From patchwork Tue Aug 3 20:13:42 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417317 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id B9B3AC4338F for ; Tue, 3 Aug 2021 20:14:07 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 80A4860F45 for ; Tue, 3 Aug 2021 20:14:07 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 80A4860F45 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 267946E8C8; Tue, 3 Aug 2021 20:13:54 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 302DE6E8BE; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499168" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499168" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776533" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 1/8] drm/i915/gt: Lock intel_engine_apply_whitelist with uncore->lock Date: Tue, 3 Aug 2021 13:13:42 -0700 Message-Id: <20210803201349.31031-2-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" Refactor intel_engine_apply_whitelist into locked and unlocked versions so that a caller who already has the lock can apply whitelist. v2: Fix sparse warning v3: (Chris) - Drop prefix and suffix for static function - Use longest to shortest line ordering for variable declaration Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Chris Wilson --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 46 ++++++++++++++------- 1 file changed, 32 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 053fa7251cd0..011cde0bc38d 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -1108,7 +1108,8 @@ void intel_gt_init_workarounds(struct drm_i915_private *i915) } static enum forcewake_domains -wal_get_fw_for_rmw(struct intel_uncore *uncore, const struct i915_wa_list *wal) +wal_get_fw(struct intel_uncore *uncore, const struct i915_wa_list *wal, + unsigned int op) { enum forcewake_domains fw = 0; struct i915_wa *wa; @@ -1117,8 +1118,7 @@ wal_get_fw_for_rmw(struct intel_uncore *uncore, const struct i915_wa_list *wal) for (i = 0, wa = wal->list; i < wal->count; i++, wa++) fw |= intel_uncore_forcewake_for_reg(uncore, wa->reg, - FW_REG_READ | - FW_REG_WRITE); + op); return fw; } @@ -1149,7 +1149,7 @@ wa_list_apply(struct intel_gt *gt, const struct i915_wa_list *wal) if (!wal->count) return; - fw = wal_get_fw_for_rmw(uncore, wal); + fw = wal_get_fw(uncore, wal, FW_REG_READ | FW_REG_WRITE); spin_lock_irqsave(&uncore->lock, flags); intel_uncore_forcewake_get__locked(uncore, fw); @@ -1188,7 +1188,7 @@ static bool wa_list_verify(struct intel_gt *gt, unsigned int i; bool ok = true; - fw = wal_get_fw_for_rmw(uncore, wal); + fw = wal_get_fw(uncore, wal, FW_REG_READ | FW_REG_WRITE); spin_lock_irqsave(&uncore->lock, flags); intel_uncore_forcewake_get__locked(uncore, fw); @@ -1477,27 +1477,45 @@ void intel_engine_init_whitelist(struct intel_engine_cs *engine) wa_init_finish(w); } -void intel_engine_apply_whitelist(struct intel_engine_cs *engine) +static void __engine_apply_whitelist(struct intel_engine_cs *engine) { const struct i915_wa_list *wal = &engine->whitelist; struct intel_uncore *uncore = engine->uncore; const u32 base = engine->mmio_base; + enum forcewake_domains fw; struct i915_wa *wa; unsigned int i; - if (!wal->count) - return; + lockdep_assert_held(&uncore->lock); + + fw = wal_get_fw(uncore, wal, FW_REG_WRITE); + intel_uncore_forcewake_get__locked(uncore, fw); for (i = 0, wa = wal->list; i < wal->count; i++, wa++) - intel_uncore_write(uncore, - RING_FORCE_TO_NONPRIV(base, i), - i915_mmio_reg_offset(wa->reg)); + intel_uncore_write_fw(uncore, + RING_FORCE_TO_NONPRIV(base, i), + i915_mmio_reg_offset(wa->reg)); /* And clear the rest just in case of garbage */ for (; i < RING_MAX_NONPRIV_SLOTS; i++) - intel_uncore_write(uncore, - RING_FORCE_TO_NONPRIV(base, i), - i915_mmio_reg_offset(RING_NOPID(base))); + intel_uncore_write_fw(uncore, + RING_FORCE_TO_NONPRIV(base, i), + i915_mmio_reg_offset(RING_NOPID(base))); + + intel_uncore_forcewake_put__locked(uncore, fw); +} + +void intel_engine_apply_whitelist(struct intel_engine_cs *engine) +{ + const struct i915_wa_list *wal = &engine->whitelist; + unsigned long flags; + + if (!wal->count) + return; + + spin_lock_irqsave(&engine->uncore->lock, flags); + __engine_apply_whitelist(engine); + spin_unlock_irqrestore(&engine->uncore->lock, flags); } static void From patchwork Tue Aug 3 20:13:43 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417315 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 78FAAC432BE for ; Tue, 3 Aug 2021 20:14:06 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 49890601FD for ; Tue, 3 Aug 2021 20:14:06 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 49890601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id F0AFF6E8C6; Tue, 3 Aug 2021 20:13:53 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 5461D6E12B; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499169" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499169" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776536" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 2/8] drm/i915/gt: Refactor _wa_add to reuse wa_index and wa_list_grow Date: Tue, 3 Aug 2021 13:13:43 -0700 Message-Id: <20210803201349.31031-3-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Chris Wilson Switch the search and grow code of the _wa_add to use _wa_index and _wa_list_grow. Signed-off-by: Chris Wilson Reviewed-by: Umesh Nerlige Ramappa --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 124 +++++++++++--------- 1 file changed, 71 insertions(+), 53 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 011cde0bc38d..94540cdb90c4 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -60,20 +60,19 @@ static void wa_init_start(struct i915_wa_list *wal, const char *name, const char #define WA_LIST_CHUNK (1 << 4) -static void wa_init_finish(struct i915_wa_list *wal) +static void wa_trim(struct i915_wa_list *wal, gfp_t gfp) { + struct i915_wa *list; + /* Trim unused entries. */ - if (!IS_ALIGNED(wal->count, WA_LIST_CHUNK)) { - struct i915_wa *list = kmemdup(wal->list, - wal->count * sizeof(*list), - GFP_KERNEL); - - if (list) { - kfree(wal->list); - wal->list = list; - } - } + list = krealloc(wal->list, wal->count * sizeof(*list), gfp); + if (list) + wal->list = list; +} +static void wa_init_finish(struct i915_wa_list *wal) +{ + wa_trim(wal, GFP_KERNEL); if (!wal->count) return; @@ -81,57 +80,60 @@ static void wa_init_finish(struct i915_wa_list *wal) wal->wa_count, wal->name, wal->engine_name); } -static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) +static int wa_index(struct i915_wa_list *wal, i915_reg_t reg) { - unsigned int addr = i915_mmio_reg_offset(wa->reg); - unsigned int start = 0, end = wal->count; - const unsigned int grow = WA_LIST_CHUNK; - struct i915_wa *wa_; + unsigned int addr = i915_mmio_reg_offset(reg); + int start = 0, end = wal->count; - GEM_BUG_ON(!is_power_of_2(grow)); + /* addr and wal->list[].reg, both include the R/W flags */ + while (start < end) { + unsigned int mid = start + (end - start) / 2; - if (IS_ALIGNED(wal->count, grow)) { /* Either uninitialized or full. */ - struct i915_wa *list; + if (i915_mmio_reg_offset(wal->list[mid].reg) < addr) + start = mid + 1; + else if (i915_mmio_reg_offset(wal->list[mid].reg) > addr) + end = mid; + else + return mid; + } - list = kmalloc_array(ALIGN(wal->count + 1, grow), sizeof(*wa), - GFP_KERNEL); - if (!list) { - DRM_ERROR("No space for workaround init!