc_state->port_clock); > > + if (new_link_rate < 0) { > > + new_lane_count = reduce_lane_count(intel_dp, crtc_state->lane_count); > > + new_link_rate = intel_dp_max_common_rate(intel_dp); > > + } > > + > > + if (new_lane_count < 0) { > > drm_err(&i915->drm, "Link Training Unsuccessful\n"); > > return -1; > > } > > > > + if (intel_dp_is_edp(intel_dp) && > > + !intel_dp_can_link_train_fallback_for_edp(intel_dp, new_link_rate, new_lane_count)) { > > + drm_dbg_kms(&i915->drm, > > + "Retrying Link training for eDP with same parameters\n"); > > + return 0; > > + } > > + > > + drm_dbg_kms(&i915->drm, "Reducing link parameters from %dx%d to %dx%d\n", > > + crtc_state->port_clock, crtc_state->lane_count, > > + new_link_rate, new_lane_count); > > + > > + intel_dp->link_train.max_rate = new_link_rate; > > + intel_dp->link_train.max_lane_count = new_lane_count; > > + > > return 0; > > } > > > > @@ -1178,9 +1200,7 @@ static void intel_dp_schedule_fallback_link_training(struct intel_dp *intel_dp, > > lt_dbg(intel_dp, DP_PHY_DPRX, > > "Link Training failed with HOBL active, not enabling it from now on\n"); > > intel_dp->hobl_failed = true; > > - } else if (intel_dp_get_link_train_fallback_values(intel_dp, > > - crtc_state->port_clock, > > - crtc_state->lane_count)) { > > + } else if (intel_dp_get_link_train_fallback_values(intel_dp, crtc_state)) { > > return; > > } > > -- > Jani Nikula, Intel