From mboxrd@z Thu Jan 1 00:00:00 1970 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=-8.8 required=3.0 tests=DKIM_INVALID,DKIM_SIGNED, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, 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 9697AC43381 for ; Fri, 1 Mar 2019 19:40:07 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 40D5F2084D for ; Fri, 1 Mar 2019 19:40:07 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="key not found in DNS" (0-bit key) header.d=codeaurora.org header.i=@codeaurora.org header.b="cbrOVO3l"; dkim=fail reason="key not found in DNS" (0-bit key) header.d=codeaurora.org header.i=@codeaurora.org header.b="VDhB5/Su" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727920AbfCATkA (ORCPT ); Fri, 1 Mar 2019 14:40:00 -0500 Received: from smtp.codeaurora.org ([198.145.29.96]:44110 "EHLO smtp.codeaurora.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727896AbfCATj6 (ORCPT ); Fri, 1 Mar 2019 14:39:58 -0500 Received: by smtp.codeaurora.org (Postfix, from userid 1000) id 760EA61A38; Fri, 1 Mar 2019 19:38:59 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1551469197; bh=Nm4UTiCkfk4wF8BgGF4YWlBhH+oZweJDfMOlYTbIFWQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=cbrOVO3lc3Vc16XVjdVUHAFKhtOgnPi16WpV//NG5KMJPrLHWtm8ieFCGmjBjntrU Z3jLZcPK/uUb1aaiuOe3DHwlWYtc/0RZ0ZTc1MuDY15c63ZkHh1AxJ2dONRdsv39EO 3DLek6b8Z0CKv7sKOLHPaJw4/Mj/03lRgBhv+oZo= Received: from jcrouse1-lnx.qualcomm.com (i-global254.qualcomm.com [199.106.103.254]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: jcrouse@smtp.codeaurora.org) by smtp.codeaurora.org (Postfix) with ESMTPSA id ECC4E61706; Fri, 1 Mar 2019 19:38:54 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1551469136; bh=Nm4UTiCkfk4wF8BgGF4YWlBhH+oZweJDfMOlYTbIFWQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=VDhB5/SuBLHKjQKgStgXRW3RfEvL56QiCr7VlTp5z6NrE5IeSrRLyY9GbDz6hwAkE b9ncqEo1ggXORfjU8HhMAGIH/WbdcKHyC3Jys6Gh3/54jz+Kc3XX/je0V9ZueQougs +sjoniw6ztV4D32SYCopuh4dw/oqqGXpWriH8/pY= DMARC-Filter: OpenDMARC Filter v1.3.2 smtp.codeaurora.org ECC4E61706 Authentication-Results: pdx-caf-mail.web.codeaurora.org; dmarc=none (p=none dis=none) header.from=codeaurora.org Authentication-Results: pdx-caf-mail.web.codeaurora.org; spf=none smtp.mailfrom=jcrouse@codeaurora.org From: Jordan Crouse To: freedreno@lists.freedesktop.org Cc: jean-philippe.brucker@arm.com, linux-arm-msm@vger.kernel.org, dianders@chromimum.org, hoegsberg@google.com, baolu.lu@linux.intel.com, Bjorn Andersson , Kees Cook , Thomas Zimmermann , Sean Paul , Sharat Masetty , dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org, Rob Clark , David Airlie , Jonathan Marek , Mamta Shukla , Daniel Vetter Subject: [RFC PATCH v1 09/15] drm/msm/gpu: Move address space setup to the GPU targets Date: Fri, 1 Mar 2019 12:38:31 -0700 Message-Id: <1551469117-3404-10-git-send-email-jcrouse@codeaurora.org> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1551469117-3404-1-git-send-email-jcrouse@codeaurora.org> References: <1551469117-3404-1-git-send-email-jcrouse@codeaurora.org> Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Move the address space steup code out of the generic msm GPU code to to the individual GPU targets. This allows us to do target specific setup such as gpummu for a2xx or split pagetables and per-instance pagetables for newer a5xx and a6xx targets. All this is at the expense of duplicated code in some of the target files but I