From nobody Sun Oct 5 05:31:37 2025 Received: from bali.collaboradmins.com (bali.collaboradmins.com [148.251.105.195]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4C40024EA9D; Fri, 8 Aug 2025 20:03:59 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=148.251.105.195 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1754683441; cv=none; b=NXHK7qjk0/lTJpYw6rMm5l3XewDoMJQw7suK4I9RDXNxYbBJ/R9bTZjf5VUjguDbWwFkM/NZ3b9IVi/d9mA5bX8fFVrhnxOES+0v4gUpQpQFISfMeG5RsnuD08vZuNhMVbfmEWXMBLaoCnFuuNFkz0mo91CyrvIYs4UiOopac4U= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1754683441; c=relaxed/simple; bh=YrvUtegmo23fxwFf79xVFdziS5YscXEdSSJGPLuYr6k=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=WZia+IWTnqDIVmgYLsRM/f+Txb//uROqybc0M9WU658u2Osjti3ptFbCr36BmPeurTVd1BzC4faBjLKCb2aiBT02PGidBUCpfgFVFDiEwW5ZGoz0lWMu5sStyCy494yirLZundjX2XM+nYKvitwoW6hjAWTJYtcI828/ZqAkiPs= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com; spf=pass smtp.mailfrom=collabora.com; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b=VIULNBAm; arc=none smtp.client-ip=148.251.105.195 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=collabora.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b="VIULNBAm" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=collabora.com; s=mail; t=1754683436; bh=YrvUtegmo23fxwFf79xVFdziS5YscXEdSSJGPLuYr6k=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=VIULNBAmi1UWAN84xlqiSUIeObKgnU+OT1+7+Femq3kY4Fp3wumyfWaFOGSWRsccl B+PM3wKIdW5bqJi8ICve2U3DJnoYBjgmYYRRg/m6HA21UAVnghrfUMqZNtJpsPT0/U zKRwC5xYWnFbgdDvT7TyibKbjDvWYWAtUBbG2XfDiInOJuxp47RBe62cG+jYzcupB/ mq/XE00vsWBOtSH1gqLKMqsErOX2NMXSMCqM8EyHJ3jQPtISrQFLb+tMM3DDT6rmki cMWU+WEX6OrP4iG1Y37KjCYggF0R+tuXlDET6xIWNWqMd3tJtypOo6YJcWi9nE5hoJ DXEs4SKrkOkdQ== Received: from earth.mtl.collabora.ca (mtl.collabora.ca [66.171.169.34]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) (Authenticated sender: detlev) by bali.collaboradmins.com (Postfix) with ESMTPSA id 4508C17E1322; Fri, 8 Aug 2025 22:03:55 +0200 (CEST) From: Detlev Casanova To: linux-kernel@vger.kernel.org Cc: Detlev Casanova , Mauro Carvalho Chehab , Heiko Stuebner , linux-media@vger.kernel.org, linux-rockchip@lists.infradead.org, linux-arm-kernel@lists.infradead.org, kernel@collabora.com Subject: [PATCH v2 06/12] media: rkvdec: Add RCB and SRAM support Date: Fri, 8 Aug 2025 16:03:28 -0400 Message-ID: <20250808200340.156393-7-detlev.casanova@collabora.com> X-Mailer: git-send-email 2.50.1 In-Reply-To: <20250808200340.156393-1-detlev.casanova@collabora.com> References: <20250808200340.156393-1-detlev.casanova@collabora.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" The RCB (Rows and Cols Buffers) are a set of buffers used by other variations of the decoder to store temporary data. Those variation come with a dedicated SRAM area used to store those buffers for better performances. The buffer sizes are either the width or height of the frame being decoded multiplied by a documented factor and can be stored either in SRAM or RAM. A fallback to RAM is provided if the SRAM is full (e.g.: multiple streams are being decoded at the same time). To manage the different kind of allocation, an enum is added to the rkvdec_aux_buf struct to specify how the buffer was allocated, and so, how to free it. This commit is in preparation of other variants support. Signed-off-by: Detlev Casanova --- .../media/platform/rockchip/rkvdec/Makefile | 1 + .../platform/rockchip/rkvdec/rkvdec-rcb.c | 174 ++++++++++++++++++ .../platform/rockchip/rkvdec/rkvdec-rcb.h | 29 +++ .../media/platform/rockchip/rkvdec/rkvdec.c | 29 ++- .../media/platform/rockchip/rkvdec/rkvdec.h | 13 ++ 5 files changed, 242 insertions(+), 4 deletions(-) create mode 100644 drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.c create mode 100644 drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.h diff --git a/drivers/media/platform/rockchip/rkvdec/Makefile b/drivers/medi= a/platform/rockchip/rkvdec/Makefile index 94c067084acd1..2b7ba9f7de286 100644 --- a/drivers/media/platform/rockchip/rkvdec/Makefile +++ b/drivers/media/platform/rockchip/rkvdec/Makefile @@ -5,4 +5,5 @@ rockchip-vdec-y +=3D \ rkvdec-cabac.o \ rkvdec-h264.o \ rkvdec-h264-common.o \ + rkvdec-rcb.o \ rkvdec-vp9.o diff --git a/drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.c b/drivers/= media/platform/rockchip/rkvdec/rkvdec-rcb.c new file mode 100644 index 0000000000000..210e59db2c622 --- /dev/null +++ b/drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.c @@ -0,0 +1,174 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Rockchip video decoder Rows and Cols Buffers manager + * + * Copyright (C) 2025 Collabora, Ltd. + * Detlev Casanova + */ + +#include "rkvdec.h" +#include "rkvdec-rcb.h" + +#include +#include +#include + +struct rkvdec_rcb_config { + struct rkvdec_aux_buf *rcb_bufs; + size_t rcb_count; +}; + +static size_t rkvdec_rcb_size(const struct rcb_size_info *size_info, + unsigned int width, unsigned int height) +{ + return size_info->multiplier * (size_info->axis =3D=3D PIC_HEIGHT ? heigh= t : width); +} + +dma_addr_t rkvdec_rcb_buf_dma_addr(struct rkvdec_ctx *ctx, int id) +{ + return ctx->rcb_config->rcb_bufs[id].dma; +} + +size_t rkvdec_rcb_buf_size(struct rkvdec_ctx *ctx, int id) +{ + return ctx->rcb_config->rcb_bufs[id].size; +} + +int rkvdec_rcb_buf_count(struct rkvdec_ctx *ctx) +{ + return ctx->rcb_config->rcb_count; +} + +void rkvdec_free_rcb(struct rkvdec_ctx *ctx) +{ + struct rkvdec_dev *dev =3D ctx->dev; + struct rkvdec_rcb_config *cfg =3D ctx->rcb_config; + unsigned long virt_addr; + int i; + + if (!cfg) + return; + + for (i =3D 0; i < cfg->rcb_count; i++) { + size_t rcb_size =3D cfg->rcb_bufs[i].size; + + if (!cfg->rcb_bufs[i].cpu) + continue; + + switch (cfg->rcb_bufs[i].type) { + case RKVDEC2_ALLOC_SRAM: + virt_addr =3D (unsigned long)cfg->rcb_bufs[i].cpu; + + if (dev->iommu_domain) + iommu_unmap(dev->iommu_domain, virt_addr, rcb_size); + gen_pool_free(dev->sram_pool, virt_addr, rcb_size); + break; + case RKVDEC2_ALLOC_DMA: + dma_free_coherent(dev->dev, + rcb_size, + cfg->rcb_bufs[i].cpu, + cfg->rcb_bufs[i].dma); + break; + } + } + + if (cfg->rcb_bufs) + devm_kfree(dev->dev, cfg->rcb_bufs); + + devm_kfree(dev->dev, cfg); +} + +int rkvdec_allocate_rcb(struct rkvdec_ctx *ctx, + const struct rcb_size_info *size_info, + size_t rcb_count) +{ + int ret, i; + u32 width, height; + struct rkvdec_dev *rkvdec =3D ctx->dev; + struct rkvdec_rcb_config *cfg; + + ctx->rcb_config =3D devm_kzalloc(rkvdec->dev, sizeof(*ctx->rcb_config), G= FP_KERNEL); + if (!ctx->rcb_config) + return -ENOMEM; + + cfg =3D ctx->rcb_config; + + cfg->rcb_bufs =3D devm_kzalloc(rkvdec->dev, sizeof(*cfg->rcb_bufs) * rcb_= count, GFP_KERNEL); + if (!cfg->rcb_bufs) { + ret =3D -ENOMEM; + goto err_alloc; + } + + width =3D ctx->decoded_fmt.fmt.pix_mp.width; + height =3D ctx->decoded_fmt.fmt.pix_mp.height; + + for (i =3D 0; i < rcb_count; i++) { + void *cpu =3D NULL; + dma_addr_t dma; + size_t rcb_size =3D rkvdec_rcb_size(&size_info[i], width, height); + enum rkvdec_alloc_type alloc_type =3D RKVDEC2_ALLOC_SRAM; + + /* Try allocating an SRAM buffer */ + if (ctx->dev->sram_pool) { + if (rkvdec->iommu_domain) + rcb_size =3D ALIGN(rcb_size, 0x1000); + + cpu =3D gen_pool_dma_zalloc_align(ctx->dev->sram_pool, + rcb_size, + &dma, + 0x1000); + } + + /* If an IOMMU is used, map the SRAM address through it */ + if (cpu && rkvdec->iommu_domain) { + unsigned long virt_addr =3D (unsigned long)cpu; + phys_addr_t phys_addr =3D dma; + + ret =3D iommu_map(rkvdec->iommu_domain, virt_addr, phys_addr, + rcb_size, IOMMU_READ | IOMMU_WRITE, 0); + if (ret) { + gen_pool_free(ctx->dev->sram_pool, + (unsigned long)cpu, + rcb_size); + cpu =3D NULL; + goto ram_fallback; + } + + /* + * The registers will be configured with the virtual + * address so that it goes through the IOMMU + */ + dma =3D virt_addr; + } + +ram_fallback: + /* Fallback to RAM */ + if (!cpu) { + cpu =3D dma_alloc_coherent(ctx->dev->dev, + rcb_size, + &dma, + GFP_KERNEL); + alloc_type =3D RKVDEC2_ALLOC_DMA; + } + + if (!cpu) { + ret =3D -ENOMEM; + goto err_alloc; + } + + cfg->rcb_bufs[i].cpu =3D cpu; + cfg->rcb_bufs[i].dma =3D dma; + cfg->rcb_bufs[i].size =3D rcb_size; + cfg->rcb_bufs[i].type =3D alloc_type; + + cfg->rcb_count +=3D 1; + } + + return 0; + +err_alloc: + rkvdec_free_rcb(ctx); + + return ret; +} + diff --git a/drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.h b/drivers/= media/platform/rockchip/rkvdec/rkvdec-rcb.h new file mode 100644 index 0000000000000..30e8002555c8a --- /dev/null +++ b/drivers/media/platform/rockchip/rkvdec/rkvdec-rcb.h @@ -0,0 +1,29 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Rockchip video decoder Rows and Cols Buffers manager + * + * Copyright (C) 2025 Collabora, Ltd. + * Detlev Casanova + */ + +#include + +struct rkvdec_ctx; + +enum rcb_axis { + PIC_WIDTH =3D 0, + PIC_HEIGHT =3D 1 +}; + +struct rcb_size_info { + u8 multiplier; + enum rcb_axis axis; +}; + +int rkvdec_allocate_rcb(struct rkvdec_ctx *ctx, + const struct rcb_size_info *size_info, + size_t rcb_count); +dma_addr_t rkvdec_rcb_buf_dma_addr(struct rkvdec_ctx *ctx, int id); +size_t rkvdec_rcb_buf_size(struct rkvdec_ctx *ctx, int id); +int rkvdec_rcb_buf_count(struct rkvdec_ctx *ctx); +void rkvdec_free_rcb(struct rkvdec_ctx *ctx); diff --git a/drivers/media/platform/rockchip/rkvdec/rkvdec.c b/drivers/medi= a/platform/rockchip/rkvdec/rkvdec.c index 28585522523c7..2602676a047a1 100644 --- a/drivers/media/platform/rockchip/rkvdec/rkvdec.c +++ b/drivers/media/platform/rockchip/rkvdec/rkvdec.c @@ -10,6 +10,7 @@ */ =20 #include +#include #include #include #include @@ -27,6 +28,7 @@ =20 #include "rkvdec.h" #include "rkvdec-regs.h" +#include "rkvdec-rcb.h" =20 static bool rkvdec_image_fmt_match(enum rkvdec_image_fmt fmt1, enum rkvdec_image_fmt fmt2) @@ -668,6 +670,7 @@ static int rkvdec_start_streaming(struct vb2_queue *q, = unsigned int count) { struct rkvdec_ctx *ctx =3D vb2_get_drv_priv(q); const struct rkvdec_coded_fmt_desc *desc; + struct rkvdec_config *cfg =3D ctx->dev->config; int ret; =20 if (V4L2_TYPE_IS_CAPTURE(q->type)) @@ -677,13 +680,22 @@ static int rkvdec_start_streaming(struct