\n"); - return; - } + return -ENOENT; +} - if (wal->list) { - memcpy(list, wal->list, sizeof(*wa) * wal->count); - kfree(wal->list); - } +static int wa_list_grow(struct i915_wa_list *wal, size_t count, gfp_t gfp) +{ + struct i915_wa *list; - wal->list = list; - } + list = krealloc(wal->list, count * sizeof(*list), gfp); + if (!list) + return -ENOMEM; - while (start < end) { - unsigned int mid = start + (end - start) / 2; + wal->list = list; + return 0; +} - if (i915_mmio_reg_offset(wal->list[mid].reg) < addr) { - start = mid + 1; - } else if (i915_mmio_reg_offset(wal->list[mid].reg) > addr) { - end = mid; - } else { - wa_ = &wal->list[mid]; - - if ((wa->clr | wa_->clr) && !(wa->clr & ~wa_->clr)) { - DRM_ERROR("Discarding overwritten w/a for reg %04x (clear: %08x, set: %08x)\n", - i915_mmio_reg_offset(wa_->reg), - wa_->clr, wa_->set); - - wa_->set &= ~wa->clr; - } - - wal->wa_count++; - wa_->set |= wa->set; - wa_->clr |= wa->clr; - wa_->read |= wa->read; - return; +static void __wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) +{ + struct i915_wa *wa_; + int index; + + index = wa_index(wal, wa->reg); + if (index >= 0) { + wa_ = &wal->list[index]; + + if ((wa->clr | wa_->clr) && !(wa->clr & ~wa_->clr)) { + DRM_ERROR("Discarding overwritten w/a for reg %04x (clear: %08x, set: %08x)\n", + i915_mmio_reg_offset(wa_->reg), + wa_->clr, wa_->set); + + wa_->set &= ~wa->clr; } + + wal->wa_count++; + wa_->set |= wa->set; + wa_->clr |= wa->clr; + wa_->read |= wa->read; + return; } wal->wa_count++; @@ -149,6 +151,22 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) } } +static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) +{ + const unsigned int grow = WA_LIST_CHUNK; + + GEM_BUG_ON(!is_power_of_2(grow)); + + if (IS_ALIGNED(wal->count, grow) && /* Either uninitialized or full. */ + wa_list_grow(wal, ALIGN(wal->count + 1, grow), GFP_KERNEL)) { + DRM_ERROR("Unable to store w/a for reg %04x\n", + i915_mmio_reg_offset(wa->reg)); + return; + } + + __wa_add(wal, wa); +} + static void wa_add(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set, u32 read_mask, bool masked_reg) { From patchwork Tue Aug 3 20:13:44 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417311 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 72AD1C432BE for ; Tue, 3 Aug 2021 20:14:03 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 416AF601FD for ; Tue, 3 Aug 2021 20:14:03 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 416AF601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id EC5556E8C4; Tue, 3 Aug 2021 20:13:53 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 9F39C6E12B; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499171" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499171" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776539" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 3/8] drm/i915/gt: Check for conflicting RING_NONPRIV Date: Tue, 3 Aug 2021 13:13:44 -0700 Message-Id: <20210803201349.31031-4-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Chris Wilson Strip the encoded bits from the register offset so that we only use the address for looking up the RING_NONPRIV entry. Signed-off-by: Chris Wilson Reviewed-by: Umesh Nerlige Ramappa --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 66 +++++++++++++-------- 1 file changed, 42 insertions(+), 24 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 94540cdb90c4..7dda5a0a8e75 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -80,18 +80,44 @@ static void wa_init_finish(struct i915_wa_list *wal) wal->wa_count, wal->name, wal->engine_name); } +static u32 reg_offset(i915_reg_t reg) +{ + return i915_mmio_reg_offset(reg) & RING_FORCE_TO_NONPRIV_ADDRESS_MASK; +} + +static u32 reg_flags(i915_reg_t reg) +{ + return i915_mmio_reg_offset(reg) & ~RING_FORCE_TO_NONPRIV_ADDRESS_MASK; +} + +__maybe_unused +static bool is_nonpriv_flags_valid(u32 flags) +{ + /* Check only valid flag bits are set */ + if (flags & ~RING_FORCE_TO_NONPRIV_MASK_VALID) + return false; + + /* NB: Only 3 out of 4 enum values are valid for access field */ + if ((flags & RING_FORCE_TO_NONPRIV_ACCESS_MASK) == + RING_FORCE_TO_NONPRIV_ACCESS_INVALID) + return false; + + return true; +} + static int wa_index(struct i915_wa_list *wal, i915_reg_t reg) { - unsigned int addr = i915_mmio_reg_offset(reg); int start = 0, end = wal->count; + u32 addr = reg_offset(reg); /* addr and wal->list[].reg, both include the R/W flags */ while (start < end) { unsigned int mid = start + (end - start) / 2; + u32 pos = reg_offset(wal->list[mid].reg); - if (i915_mmio_reg_offset(wal->list[mid].reg) < addr) + if (pos < addr) start = mid + 1; - else if (i915_mmio_reg_offset(wal->list[mid].reg) > addr) + else if (pos > addr) end = mid; else return mid; @@ -117,13 +143,22 @@ static void __wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) struct i915_wa *wa_; int index; + GEM_BUG_ON(!is_nonpriv_flags_valid(reg_flags(wa->reg))); + index = wa_index(wal, wa->reg); if (index >= 0) { wa_ = &wal->list[index]; + if (i915_mmio_reg_offset(wa->reg) != + i915_mmio_reg_offset(wa_->reg)) { + DRM_ERROR("Discarding incompatible w/a for reg %04x\n", + reg_offset(wa->reg)); + return; + } + if ((wa->clr | wa_->clr) && !(wa->clr & ~wa_->clr)) { DRM_ERROR("Discarding overwritten w/a for reg %04x (clear: %08x, set: %08x)\n", - i915_mmio_reg_offset(wa_->reg), + reg_offset(wa_->reg), wa_->clr, wa_->set); wa_->set &= ~wa->clr; @@ -141,10 +176,8 @@ static void __wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) *wa_ = *wa; while (wa_-- > wal->list) { - GEM_BUG_ON(i915_mmio_reg_offset(wa_[0].reg) == - i915_mmio_reg_offset(wa_[1].reg)); - if (i915_mmio_reg_offset(wa_[1].reg) > - i915_mmio_reg_offset(wa_[0].reg)) + GEM_BUG_ON(reg_offset(wa_[0].reg) == reg_offset(wa_[1].reg)); + if (reg_offset(wa_[1].reg) > reg_offset(wa_[0].reg)) break; swap(wa_[1], wa_[0]); @@ -160,7 +193,7 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) if (IS_ALIGNED(wal->count, grow) && /* Either uninitialized or full. */ wa_list_grow(wal, ALIGN(wal->count + 1, grow), GFP_KERNEL)) { DRM_ERROR("Unable to store w/a for reg %04x\n", - i915_mmio_reg_offset(wa->reg)); + reg_offset(wa->reg)); return; } @@ -1227,21 +1260,6 @@ bool intel_gt_verify_workarounds(struct