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 E47A8CD6E5D for ; Tue, 2 Jun 2026 18:14:51 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id A782D10F2B5; Tue, 2 Jun 2026 18:14:51 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=lankhorst.se header.i=@lankhorst.se header.b="n47lbTDl"; dkim-atps=neutral Received: from lankhorst.se (unknown [141.105.120.124]) by gabe.freedesktop.org (Postfix) with ESMTPS id 7AB4310F2B5 for ; Tue, 2 Jun 2026 18:14:50 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=lankhorst.se; s=default; t=1780424089; bh=FJKULz51hmWimbxX7f1e6Ft33WJ4y8pcvvX1TuzH+j0=; h=From:To:Cc:Subject:Date:From; b=n47lbTDlBCY0zJO0DoYWgjEyJ7R7RhWGu9ZGJe9EamgVFXfHEbHm1ajdfTz3dQAas drmQ3VOs4wI9DIyuOyO4WGxH1q0YC3Irc5qNiQd9ubRyV9DpizsEqUun9pk01cRUdC w7mTHVSUCYJGTm0DXJ5CT5ymwaAHQl+Dp4wJQUqPLIvYt+AbbHltMWrEy6vxtgqqjF IMdsoxrt8GhQahvTPLaoL3LQyGRbuBr5BQ9U12D6KppVI/pMhOue+1XetDqaRORy6E W+4Ayzi3dq/+MUI2JXLsh69WKgXsIEk0HsnXlh0zJFxkGvfqpSpxZ2Pf33CBnYmD/m JULESgAp7gx6A== From: Maarten Lankhorst To: intel-xe@lists.freedesktop.org Cc: Maarten Lankhorst Subject: [RFC 0/2] drm/xe: Allow pushing more jobs into LRC. Date: Tue, 2 Jun 2026 20:14:45 +0200 Message-ID: <20260602181447.131216-1-dev@lankhorst.se> X-Mailer: git-send-email 2.53.0 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-BeenThere: intel-xe@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel Xe graphics driver List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" We don't know beforehand how big a job will be, so we always assume the worst case scenario and fill the LRC based on that. Try an alternative approach, where we determine the space required in LRC right before submitting the job and pass that info to the drm scheduler. This does increase the size of xe_sched_job slightly, but does allow a bigger queue depth. Noticed when rebasing Matt Brost's series where queue depth dropped below 64. Maarten Lankhorst (2): drm/xe: Convert job preparation in 2 stages drm/xe: Set job credits based on LRC size drivers/gpu/drm/xe/xe_execlist.c | 2 +- drivers/gpu/drm/xe/xe_guc_submit.c | 2 +- drivers/gpu/drm/xe/xe_ring_ops.c | 150 ++++++++++++++---------- drivers/gpu/drm/xe/xe_ring_ops_types.h | 3 + drivers/gpu/drm/xe/xe_sched_job.c | 9 +- drivers/gpu/drm/xe/xe_sched_job_types.h | 5 + drivers/gpu/drm/xe/xe_tlb_inval_job.c | 5 +- 7 files changed, 105 insertions(+), 71 deletions(-) -- 2.53.0