think it pays for itself in improved code flow and flexibility. Signed-off-by: Jordan Crouse --- drivers/gpu/drm/msm/adreno/a2xx_gpu.c | 37 ++++++++++++++++------ drivers/gpu/drm/msm/adreno/a3xx_gpu.c | 50 ++++++++++++++++++++++-------- drivers/gpu/drm/msm/adreno/a4xx_gpu.c | 51 +++++++++++++++++++++++-------- drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 37 +++++++++++++++++++--- drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 37 +++++++++++++++++++--- drivers/gpu/drm/msm/adreno/adreno_gpu.c | 7 ----- drivers/gpu/drm/msm/msm_gem.h | 1 + drivers/gpu/drm/msm/msm_gpu.c | 54 ++------------------------------- drivers/gpu/drm/msm/msm_gpu.h | 2 ++ 9 files changed, 173 insertions(+), 103 deletions(-) diff --git a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c index 1f83bc1..49241d0 100644 --- a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c @@ -401,6 +401,30 @@ static struct msm_gpu_state *a2xx_gpu_state_get(struct msm_gpu *gpu) return state; } +static struct msm_gem_address_space * +a2xx_create_address_space(struct msm_gpu *gpu) +{ + struct msm_gem_address_space *aspace; + int ret; + + aspace = msm_gem_address_space_create_a2xx(&gpu->pdev->dev, gpu, + "gpu", SZ_16M, SZ_16M + 0xff * SZ_64K); + if (IS_ERR(aspace)) { + DRM_DEV_ERROR(gpu->dev->dev, + "No memory protection without MMU\n"); + return ERR_PTR(-ENXIO); + } + + ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); + if (ret) { + msm_gem_address_space_put(aspace); + return ERR_PTR(ret); + } + + return aspace; +} + + /* Register offset defines for A2XX - copy of A3XX */ static const unsigned int a2xx_register_offsets[REG_ADRENO_REGISTER_MAX] = { REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_AXXX_CP_RB_BASE), @@ -429,6 +453,7 @@ static const struct adreno_gpu_funcs funcs = { #endif .gpu_state_get = a2xx_gpu_state_get, .gpu_state_put = adreno_gpu_state_put, + .create_address_space = a2xx_create_address_space, }, }; @@ -473,16 +498,8 @@ struct msm_gpu *a2xx_gpu_init(struct drm_device *dev) adreno_gpu->reg_offsets = a2xx_register_offsets; ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, 1); - if (ret) - goto fail; - - if (!gpu->aspace) { - dev_err(dev->dev, "No memory protection without MMU\n"); - ret = -ENXIO; - goto fail; - } - - return gpu; + if (!ret) + return gpu; fail: if (a2xx_gpu) diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c index c3b4bc6..33ab5e8 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c @@ -21,6 +21,7 @@ # include #endif +#include "msm_gem.h" #include "a3xx_gpu.h" #define A3XX_INT0_MASK \ @@ -433,6 +434,41 @@ static struct msm_gpu_state *a3xx_gpu_state_get(struct msm_gpu *gpu) return state; } +static struct msm_gem_address_space * +a3xx_create_address_space(struct msm_gpu *gpu) +{ + struct msm_gem_address_space *aspace; + struct iommu_domain *iommu; + int ret; + + iommu = iommu_domain_alloc(&platform_bus_type); + if (!iommu) { + DRM_DEV_ERROR(gpu->dev->dev, + "No memory protection without IOMMU\n"); + return ERR_PTR(-ENXIO); + } + + iommu->geometry.aperture_start = SZ_16M; + iommu->geometry.aperture_end = 0xffffffff; + + aspace = msm_gem_address_space_create(&gpu->pdev->dev, iommu, "gpu"); + if (IS_ERR(aspace)) { + iommu_domain_free(iommu); + DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n", + PTR_ERR(aspace)); + return aspace; + } + + ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); + if (ret) { + msm_gem_address_space_put(aspace); + return