From patchwork Fri Jun 15 10:18:28 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mika Kuoppala X-Patchwork-Id: 10466071 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 9F082600F4 for ; Fri, 15 Jun 2018 10:18:58 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id BA99028A48 for ; Fri, 15 Jun 2018 10:18:57 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id AF3C728BD0; Fri, 15 Jun 2018 10:18:57 +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=-5.2 required=2.0 tests=BAYES_00, MAILING_LIST_MULTI, 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 71E8128A48 for ; Fri, 15 Jun 2018 10:18:56 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id A80DF6E97E; Fri, 15 Jun 2018 10:18:55 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mga18.intel.com (mga18.intel.com [134.134.136.126]) by gabe.freedesktop.org (Postfix) with ESMTPS id A612B6E97E for ; Fri, 15 Jun 2018 10:18:53 +0000 (UTC) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by orsmga106.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 15 Jun 2018 03:18:52 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.51,226,1526367600"; d="scan'208";a="57665128" Received: from rosetta.fi.intel.com ([10.237.72.186]) by fmsmga002.fm.intel.com with ESMTP; 15 Jun 2018 03:18:50 -0700 Received: by rosetta.fi.intel.com (Postfix, from userid 1000) id 0E69D8407C8; Fri, 15 Jun 2018 13:18:29 +0300 (EEST) From: Mika Kuoppala To: intel-gfx@lists.freedesktop.org Date: Fri, 15 Jun 2018 13:18:28 +0300 Message-Id: <20180615101828.24195-1-mika.kuoppala@linux.intel.com> X-Mailer: git-send-email 2.17.0 Subject: [Intel-gfx] [PATCH] drm/i915: Fix context ban and hang accounting for client X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 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 If client is smart or lucky enough to create a new context after each hang, our context banning mechanism will never catch up, and as a result of that it will be saved from client banning. This can result in a never ending streak of gpu hangs caused by bad or malicious client, preventing access from other legit gpu clients. Fix this by always incrementing per client ban score if it hangs in short successions regardless of context ban scoring. The exception are non bannable contexts. They remain detached from client ban scoring mechanism. v2: xchg timestamp, tidyup (Chris) Fixes: b083a0870c79 ("drm/i915: Add per client max context ban limit") Cc: Chris Wilson Signed-off-by: Mika Kuoppala Reviewed-by: Chris Wilson --- drivers/gpu/drm/i915/i915_drv.h | 20 ++++++--- drivers/gpu/drm/i915/i915_gem.c | 57 +++++++++++++++++-------- drivers/gpu/drm/i915/i915_gem_context.c | 2 +- 3 files changed, 54 insertions(+), 25 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 74dd88d8563e..93aa8e7dfaba 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -352,14 +352,20 @@ struct drm_i915_file_private { unsigned int bsd_engine; -/* Client can have a maximum of 3 contexts banned before - * it is denied of creating new contexts. As one context - * ban needs 4 consecutive hangs, and more if there is - * progress in between, this is a last resort stop gap measure - * to limit the badly behaving clients access to gpu. +/* Every context ban increments per client ban score. Also + * hangs in short succession increments ban score. If client suffers 3 + * context bans, 9 hangs in quick succession or combination of those, + * it is banned and submitting more work will fail. This is a stop gap + * measure to limit the badly behaving clients access to gpu. + * Note that unbannable contexts never increment the client ban score. */ -#define I915_MAX_CLIENT_CONTEXT_BANS 3 - atomic_t context_bans; +#define I915_CLIENT_SCORE_HANG_FAST 1 +#define I915_CLIENT_FAST_HANG_JIFFIES (60 * HZ) +#define I915_CLIENT_SCORE_CONTEXT_BAN 3 +#define I915_CLIENT_SCORE_BANNED 9 + /** ban_score: Accumulated score of all ctx bans and fast hangs. */ + atomic_t ban_score; + unsigned long hang_timestamp; }; /* Interface history: diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 8dd4d35655af..f06fe1c636e5 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2941,32 +2941,54 @@ i915_gem_object_pwrite_gtt(struct drm_i915_gem_object *obj, return 0; } +static void i915_gem_client_mark_guilty(struct drm_i915_file_private *file_priv, + const struct i915_gem_context *ctx) +{ + unsigned int score; + unsigned long prev_hang; + + if (i915_gem_context_is_banned(ctx)) + score = I915_CLIENT_SCORE_CONTEXT_BAN; + else + score = 0; + + prev_hang = xchg(&file_priv->hang_timestamp, jiffies); + if (time_before(jiffies, prev_hang + I915_CLIENT_FAST_HANG_JIFFIES)) + score += I915_CLIENT_SCORE_HANG_FAST; + + if (score) { + atomic_add(score, &file_priv->ban_score); + + DRM_DEBUG_DRIVER("client %s: gained %u ban score, now %u\n", + ctx->name, score, + atomic_read(&file_priv->ban_score)); + } +} + static void i915_gem_context_mark_guilty(struct i915_gem_context *ctx) { - bool banned; + unsigned int score; + bool banned, bannable; atomic_inc(&ctx->guilty_count); - banned = false; - if (i915_gem_context_is_bannable(ctx)) { - unsigned int score; + bannable = i915_gem_context_is_bannable(ctx); + score = atomic_add_return(CONTEXT_SCORE_GUILTY, &ctx->ban_score); + banned = score >= CONTEXT_SCORE_BAN_THRESHOLD; - score = atomic_add_return(CONTEXT_SCORE_GUILTY, - &ctx->ban_score); - banned = score >= CONTEXT_SCORE_BAN_THRESHOLD; + DRM_DEBUG_DRIVER("context %s: guilty %d, score %u, ban %s:%s\n", + ctx->name, atomic_read(&ctx->guilty_count), + score, yesno(bannable), yesno(banned)); - DRM_DEBUG_DRIVER("context %s marked guilty (score %d) banned? %s\n", - ctx->name, score, yesno(banned)); - } - if (!banned) + /* Cool contexts don't accumulate client ban score */ + if (!bannable) return; - i915_gem_context_set_banned(ctx); - if (!IS_ERR_OR_NULL(ctx->file_priv)) { - atomic_inc(&ctx->file_priv->context_bans); - DRM_DEBUG_DRIVER("client %s has had %d context banned\n", - ctx->name, atomic_read(&ctx->file_priv->context_bans)); - } + if (banned) + i915_gem_context_set_banned(ctx); + + if (!IS_ERR_OR_NULL(ctx->file_priv)) + i915_gem_client_mark_guilty(ctx->file_priv, ctx); } static void i915_gem_context_mark_innocent(struct i915_gem_context *ctx) @@ -5820,6 +5842,7 @@ int i915_gem_open(struct drm_i915_private *i915, struct drm_file *file) INIT_LIST_HEAD(&file_priv->mm.request_list); file_priv->bsd_engine = -1; + file_priv->hang_timestamp = jiffies; ret = i915_gem_context_open(i915, file); if (ret) diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index ef6ea4bcd773..ccf463ab6562 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -708,7 +708,7 @@ int i915_gem_switch_to_kernel_context(struct drm_i915_private *i915) static bool client_is_banned(struct drm_i915_file_private *file_priv) { - return atomic_read(&file_priv->context_bans) > I915_MAX_CLIENT_CONTEXT_BANS; + return atomic_read(&file_priv->ban_score) >= I915_CLIENT_SCORE_BANNED; } int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,