intel_gt *gt, const char *from) return wa_list_verify(gt, >->i915->gt_wa_list, from); } -__maybe_unused -static bool is_nonpriv_flags_valid(u32 flags) -{ - /* Check only valid flag bits are set */ - if (flags & ~RING_FORCE_TO_NONPRIV_MASK_VALID) - return false; - - /* NB: Only 3 out of 4 enum values are valid for access field */ - if ((flags & RING_FORCE_TO_NONPRIV_ACCESS_MASK) == - RING_FORCE_TO_NONPRIV_ACCESS_INVALID) - return false; - - return true; -} - static void whitelist_reg_ext(struct i915_wa_list *wal, i915_reg_t reg, u32 flags) { From patchwork Tue Aug 3 20:13:45 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417321 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id F27DDC43214 for ; Tue, 3 Aug 2021 20:14:10 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id C5EE8601FD for ; Tue, 3 Aug 2021 20:14:10 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org C5EE8601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id DCD906E8C9; Tue, 3 Aug 2021 20:13:54 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id ECBFA6E12B; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499173" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499173" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:51 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776546" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 4/8] drm/i915/gt: Enable dynamic adjustment of RING_NONPRIV Date: Tue, 3 Aug 2021 13:13:45 -0700 Message-Id: <20210803201349.31031-5-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Chris Wilson The OA subsystem would like to enable its privileged clients access to the OA registers from execbuf. This requires temporarily removing the HW validation from those registers for the duration of the OA client, for which we need to allow OA to dynamically adjust the set of RING_NONPRIV. Care must still be taken since the RING_NONPRIV are global, so any and all contexts that run at the same time as the OA client, will also be able to adjust the registers from their execbuf. v2: Fix memmove size (Umesh) Signed-off-by: Chris Wilson Reviewed-by: Umesh Nerlige Ramappa --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 59 +++++ drivers/gpu/drm/i915/gt/intel_workarounds.h | 7 + .../gpu/drm/i915/gt/selftest_workarounds.c | 237 ++++++++++++++++++ 3 files changed, 303 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 7dda5a0a8e75..3da7b5486251 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -200,6 +200,18 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) __wa_add(wal, wa); } +static void _wa_del(struct i915_wa_list *wal, i915_reg_t reg) +{ + struct i915_wa *wa = wal->list; + int index; + + index = wa_index(wal, reg); + if (GEM_DEBUG_WARN_ON(index < 0)) + return; + + memmove(wa + index, wa + index + 1, (--wal->count - index) * sizeof(*wa)); +} + static void wa_add(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set, u32 read_mask, bool masked_reg) { @@ -2012,6 +2024,53 @@ void intel_engine_init_workarounds(struct intel_engine_cs *engine) wa_init_finish(wal); } +int intel_engine_allow_user_register_access(struct intel_engine_cs *engine, + const i915_reg_t *reg, + unsigned int count) +{ + struct i915_wa_list *wal = &engine->whitelist; + unsigned long flags; + int err; + + if (GEM_DEBUG_WARN_ON(wal->count + count >= RING_MAX_NONPRIV_SLOTS)) + return -ENOSPC; + + spin_lock_irqsave(&engine->uncore->lock, flags); + + err = wa_list_grow(wal, wal->count + count, GFP_ATOMIC | __GFP_NOWARN); + if (err) + goto out; + + while (count--) { + struct i915_wa wa = { .reg = *reg++ }; + + __wa_add(wal, &wa); + } + + __engine_apply_whitelist(engine); + +out: + spin_unlock_irqrestore(&engine->uncore->lock, flags); + return err; +} + +void intel_engine_deny_user_register_access(struct intel_engine_cs *engine, + const i915_reg_t *reg, + unsigned int count) +{ + struct i915_wa_list *wal = &engine->whitelist; + unsigned long flags; + + spin_lock_irqsave(&engine->uncore->lock, flags); + + while (count--) + _wa_del(wal, *reg++); + + __engine_apply_whitelist(engine); + + spin_unlock_irqrestore(&engine->uncore->lock, flags); +} + void intel_engine_apply_workarounds(struct intel_engine_cs *engine) { wa_list_apply(engine->gt, &engine->wa_list); diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.h b/drivers/gpu/drm/i915/gt/intel_workarounds.h index 15abb68b6c00..3c50390e3a7f 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.h +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.h @@ -36,4 +36,11 @@ void intel_engine_apply_workarounds(struct intel_engine_cs *engine); int intel_engine_verify_workarounds(struct intel_engine_cs *engine, const char *from); +int intel_engine_allow_user_register_access(struct intel_engine_cs *engine, + const i915_reg_t *reg, + unsigned int count); +void intel_engine_deny_user_register_access(struct intel_engine_cs *engine, + const i915_reg_t *reg, + unsigned int count); + #endif diff --git a/drivers/gpu/drm/i915/gt/selftest_workarounds.c b/drivers/gpu/drm/i915/gt/selftest_workarounds.c index e623ac45f4aa..8290e02d4663 100644 --- a/drivers/gpu/drm/i915/gt/selftest_workarounds.c +++ b/drivers/gpu/drm/i915/gt/selftest_workarounds.c @@ -1177,6 +1177,242 @@ static int live_isolated_whitelist(void *arg) return err; } +static int rmw_reg(struct intel_engine_cs *engine, const i915_reg_t reg) +{ + const u32 values[] = { + 0x00000000, + 0x01010101, + 0x10100101, + 0x03030303, + 0x30300303, + 0x05050505, + 0x50500505, + 0x0f0f0f0f, + 0xf00ff00f, + 0x10101010, + 0xf0f01010, + 0x30303030, + 0xa0a03030, + 0x50505050, + 0xc0c05050, + 0xf0f0f0f0, + 0x11111111, + 0x33333333, + 0x55555555, + 0x0000ffff, + 0x00ff00ff, + 0xff0000ff, + 0xffff00ff, + 0xffffffff, + }; + struct i915_vma *vma, *batch; + struct intel_context *ce; + struct i915_request *rq; + u32 srm, lrm, idx; + u32 *cs, *results; + u64 addr; + int err; + int v; + + ce = intel_context_create(engine); + if (IS_ERR(ce)) + return PTR_ERR(ce); + + vma = __vm_create_scratch_for_read(&ce->engine->gt->ggtt->vm, + sizeof(u32)); + if (IS_ERR(vma)) { + err = PTR_ERR(vma); + goto out_context; + } + + batch = create_batch(ce->vm); + if (IS_ERR(batch)) { + err = PTR_ERR(batch); + goto out_vma; + } + + srm = MI_STORE_REGISTER_MEM; + lrm = MI_LOAD_REGISTER_MEM; + if (GRAPHICS_VER(ce->vm->i915) >= 8) + lrm++, srm++; + + addr = vma->node.start; + + cs = i915_gem_object_pin_map(batch->obj, I915_MAP_WC); + if (IS_ERR(cs)) { + err = PTR_ERR(cs); + goto out_batch; + } + + /* SRM original */ + *cs++ = srm; + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = lower_32_bits(addr); + *cs++ = upper_32_bits(addr); + + idx = 1; + for (v = 0; v < ARRAY_SIZE(values); v++) { + /* LRI garbage */ + *cs++ = MI_LOAD_REGISTER_IMM(1); + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = values[v]; + + /* SRM result */ + *cs++ = srm; + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = lower_32_bits(addr + sizeof(u32) * idx); + *cs++ = upper_32_bits(addr + sizeof(u32) * idx); + idx++; + } + for (v = 0; v < ARRAY_SIZE(values); v++) { + /* LRI garbage */ + *cs++ = MI_LOAD_REGISTER_IMM(1); + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = ~values[v]; + + /* SRM result */ + *cs++ = srm; + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = lower_32_bits(addr + sizeof(u32) * idx); + *cs++ = upper_32_bits(addr + sizeof(u32) * idx); + idx++; + } + + /* LRM original -- don't leave garbage in the context! */ + *cs++ = lrm; + *cs++ = i915_mmio_reg_offset(reg); + *cs++ = lower_32_bits(addr); + *cs++ = upper_32_bits(addr); + + *cs++ = MI_BATCH_BUFFER_END; + + i915_gem_object_flush_map(batch->obj); + i915_gem_object_unpin_map(batch->obj); + intel_gt_chipset_flush(engine->gt); + + rq = intel_context_create_request(ce); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_batch; + } + + if (engine->emit_init_breadcrumb) { /* Be nice if we hang */ + err = engine->emit_init_breadcrumb(rq); + if (err) + goto err_request; + } + + i915_vma_lock(batch); + err = i915_request_await_object(rq, batch->obj, false); + if (err == 0) + err = i915_vma_move_to_active(batch, rq, 0); + i915_vma_unlock(batch); + if (err) + goto err_request; + + i915_vma_lock(vma); + err = i915_request_await_object(rq, vma->obj, true); + if (err == 0) + err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE); + i915_vma_unlock(vma); + if (err) + goto err_request; + + err = engine->emit_bb_start(rq, batch->node.start, PAGE_SIZE, 0); + if (err) + goto err_request; + +err_request: + err = request_add_sync(rq, err); + if (err) { + pr_err("%s: Futzing %04x timedout; cancelling test\n", + engine->name, i915_mmio_reg_offset(reg)); + intel_gt_set_wedged(engine->gt); + goto out_batch; + } + + results = i915_gem_object_pin_map(vma->obj, I915_MAP_WB); + if (IS_ERR(results)) { + err = PTR_ERR(results); + goto out_batch; + } + + for (v = 0, idx = 0; v < 2 * ARRAY_SIZE(values); v++) { + if (results[++idx] != results[0]) { + err = idx; + break; + } + } + +out_batch: + i915_vma_put(batch); +out_vma: + i915_vma_put(vma); +out_context: + intel_context_put(ce); + return err; +} + +static int live_dynamic_whitelist(void *arg) +{ + struct intel_gt *gt = arg; + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err = 0; + + if (GRAPHICS_VER(gt->i915) < 8) + return 0; + + for_each_engine(engine, gt, id) { + const i915_reg_t reg = RING_MAX_IDLE(engine->mmio_base); + + err = rmw_reg(engine, reg); + if (err < 0) + break; + + if (err) { + pr_err("%s: Able to write to protected reg:%04x!\n", + engine->name, i915_mmio_reg_offset(reg)); + err = -EINVAL; + break; + } + + err = intel_engine_allow_user_register_access(engine, ®, 1); + if (err) + break; + + err = rmw_reg(engine, reg); + intel_engine_deny_user_register_access(engine, ®, 1); + if (err < 0) + break; + + if (!err) { + pr_err("%s: Unable to write to allowed reg:%04x!\n", + engine->name, i915_mmio_reg_offset(reg)); + err = -EINVAL; + break; + } + + err = rmw_reg(engine, reg); + if (err < 0) + break; + + if (err) { + pr_err("%s: Able to write to denied reg:%04x!\n", + engine->name, i915_mmio_reg_offset(reg)); + err = -EINVAL; + break; + } + + err = 0; + } + + if (igt_flush_test(gt->i915)) + err = -EIO; + + return err; +} + static bool verify_wa_lists(struct intel_gt *gt, struct wa_lists *lists, const char *str) @@ -1383,6 +1619,7 @@ int intel_workarounds_live_selftests(struct drm_i915_private *i915) SUBTEST(live_dirty_whitelist), SUBTEST(live_reset_whitelist), SUBTEST(live_isolated_whitelist), + SUBTEST(live_dynamic_whitelist), SUBTEST(live_gpu_reset_workarounds), SUBTEST(live_engine_reset_workarounds), }; From patchwork Tue Aug 3 20:13:46 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417323 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 04F71C432BE for ; Tue, 3 Aug 2021 20:14:13 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id C0AC7601FD for ; Tue, 3 Aug 2021 20:14:12 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org C0AC7601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id A5F476E8DB; Tue, 3 Aug 2021 20:13:56 +0000 (UTC) Received: from mga12.intel.com (mga12.intel.com [192.55.52.136]) by gabe.freedesktop.org (Postfix) with ESMTPS id 07B9F6E8BF; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="193364059" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="193364059" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga106.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:51 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776551" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 5/8] drm/i915/perf: Ensure observation logic is not clock gated Date: Tue, 3 Aug 2021 13:13:46 -0700 Message-Id: <20210803201349.31031-6-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Piotr Maciejewski A clock gating switch can control if the performance monitoring and observation logic is enaled or not. Ensure that we enable the clocks. v2: Separate code from other patches (Lionel) v3: Reset PMON enable when disabling perf to save power (Lionel) v4: Use intel_uncore_rmw and REG_BIT (Chris) Fixes: 00a7f0d7155c ("drm/i915/tgl: Add perf support on TGL") Signed-off-by: Piotr Maciejewski Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Lionel Landwerlin --- drivers/gpu/drm/i915/i915_perf.c | 9 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 2 ++ 2 files changed, 11 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 2f01b8c0284c..3ded6e7d8526 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -2553,6 +2553,12 @@ gen12_enable_metric_set(struct i915_perf_stream *stream, (period_exponent << GEN12_OAG_OAGLBCTXCTRL_TIMER_PERIOD_SHIFT)) : 0); + /* + * Initialize Super Queue Internal Cnt Register + * Set PMON Enable in order to collect valid metrics. + */ + intel_uncore_rmw(uncore, GEN12_SQCNT1, 0, GEN12_SQCNT1_PMON_ENABLE); + /* * Update all contexts prior writing the mux configurations as we need * to make sure all slices/subslices are ON before writing to NOA @@ -2612,6 +2618,9 @@ static void gen12_disable_metric_set(struct i915_perf_stream *stream) /* Make sure we disable noa to save power. */ intel_uncore_rmw(uncore, RPM_CONFIG1, GEN10_GT_NOA_ENABLE, 0); + + /* Reset PMON Enable to save power. */ + intel_uncore_rmw(uncore, GEN12_SQCNT1, GEN12_SQCNT1_PMON_ENABLE, 0); } static void gen7_oa_enable(struct i915_perf_stream *stream) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index bd395d967634..138b35da3057 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -718,6 +718,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OABUFFER_SIZE_16M (7 << 3) #define GEN12_OA_TLB_INV_CR _MMIO(0xceec) +#define GEN12_SQCNT1 _MMIO(0x8718) +#define GEN12_SQCNT1_PMON_ENABLE REG_BIT(30) /* Gen12 OAR unit */ #define