vb2_queue *q= , unsigned int count) if (WARN_ON(!desc)) return -EINVAL; =20 + ret =3D rkvdec_allocate_rcb(ctx, cfg->rcb_size_info, cfg->rcb_num); + if (ret) + return ret; + if (desc->ops->start) { ret =3D desc->ops->start(ctx); if (ret) - return ret; + goto err_ops_start; } =20 return 0; + +err_ops_start: + rkvdec_free_rcb(ctx); + + return ret; } =20 static void rkvdec_queue_cleanup(struct vb2_queue *vq, u32 state) @@ -719,6 +731,8 @@ static void rkvdec_stop_streaming(struct vb2_queue *q) =20 if (desc->ops->stop) desc->ops->stop(ctx); + + rkvdec_free_rcb(ctx); } =20 rkvdec_queue_cleanup(q, VB2_BUF_STATE_ERROR); @@ -1183,15 +1197,14 @@ static int rkvdec_probe(struct platform_device *pde= v) return ret; } =20 - if (iommu_get_domain_for_dev(&pdev->dev)) { + rkvdec->iommu_domain =3D iommu_get_domain_for_dev(&pdev->dev); + if (rkvdec->iommu_domain) { rkvdec->empty_domain =3D iommu_paging_domain_alloc(rkvdec->dev); =20 if (!rkvdec->empty_domain) dev_warn(rkvdec->dev, "cannot alloc new empty domain\n"); } =20 - vb2_dma_contig_set_max_seg_size(&pdev->dev, DMA_BIT_MASK(32)); - irq =3D platform_get_irq(pdev, 0); if (irq <=3D 0) return -ENXIO; @@ -1204,6 +1217,10 @@ static int rkvdec_probe(struct platform_device *pdev) return ret; } =20 + rkvdec->sram_pool =3D of_gen_pool_get(pdev->dev.of_node, "sram", 0); + if (!rkvdec->sram_pool && rkvdec->config->rcb_num > 0) + dev_info(&pdev->dev, "No sram node, RCB will be stored in RAM\n"); + pm_runtime_set_autosuspend_delay(&pdev->dev, 100); pm_runtime_use_autosuspend(&pdev->dev); pm_runtime_enable(&pdev->dev); @@ -1217,6 +1234,10 @@ static int rkvdec_probe(struct platform_device *pdev) err_disable_runtime_pm: pm_runtime_dont_use_autosuspend(&pdev->dev); pm_runtime_disable(&pdev->dev); + + if (rkvdec->sram_pool) + gen_pool_destroy(rkvdec->sram_pool); + return ret; } =20 diff --git a/drivers/media/platform/rockchip/rkvdec/rkvdec.h b/drivers/medi= a/platform/rockchip/rkvdec/rkvdec.h index 14717c646c665..3a073883d21eb 100644 --- a/drivers/media/platform/rockchip/rkvdec/rkvdec.h +++ b/drivers/media/platform/rockchip/rkvdec/rkvdec.h @@ -19,10 +19,12 @@ #include #include #include +#include #include #include =20 struct rkvdec_ctx; +struct rkvdec_rcb_config; =20 struct rkvdec_ctrl_desc { struct v4l2_ctrl_config cfg; @@ -103,6 +105,8 @@ struct rkvdec_coded_fmt_desc { struct rkvdec_config { struct rkvdec_coded_fmt_desc *coded_fmts; size_t coded_fmts_num; + struct rcb_size_info *rcb_size_info; + size_t rcb_num; }; =20 struct rkvdec_dev { @@ -115,6 +119,8 @@ struct rkvdec_dev { void __iomem *regs; struct mutex vdev_lock; /* serializes ioctls */ struct delayed_work watchdog_work; + struct gen_pool *sram_pool; + struct iommu_domain *iommu_domain; struct iommu_domain *empty_domain; struct rkvdec_config *config; }; @@ -127,6 +133,7 @@ struct rkvdec_ctx { struct v4l2_ctrl_handler ctrl_hdl; struct rkvdec_dev *dev; enum rkvdec_image_fmt image_fmt; + struct rkvdec_rcb_config *rcb_config; void *priv; }; =20 @@ -135,10 +142,16 @@ static inline struct rkvdec_ctx *fh_to_rkvdec_ctx(str= uct v4l2_fh *fh) return container_of(fh, struct rkvdec_ctx, fh); } =20 +enum rkvdec_alloc_type { + RKVDEC2_ALLOC_DMA =3D 0, + RKVDEC2_ALLOC_SRAM =3D 1, +}; + struct rkvdec_aux_buf { void *cpu; dma_addr_t dma; size_t size; + enum rkvdec_alloc_type type; }; =20 void rkvdec_run_preamble(struct rkvdec_ctx *ctx, struct rkvdec_run *run); --=20 2.50.1