ERR_PTR(ret); + } + + return aspace; +} + + /* Register offset defines for A3XX */ static const unsigned int a3xx_register_offsets[REG_ADRENO_REGISTER_MAX] = { REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_AXXX_CP_RB_BASE), @@ -461,6 +497,7 @@ static const struct adreno_gpu_funcs funcs = { #endif .gpu_state_get = a3xx_gpu_state_get, .gpu_state_put = adreno_gpu_state_put, + .create_address_space = a3xx_create_address_space, }, }; @@ -520,19 +557,6 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev) #endif } - if (!gpu->aspace) { - /* TODO we think it is possible to configure the GPU to - * restrict access to VRAM carveout. But the required - * registers are unknown. For now just bail out and - * limp along with just modesetting. If it turns out - * to not be possible to restrict access, then we must - * implement a cmdstream validator. - */ - DRM_DEV_ERROR(dev->dev, "No memory protection without IOMMU\n"); - ret = -ENXIO; - goto fail; - } - return gpu; fail: diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c index 18f9a8e..08a5729 100644 --- a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c @@ -15,6 +15,8 @@ # include #endif +#include "msm_gem.h" + #define A4XX_INT0_MASK \ (A4XX_INT0_RBBM_AHB_ERROR | \ A4XX_INT0_RBBM_ATB_BUS_OVERFLOW | \ @@ -530,6 +532,41 @@ static int a4xx_get_timestamp(struct msm_gpu *gpu, uint64_t *value) return 0; } +static struct msm_gem_address_space * +a4xx_create_address_space(struct msm_gpu *gpu) +{ + struct msm_gem_address_space *aspace; + struct iommu_domain *iommu; + int ret; + + iommu = iommu_domain_alloc(&platform_bus_type); + if (!iommu) { + DRM_DEV_ERROR(gpu->dev->dev, + "No memory protection without IOMMU\n"); + return ERR_PTR(-ENXIO); + } + + iommu->geometry.aperture_start = SZ_16M; + iommu->geometry.aperture_end = 0xffffffff; + + aspace = msm_gem_address_space_create(&gpu->pdev->dev, iommu, "gpu"); + if (IS_ERR(aspace)) { + iommu_domain_free(iommu); + DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n", + PTR_ERR(aspace)); + return aspace; + } + + ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); + if (ret) { + msm_gem_address_space_put(aspace); + return ERR_PTR(ret); + } + + return aspace; +} + + static const struct adreno_gpu_funcs funcs = { .base = { .get_param = adreno_get_param, @@ -547,6 +584,7 @@ static const struct adreno_gpu_funcs funcs = { #endif .gpu_state_get = a4xx_gpu_state_get, .gpu_state_put = adreno_gpu_state_put, + .create_address_space = a4xx_create_address_space, }, .get_timestamp = a4xx_get_timestamp, }; @@ -600,19 +638,6 @@ struct msm_gpu *a4xx_gpu_init(struct drm_device *dev) #endif } - if (!gpu->aspace) { - /* TODO we think it is possible to configure the GPU to - * restrict access to VRAM carveout. But the required - * registers are unknown. For now just bail out and - * limp along with just modesetting. If it turns out - * to not be possible to restrict access, then we must - * implement a cmdstream validator. - */ - DRM_DEV_ERROR(dev->dev, "No memory protection without IOMMU\n"); - ret = -ENXIO; - goto fail; - } - return gpu; fail: diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index 45662d3..3d6f414 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -1456,6 +1456,38 @@ static unsigned long a5xx_gpu_busy(struct msm_gpu *gpu) return (unsigned long)busy_time; } +static struct msm_gem_address_space * +a5xx_create_address_space(struct msm_gpu *gpu) +{ + struct msm_gem_address_space *aspace; + struct iommu_domain *iommu; + int ret; + + iommu = iommu_domain_alloc(&platform_bus_type); + if (!iommu) + return NULL; + + iommu->geometry.aperture_start = 0x100000000ULL; + iommu->geometry.aperture_end = 0x1ffffffffULL; + + aspace = msm_gem_address_space_create(&gpu->pdev->dev, iommu, "gpu"); + if (IS_ERR(aspace)) { + iommu_domain_free(iommu); + DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n", + PTR_ERR(aspace)); + return aspace; + } + + ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); + if (ret) { + msm_gem_address_space_put(aspace); + return ERR_PTR(ret); + } + + msm_mmu_set_fault_handler(aspace->mmu, gpu, a5xx_fault_handler); + return aspace; +} + static const struct adreno_gpu_funcs funcs = { .base = { .get_param = adreno_get_param, @@ -1477,6 +1509,7 @@ static const struct adreno_gpu_funcs funcs = { .gpu_busy = a5xx_gpu_busy, .gpu_state_get = a5xx_gpu_state_get, .gpu_state_put = a5xx_gpu_state_put, + .create_address_space = a5xx_create_address_space, }, .get_timestamp = a5xx_get_timestamp, }; @@ -1523,7 +1556,6 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev) adreno_gpu->registers = a5xx_registers; adreno_gpu->reg_offsets = a5xx_register_offsets; - a5xx_gpu->lm_leakage = 0x4E001A; check_speed_bin(&pdev->dev); @@ -1534,9 +1566,6 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev) return ERR_PTR(ret); } - if (gpu->aspace) - msm_mmu_set_fault_handler(gpu->aspace->mmu, gpu, a5xx_fault_handler); - /* Set up the preemption specific bits and pieces for each ringbuffer */ a5xx_preempt_init(gpu); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index 1c20d59..f2e0800 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -783,6 +783,38 @@ static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu) return (unsigned long)busy_time; } +static struct msm_gem_address_space * +a6xx_create_address_space(struct msm_gpu *gpu) +{ + struct msm_gem_address_space *aspace; + struct iommu_domain *iommu; + int ret; + + iommu = iommu_domain_alloc(&platform_bus_type); + if (!iommu) + return NULL; + + iommu->geometry.aperture_start = 0x100000000ULL; + iommu->geometry.aperture_end = 0x1ffffffffULL; + + aspace = msm_gem_address_space_create(&gpu->pdev->dev, iommu, "gpu"); + if (IS_ERR(aspace)) { + iommu_domain_free(iommu); + DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n", + PTR_ERR(aspace)); + return aspace; + } + + ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); + if (ret) { + msm_gem_address_space_put(aspace); + return ERR_PTR(ret); + } + + msm_mmu_set_fault_handler(aspace->mmu, gpu, a6xx_fault_handler); + return aspace; +} + static const struct adreno_gpu_funcs funcs = { .base = { .get_param = adreno_get_param, @@ -803,6 +835,7 @@ static const struct adreno_gpu_funcs funcs = { .gpu_set_freq = a6xx_gmu_set_freq, .gpu_state_get = a6xx_gpu_state_get, .gpu_state_put = a6xx_gpu_state_put, + .create_address_space = a6xx_create_address_space, }, .get_timestamp = a6xx_get_timestamp, }; @@ -845,9 +878,5 @@ struct msm_gpu *a6xx_gpu_init(struct drm_device *dev) return ERR_PTR(ret); } - if (gpu->aspace) - msm_mmu_set_fault_handler(gpu->aspace->mmu, gpu, - a6xx_fault_handler); - return gpu; } diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index 2cfee1a..dc9ea82 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -765,13 +765,6 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, adreno_gpu->rev = config->rev; adreno_gpu_config.ioname = "kgsl_3d0_reg_memory"; - - adreno_gpu_config.va_start = SZ_16M; - adreno_gpu_config.va_end = 0xffffffff; - /* maximum range of a2xx mmu */ - if (adreno_is_a2xx(adreno_gpu)) - adreno_gpu_config.va_end = SZ_16M + 0xfff * SZ_64K; - adreno_gpu_config.nr_rings = nr_rings; adreno_get_pwrlevels(&pdev->dev, gpu); diff --git a/drivers/gpu/drm/msm/msm_gem.h b/drivers/gpu/drm/msm/msm_gem.h index 5e21d01..777f5fb 100644 --- a/drivers/gpu/drm/msm/msm_gem.h +++ b/drivers/gpu/drm/msm/msm_gem.h @@ -21,6 +21,7 @@ #include #include #include "msm_drv.h" +#include "msm_mmu.h" /* Additional internal-use only BO flags: */ #define MSM_BO_STOLEN 0x10000000 /* try to use stolen/splash memory */ diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c index 79b71b1..ec48bb3 100644 --- a/drivers/gpu/drm/msm/msm_gpu.c +++ b/drivers/gpu/drm/msm/msm_gpu.c @@ -20,7 +20,6 @@ #include "msm_mmu.h" #include "msm_fence.h" #include "msm_gpu_trace.h" -#include "adreno/adreno_gpu.h" #include #include @@ -821,51 +820,6 @@ static int get_clocks(struct platform_device *pdev, struct msm_gpu *gpu) return 0; } -static struct msm_gem_address_space * -msm_gpu_create_address_space(struct msm_gpu *gpu, struct platform_device *pdev, - uint64_t va_start, uint64_t va_end) -{ - struct msm_gem_address_space *aspace; - int ret; - - /* - * Setup IOMMU.. eventually we will (I think) do this once per context - * and have separate page tables per context. For now, to keep things - * simple and to get something working, just use a single address space: - */ - if (!adreno_is_a2xx(to_adreno_gpu(gpu))) { - struct iommu_domain *iommu = iommu_domain_alloc(&platform_bus_type); - if (!iommu) - return NULL; - - iommu->geometry.aperture_start = va_start; - iommu->geometry.aperture_end = va_end; - - DRM_DEV_INFO(gpu->dev->dev, "%s: using IOMMU\n", gpu->name); - - aspace = msm_gem_address_space_create(&pdev->dev, iommu, "gpu"); - if (IS_ERR(aspace)) - iommu_domain_free(iommu); - } else { - aspace = msm_gem_address_space_create_a2xx(&pdev->dev, gpu, "gpu", - va_start, va_end); - } - - if (IS_ERR(aspace)) { - DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n", - PTR_ERR(aspace)); - return ERR_CAST(aspace); - } - - ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0); - if (ret) { - msm_gem_address_space_put(aspace); - return ERR_PTR(ret); - } - - return aspace; -} - int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev, struct msm_gpu *gpu, const struct msm_gpu_funcs *funcs, const char *name, struct msm_gpu_config *config) @@ -938,12 +892,8 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev, msm_devfreq_init(gpu); - gpu->aspace = msm_gpu_create_address_space(gpu, pdev, - config->va_start, config->va_end); - - if (gpu->aspace == NULL) - DRM_DEV_INFO(drm->dev, "%s: no IOMMU, fallback to VRAM carveout!\n", name); - else if (IS_ERR(gpu->aspace)) { + gpu->aspace = gpu->funcs->create_address_space(gpu); + if (IS_ERR(gpu->aspace)) { ret = PTR_ERR(gpu->aspace); goto fail; } diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h index ca17086..81b9861 100644 --- a/drivers/gpu/drm/msm/msm_gpu.h +++ b/drivers/gpu/drm/msm/msm_gpu.h @@ -74,6 +74,8 @@ struct msm_gpu_funcs { int (*gpu_state_put)(struct msm_gpu_state *state); unsigned long (*gpu_get_freq)(struct msm_gpu *gpu); void (*gpu_set_freq)(struct msm_gpu *gpu, unsigned long freq); + struct msm_gem_address_space *(*create_address_space) + (struct msm_gpu *gpu); }; struct msm_gpu { -- 2.7.4