GEN12_OAR_OACONTROL _MMIO(0x2960) From patchwork Tue Aug 3 20:13:47 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417307 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 90C3CC4320A for ; Tue, 3 Aug 2021 20:13:53 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 600CD601FD for ; Tue, 3 Aug 2021 20:13:53 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 600CD601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 469916E8C0; Tue, 3 Aug 2021 20:13:52 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 79A4A6E8BE; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499170" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499170" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776545" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 6/8] drm/i915/perf: Whitelist OA report trigger registers Date: Tue, 3 Aug 2021 13:13:47 -0700 Message-Id: <20210803201349.31031-7-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" OA reports can be triggered into the OA buffer by writing into the OAREPORTTRIG registers. Whitelist the registers to allow non-privileged user to trigger reports. Whitelist registers only if perf_stream_paranoid is set to 0. In i915_perf_open_ioctl, this setting is checked and the whitelist is enabled accordingly. On closing the perf fd, the whitelist is removed. This ensures that the access to the whitelist is gated by perf_stream_paranoid. v2: - Move related change to this patch (Lionel) - Bump up perf revision (Lionel) v3: Pardon whitelisted registers for selftest (Umesh) v4: Document supported gens for the feature (Lionel) v5: Whitelist registers only if perf_stream_paranoid is set to 0 (Jon) v6: Move oa whitelist array to i915_perf (Chris) v7: Fix OA writing beyond the wal->list memory (CI) v8: Protect write to engine whitelist registers v9: (Umesh) - Use uncore->lock to protect write to forcepriv regs - In case wal->count falls to zero on _wa_remove, make sure you still clear the registers. Remove wal->count check when applying whitelist. v10: (Umesh) - Split patches modifying intel_workarounds - On some platforms there are no whitelisted regs. intel_engine_resume applies whitelist on these platforms too and the wal->count gates such platforms. Bring back the wal->count check. - intel_engine_allow/deny_user_register_access modifies the engine whitelist and the wal->count. Use uncore->lock to serialize it with intel_engine_apply_whitelist. - Grow the wal->list when adding whitelist registers after driver load. v11: - Fix memory leak in _wa_list_grow (Chris) - Serialize kfree with engine resume using uncore->lock (Umesh) - Grow the list only if wal->count is not aligned (Umesh) Signed-off-by: Piotr Maciejewski Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Lionel Landwerlin --- drivers/gpu/drm/i915/i915_perf.c | 79 +++++++++++++++++++++++++- drivers/gpu/drm/i915/i915_perf_types.h | 8 +++ drivers/gpu/drm/i915/i915_reg.h | 12 ++-- 3 files changed, 92 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 3ded6e7d8526..30f5025b2ff6 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1364,12 +1364,56 @@ free_noa_wait(struct i915_perf_stream *stream) i915_vma_unpin_and_release(&stream->noa_wait, 0); } +static const i915_reg_t gen9_oa_wl_regs[] = { + { __OAREPORTTRIG2 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, + { __OAREPORTTRIG6 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, +}; + +static const i915_reg_t gen12_oa_wl_regs[] = { + { __GEN12_OAG_OAREPORTTRIG2 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, + { __GEN12_OAG_OAREPORTTRIG6 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, +}; + +static int intel_engine_apply_oa_whitelist(struct i915_perf_stream *stream) +{ + struct intel_engine_cs *engine = stream->engine; + int ret; + + if (i915_perf_stream_paranoid || + !(stream->sample_flags & SAMPLE_OA_REPORT) || + !stream->perf->oa_wl) + return 0; + + ret = intel_engine_allow_user_register_access(engine, + stream->perf->oa_wl, + stream->perf->num_oa_wl); + if (ret < 0) + return ret; + + stream->oa_whitelisted = true; + return 0; +} + +static void intel_engine_remove_oa_whitelist(struct i915_perf_stream *stream) +{ + struct intel_engine_cs *engine = stream->engine; + + if (!stream->oa_whitelisted) + return; + + intel_engine_deny_user_register_access(engine, + stream->perf->oa_wl, + stream->perf->num_oa_wl); +} + static void i915_oa_stream_destroy(struct i915_perf_stream *stream) { struct i915_perf *perf = stream->perf; BUG_ON(stream != perf->exclusive_stream); + intel_engine_remove_oa_whitelist(stream); + /* * Unset exclusive_stream first, it will be checked while disabling * the metric set on gen8+. @@ -1465,7 +1509,8 @@ static void gen8_init_oa_buffer(struct i915_perf_stream *stream) * bit." */ intel_uncore_write(uncore, GEN8_OABUFFER, gtt_offset | - OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT); + OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT | + GEN7_OABUFFER_EDGE_TRIGGER); intel_uncore_write(uncore, GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK); /* Mark that we need updated tail pointers to read from... */ @@ -1518,7 +1563,8 @@ static void gen12_init_oa_buffer(struct i915_perf_stream *stream) * bit." */ intel_uncore_write(uncore, GEN12_OAG_OABUFFER, gtt_offset | - OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT); + OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT | + GEN7_OABUFFER_EDGE_TRIGGER); intel_uncore_write(uncore, GEN12_OAG_OATAILPTR, gtt_offset & GEN12_OAG_OATAILPTR_MASK); @@ -3530,6 +3576,20 @@ i915_perf_open_ioctl_locked(struct i915_perf *perf, if (!(param->flags & I915_PERF_FLAG_DISABLED)) i915_perf_enable_locked(stream); + /* + * OA whitelist allows non-privileged access to some OA counters for + * triggering reports into the OA buffer. This is only allowed if + * perf_stream_paranoid is set to 0 by the sysadmin. + * + * We want to make sure this is almost the last thing we do before + * returning the stream fd. If we do end up checking for errors in code + * that follows this, we MUST call intel_engine_remove_oa_whitelist in + * the error handling path to remove the whitelisted registers. + */ + ret = intel_engine_apply_oa_whitelist(stream); + if (ret < 0) + goto err_flags; + /* Take a reference on the driver that will be kept with stream_fd * until its release. */ @@ -4410,6 +4470,8 @@ void i915_perf_init(struct drm_i915_private *i915) perf->ctx_flexeu0_offset = 0x3de; perf->gen8_valid_ctx_bit = BIT(16); + perf->oa_wl = gen9_oa_wl_regs; + perf->num_oa_wl = ARRAY_SIZE(gen9_oa_wl_regs); } } else if (GRAPHICS_VER(i915) == 11) { perf->ops.is_valid_b_counter_reg = @@ -4428,6 +4490,9 @@ void i915_perf_init(struct drm_i915_private *i915) perf->ctx_oactxctrl_offset = 0x124; perf->ctx_flexeu0_offset = 0x78e; + perf->oa_wl = gen9_oa_wl_regs; + perf->num_oa_wl = ARRAY_SIZE(gen9_oa_wl_regs); + perf->gen8_valid_ctx_bit = BIT(16); } else if (GRAPHICS_VER(i915) == 12) { perf->ops.is_valid_b_counter_reg = @@ -4445,6 +4510,9 @@ void i915_perf_init(struct drm_i915_private *i915) perf->ctx_flexeu0_offset = 0; perf->ctx_oactxctrl_offset = 0x144; + + perf->oa_wl = gen12_oa_wl_regs; + perf->num_oa_wl = ARRAY_SIZE(gen12_oa_wl_regs); } } @@ -4550,8 +4618,13 @@ int i915_perf_ioctl_version(void) * * 5: Add DRM_I915_PERF_PROP_POLL_OA_PERIOD parameter that controls the * interval for the hrtimer used to check for OA data. + * + * 6: Whitelist OATRIGGER registers to allow user to trigger reports + * into the OA buffer. This applies only to gen8+. The feature can + * only be accessed if perf_stream_paranoid is set to 0 by privileged + * user. */ - return 5; + return 6; } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) diff --git a/drivers/gpu/drm/i915/i915_perf_types.h b/drivers/gpu/drm/i915/i915_perf_types.h index aa14354a5120..f5d3eece70d8 100644 --- a/drivers/gpu/drm/i915/i915_perf_types.h +++ b/drivers/gpu/drm/i915/i915_perf_types.h @@ -312,6 +312,11 @@ struct i915_perf_stream { * buffer should be checked for available data. */ u64 poll_oa_period; + + /** + * @oa_whitelisted: Indicates that the oa registers are whitelisted. + */ + bool oa_whitelisted; }; /** @@ -432,6 +437,9 @@ struct i915_perf { u32 ctx_oactxctrl_offset; u32 ctx_flexeu0_offset; + const i915_reg_t *oa_wl; + unsigned int num_oa_wl; + /** * The RPT_ID/reason field for Gen8+ includes a bit * to determine if the CTX ID in the report is valid diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 138b35da3057..114958854d67 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -894,7 +894,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OAREPORTTRIG1_THRESHOLD_MASK 0xffff #define OAREPORTTRIG1_EDGE_LEVEL_TRIGGER_SELECT_MASK 0xffff0000 /* 0=level */ -#define OAREPORTTRIG2 _MMIO(0x2744) +#define __OAREPORTTRIG2 0x2744 +#define OAREPORTTRIG2 _MMIO(__OAREPORTTRIG2) #define OAREPORTTRIG2_INVERT_A_0 (1 << 0) #define OAREPORTTRIG2_INVERT_A_1 (1 << 1) #define OAREPORTTRIG2_INVERT_A_2 (1 << 2) @@ -947,7 +948,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OAREPORTTRIG5_THRESHOLD_MASK 0xffff #define OAREPORTTRIG5_EDGE_LEVEL_TRIGGER_SELECT_MASK 0xffff0000 /* 0=level */ -#define OAREPORTTRIG6 _MMIO(0x2754) +#define __OAREPORTTRIG6 0x2754 +#define OAREPORTTRIG6 _MMIO(__OAREPORTTRIG6) #define OAREPORTTRIG6_INVERT_A_0 (1 << 0) #define OAREPORTTRIG6_INVERT_A_1 (1 << 1) #define OAREPORTTRIG6_INVERT_A_2 (1 << 2) @@ -1008,11 +1010,13 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) /* Same layout as OAREPORTTRIGX */ #define GEN12_OAG_OAREPORTTRIG1 _MMIO(0xd920) -#define GEN12_OAG_OAREPORTTRIG2 _MMIO(0xd924) +#define __GEN12_OAG_OAREPORTTRIG2 0xd924 +#define GEN12_OAG_OAREPORTTRIG2 _MMIO(__GEN12_OAG_OAREPORTTRIG2) #define GEN12_OAG_OAREPORTTRIG3 _MMIO(0xd928) #define GEN12_OAG_OAREPORTTRIG4 _MMIO(0xd92c) #define GEN12_OAG_OAREPORTTRIG5 _MMIO(0xd930) -#define GEN12_OAG_OAREPORTTRIG6 _MMIO(0xd934) +#define __GEN12_OAG_OAREPORTTRIG6 0xd934 +#define GEN12_OAG_OAREPORTTRIG6 _MMIO(__GEN12_OAG_OAREPORTTRIG6) #define GEN12_OAG_OAREPORTTRIG7 _MMIO(0xd938) #define GEN12_OAG_OAREPORTTRIG8 _MMIO(0xd93c) From patchwork Tue Aug 3 20:13:48 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417313 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 0EB3CC4338F for ; Tue, 3 Aug 2021 20:14:05 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id D687D6104F for ; Tue, 3 Aug 2021 20:14:04 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org D687D6104F Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 6CAB46E8CA; Tue, 3 Aug 2021 20:13:54 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id CB1E46E8BE; Tue, 3 Aug 2021 20:13:51 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499172" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499172" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:51 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776549" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 7/8] drm/i915/perf: Whitelist OA counter and buffer registers Date: Tue, 3 Aug 2021 13:13:48 -0700 Message-Id: <20210803201349.31031-8-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" It is useful to have markers in the OA reports to identify triggered reports. Whitelist some OA counters that can be used as markers. A triggered report can be found faster if we can sample the HW tail and head registers when the report was triggered. Whitelist OA buffer specific registers. v2: - Bump up the perf revision (Lionel) - Use indexing for counters (Lionel) - Fix selftest for oa ticking register (Umesh) v3: Pardon whitelisted registers for selftest (Umesh) v4: - Document whitelisted registers (Lionel) - Fix live isolated whitelist for OA regs (Umesh) v5: - Free up whitelist slots. Remove GPU_TICKS and A20 counter (Piotr) - Whitelist registers only if perf_stream_paranoid is set to 0 (Jon) v6: Move oa whitelist array to i915_perf (Chris) Signed-off-by: Piotr Maciejewski Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Lionel Landwerlin --- drivers/gpu/drm/i915/i915_perf.c | 18 +++++++++++++++++- drivers/gpu/drm/i915/i915_reg.h | 16 ++++++++++++++-- 2 files changed, 31 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 30f5025b2ff6..de3d1738aabe 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1367,11 +1367,19 @@ free_noa_wait(struct i915_perf_stream *stream) static const i915_reg_t gen9_oa_wl_regs[] = { { __OAREPORTTRIG2 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, { __OAREPORTTRIG6 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, + { __OA_PERF_COUNTER_A(18) | (RING_FORCE_TO_NONPRIV_ACCESS_RW | + RING_FORCE_TO_NONPRIV_RANGE_4) }, + { __GEN8_OASTATUS | (RING_FORCE_TO_NONPRIV_ACCESS_RD | + RING_FORCE_TO_NONPRIV_RANGE_4) }, }; static const i915_reg_t gen12_oa_wl_regs[] = { { __GEN12_OAG_OAREPORTTRIG2 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, { __GEN12_OAG_OAREPORTTRIG6 | RING_FORCE_TO_NONPRIV_ACCESS_RW }, + { __GEN12_OAG_PERF_COUNTER_A(18) | (RING_FORCE_TO_NONPRIV_ACCESS_RW | + RING_FORCE_TO_NONPRIV_RANGE_4) }, + { __GEN12_OAG_OASTATUS | (RING_FORCE_TO_NONPRIV_ACCESS_RD | + RING_FORCE_TO_NONPRIV_RANGE_4) }, }; static int intel_engine_apply_oa_whitelist(struct i915_perf_stream *stream) @@ -4623,8 +4631,16 @@ int i915_perf_ioctl_version(void) * into the OA buffer. This applies only to gen8+. The feature can * only be accessed if perf_stream_paranoid is set to 0 by privileged * user. + * + * 7: Whitelist below OA registers for user to identify the location of + * triggered reports in the OA buffer. This applies only to gen8+. + * The feature can only be accessed if perf_stream_paranoid is set to + * 0 by privileged user. + * + * - OA buffer head/tail/status/buffer registers for read only + * - OA counters A18, A19, A20 for read/write */ - return 6; + return 7; } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 114958854d67..09940d7d00cf 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -695,7 +695,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN7_OASTATUS2_HEAD_MASK 0xffffffc0 #define GEN7_OASTATUS2_MEM_SELECT_GGTT (1 << 0) /* 0: PPGTT, 1: GGTT */ -#define GEN8_OASTATUS _MMIO(0x2b08) +#define __GEN8_OASTATUS 0x2b08 +#define GEN8_OASTATUS _MMIO(__GEN8_OASTATUS) #define GEN8_OASTATUS_TAIL_POINTER_WRAP (1 << 17) #define GEN8_OASTATUS_HEAD_POINTER_WRAP (1 << 16) #define GEN8_OASTATUS_OVERRUN_STATUS (1 << 3) @@ -755,7 +756,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN12_OAG_OA_DEBUG_DISABLE_GO_1_0_REPORTS (1 << 2) #define GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS (1 << 1) -#define GEN12_OAG_OASTATUS _MMIO(0xdafc) +#define __GEN12_OAG_OASTATUS 0xdafc +#define GEN12_OAG_OASTATUS _MMIO(__GEN12_OAG_OASTATUS) #define GEN12_OAG_OASTATUS_COUNTER_OVERFLOW (1 << 2) #define GEN12_OAG_OASTATUS_BUFFER_OVERFLOW (1 << 1) #define GEN12_OAG_OASTATUS_REPORT_LOST (1 << 0) @@ -998,6 +1000,16 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OAREPORTTRIG8_NOA_SELECT_6_SHIFT 24 #define OAREPORTTRIG8_NOA_SELECT_7_SHIFT 28 +/* Performance counters registers */ +#define __OA_PERF_COUNTER_A(idx) (0x2800 + 8 * (idx)) +#define OA_PERF_COUNTER_A(idx) _MMIO(__OA_PERF_COUNTER_A(idx)) +#define OA_PERF_COUNTER_A_UDW(idx) _MMIO(__OA_PERF_COUNTER_A(idx) + 4) + +/* Gen12 Performance counters registers */ +#define __GEN12_OAG_PERF_COUNTER_A(idx) (0xD980 + 8 * (idx)) +#define GEN12_OAG_PERF_COUNTER_A(idx) _MMIO(__GEN12_OAG_PERF_COUNTER_A(idx)) +#define GEN12_OAG_PERF_COUNTER_A_UDW(idx) _MMIO(__GEN12_OAG_PERF_COUNTER_A(idx) + 4) + /* Same layout as OASTARTTRIGX */ #define GEN12_OAG_OASTARTTRIG1 _MMIO(0xd900) #define GEN12_OAG_OASTARTTRIG2 _MMIO(0xd904) From patchwork Tue Aug 3 20:13:49 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Umesh Nerlige Ramappa X-Patchwork-Id: 12417319 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E2947C432BE for ; Tue, 3 Aug 2021 20:14:08 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id B1BFB601FD for ; Tue, 3 Aug 2021 20:14:08 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org B1BFB601FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 28ED26E8CC; Tue, 3 Aug 2021 20:13:55 +0000 (UTC) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 0F03F6E8BE; Tue, 3 Aug 2021 20:13:52 +0000 (UTC) X-IronPort-AV: E=McAfee;i="6200,9189,10065"; a="213499174" X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="213499174" Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:51 -0700 X-IronPort-AV: E=Sophos;i="5.84,292,1620716400"; d="scan'208";a="568776553" Received: from unerlige-ril-10.jf.intel.com ([10.165.21.208]) by orsmga004-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 03 Aug 2021 13:13:50 -0700 From: Umesh Nerlige Ramappa To: intel-gfx@lists.freedesktop.org, Lionel G Landwerlin , Ashutosh Dixit , daniel@ffwll.ch Cc: dri-devel@lists.freedesktop.org Subject: [PATCH 8/8] drm/i915/perf: Map OA buffer to user space for gen12 performance query Date: Tue, 3 Aug 2021 13:13:49 -0700 Message-Id: <20210803201349.31031-9-umesh.nerlige.ramappa@intel.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> References: <20210803201349.31031-1-umesh.nerlige.ramappa@intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" i915 used to support time based sampling mode which is good for overall system monitoring, but is not enough for query mode used to measure a single draw call or dispatch. Gen9-Gen11 are using current i915 perf implementation for query, but Gen12+ requires a new approach for query based on triggered reports within oa buffer. Triggering reports into the OA buffer is achieved by writing into a a trigger register. Optionally an unused counter/register is set with a marker value such that a triggered report can be identified in the OA buffer. Reports are usually triggered at the start and end of work that is measured. Since OA buffer is large and queries can be frequent, an efficient way to look for triggered reports is required. By knowing the current head and tail offsets into the OA buffer, it is easier to determine the locality of the reports of interest. Current perf OA interface does not expose head/tail information to the user and it filters out invalid reports before sending data to user. Also considering limited size of user buffer used during a query, creating a 1:1 copy of the OA buffer at the user space added undesired complexity. The solution was to map the OA buffer to user space provided (1) that it is accessed from a privileged user. (2) OA report filtering is not used. These 2 conditions would satisfy the safety criteria that the current perf interface addresses. To enable the query: - Add an ioctl to expose head and tail to the user - Add an ioctl to return size and offset of the OA buffer - Map the OA buffer to the user space v2: - Improve commit message (Chris) - Do not mmap based on gem object filp. Instead, use perf_fd and support mmap syscall (Chris) - Pass non-zero offset in mmap to enforce the right object is mapped (Chris) - Do not expose gpu_address (Chris) - Verify start and length of vma for page alignment (Lionel) - Move SQNTL config out (Lionel) v3: (Chris) - Omit redundant checks - Return VM_FAULT_SIGBUS is old stream is closed - Maintain reference counts to stream in vm_open and vm_close - Use switch to identify object to be mapped v4: Call kref_put on closing perf fd (Chris) v5: - Strip access to OA buffer from unprivileged child of a privileged parent. Use VM_DONTCOPY - Enforce MAP_PRIVATE by checking for VM_MAYSHARE v6: (Chris) - Use len of -1 in unmap_mapping_range - Don't use stream->oa_buffer.vma->obj in vm_fault_oa - Use kernel block comment style - do_mmap gets a reference to the file and puts it in do_munmap, so no need to maintain a reference to i915_perf_stream. Hence, remove vm_open/vm_close and stream->closed hooks/checks. (Umesh) - Do not allow mmap if SAMPLE_OA_REPORT is not set during i915_perf_open_ioctl. - Drop ioctl returning head/tail since this information is already whitelisted. Remove hooks to read head register. v7: (Chris) - unmap before destroy - change ioctl argument struct v8: Documentation and more checks (Chris) Signed-off-by: Piotr Maciejewski Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Chris Wilson --- drivers/gpu/drm/i915/gem/i915_gem_mman.c | 2 +- drivers/gpu/drm/i915/gem/i915_gem_mman.h | 2 + drivers/gpu/drm/i915/i915_perf.c | 126 ++++++++++++++++++++++- include/uapi/drm/i915_drm.h | 33 ++++++ 4 files changed, 161 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c index 5130e8ed9564..84cdce2ee447 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c @@ -213,7 +213,7 @@ compute_partial_view(const struct drm_i915_gem_object *obj, return view; } -static vm_fault_t i915_error_to_vmf_fault(int err) +vm_fault_t i915_error_to_vmf_fault(int err) { switch (err) { default: diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.h b/drivers/gpu/drm/i915/gem/i915_gem_mman.h index efee9e0d2508..1190a3a228ea 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.h @@ -29,4 +29,6 @@ void i915_gem_object_release_mmap_gtt(struct drm_i915_gem_object *obj); void i915_gem_object_release_mmap_offset(struct drm_i915_gem_object *obj); +vm_fault_t i915_error_to_vmf_fault(int err); + #endif diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index de3d1738aabe..1f8d4f3a2148 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -192,10 +192,12 @@ */ #include +#include #include #include #include "gem/i915_gem_context.h" +#include "gem/i915_gem_mman.h" #include "gt/intel_engine_pm.h" #include "gt/intel_engine_user.h" #include "gt/intel_execlists_submission.h" @@ -3322,6 +3324,44 @@ static long i915_perf_config_locked(struct i915_perf_stream *stream, return ret; } +#define I915_PERF_OA_BUFFER_MMAP_OFFSET 1 + +/** + * i915_perf_oa_buffer_info_locked - size and offset of the OA buffer + * @stream: i915 perf stream + * @cmd: ioctl command + * @arg: pointer to oa buffer info filled by this function. + */ +static int i915_perf_oa_buffer_info_locked(struct i915_perf_stream *stream, + unsigned int cmd, + unsigned long arg) +{ + struct drm_i915_perf_oa_buffer_info info; + void __user *output = (void __user *)arg; + + if (i915_perf_stream_paranoid && !perfmon_capable()) { + DRM_DEBUG("Insufficient privileges to access OA buffer info\n"); + return -EACCES; + } + + if (_IOC_SIZE(cmd) != sizeof(info)) + return -EINVAL; + + if (copy_from_user(&info, output, sizeof(info))) + return -EFAULT; + + if (info.type || info.flags || info.rsvd) + return -EINVAL; + + info.size = stream->oa_buffer.vma->size; + info.offset = I915_PERF_OA_BUFFER_MMAP_OFFSET * PAGE_SIZE; + + if (copy_to_user(output, &info, sizeof(info))) + return -EFAULT; + + return 0; +} + /** * i915_perf_ioctl_locked - support ioctl() usage with i915 perf stream FDs * @stream: An i915 perf stream @@ -3347,6 +3387,8 @@ static long i915_perf_ioctl_locked(struct i915_perf_stream *stream, return 0; case I915_PERF_IOCTL_CONFIG: return i915_perf_config_locked(stream, arg); + case I915_PERF_IOCTL_GET_OA_BUFFER_INFO: + return i915_perf_oa_buffer_info_locked(stream, cmd, arg); } return -EINVAL; @@ -3418,6 +3460,14 @@ static int i915_perf_release(struct inode *inode, struct file *file) struct i915_perf_stream *stream = file->private_data; struct i915_perf *perf = stream->perf; + /* + * User could have multiple vmas from multiple mmaps. We want to zap + * them all here. Note that a fresh fault cannot occur as the mmap holds + * a reference to the stream via the vma->vm_file, so before user's + * munmap, the stream cannot be destroyed. + */ + unmap_mapping_range(file->f_mapping, 0, -1, 1); + mutex_lock(&perf->lock); i915_perf_destroy_locked(stream); mutex_unlock(&perf->lock); @@ -3428,6 +3478,75 @@ static int i915_perf_release(struct inode *inode, struct file *file) return 0; } +static vm_fault_t vm_fault_oa(struct vm_fault *vmf) +{ + struct vm_area_struct *vma = vmf->vma; + struct i915_perf_stream *stream = vma->vm_private_data; + int err; + + err = remap_io_sg(vma, + vma->vm_start, vma->vm_end - vma->vm_start, + stream->oa_buffer.vma->pages->sgl, -1); + + return i915_error_to_vmf_fault(err); +} + +static const struct vm_operations_struct vm_ops_oa = { + .fault = vm_fault_oa, +}; + +static int i915_perf_mmap(struct file *file, struct vm_area_struct *vma) +{ + struct i915_perf_stream *stream = file->private_data; + + /* mmap-ing OA buffer to user space MUST absolutely be privileged */ + if (i915_perf_stream_paranoid && !perfmon_capable()) { + DRM_DEBUG("Insufficient privileges to map OA buffer\n"); + return -EACCES; + } + + switch (vma->vm_pgoff) { + /* + * A non-zero offset ensures that we are mapping the right object. Also + * leaves room for future objects added to this implementation. + */ + case I915_PERF_OA_BUFFER_MMAP_OFFSET: + if (!(stream->sample_flags & SAMPLE_OA_REPORT)) + return -EINVAL; + + if (vma->vm_end - vma->vm_start > stream->oa_buffer.vma->size) + return -EINVAL; + + /* + * Only support VM_READ. Enforce MAP_PRIVATE by checking for + * VM_MAYSHARE. + */ + if (vma->vm_flags & (VM_WRITE | VM_EXEC | + VM_SHARED | VM_MAYSHARE)) + return -EINVAL; + + vma->vm_flags &= ~(VM_MAYWRITE | VM_MAYEXEC); + + /* + * If the privileged parent forks and child drops root + * privilege, we do not want the child to retain access to the + * mapped OA buffer. Explicitly set VM_DONTCOPY to avoid such + * cases. + */ + vma->vm_flags |= VM_PFNMAP | VM_DONTEXPAND | + VM_DONTDUMP | VM_DONTCOPY; + break; + + default: + return -EINVAL; + } + + vma->vm_page_prot = vm_get_page_prot(vma->vm_flags); + vma->vm_private_data = stream; + vma->vm_ops = &vm_ops_oa; + + return 0; +} static const struct file_operations fops = { .owner = THIS_MODULE, @@ -3440,6 +3559,7 @@ static const struct file_operations fops = { * to handle 32bits compatibility. */ .compat_ioctl = i915_perf_ioctl, + .mmap = i915_perf_mmap, }; @@ -4639,8 +4759,12 @@ int i915_perf_ioctl_version(void) * * - OA buffer head/tail/status/buffer registers for read only * - OA counters A18, A19, A20 for read/write + * + * 8: Added an option to map oa buffer at umd driver level and trigger + * oa reports within oa buffer from command buffer. See + * I915_PERF_IOCTL_GET_OA_BUFFER_INFO. */ - return 7; + return 8; } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h index 7f13d241417f..7f6493bb5bba 100644 --- a/include/uapi/drm/i915_drm.h +++ b/include/uapi/drm/i915_drm.h @@ -2397,6 +2397,39 @@ struct drm_i915_perf_open_param { */ #define I915_PERF_IOCTL_CONFIG _IO('i', 0x2) +/* + * Returns OA buffer properties to be used with mmap. + * + * This ioctl is available in perf revision 8. + */ +#define I915_PERF_IOCTL_GET_OA_BUFFER_INFO _IOWR('i', 0x3, struct drm_i915_perf_oa_buffer_info) + +/** + * OA buffer size and offset. + * + * OA output buffer + * type: 0 + * flags: mbz + * + * After querying the info, pass (size,offset) to mmap(), + * + * mmap(0, info.size, PROT_READ, MAP_PRIVATE, perf_fd, info.offset). + * + * Note that only a private (not shared between processes, or across fork()) + * read-only mmapping is allowed. + * + * Userspace must treat the incoming data as tainted, but it conforms to the OA + * format as specified by user config. The buffer provides reports that have + * OA counters - A, B and C. + */ +struct drm_i915_perf_oa_buffer_info { + __u32 type; /* in */ + __u32 flags; /* in */ + __u64 size; /* out */ + __u64 offset; /* out */ + __u64 rsvd; /* mbz */ +}; + /* * Common to all i915 perf records */