Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752486AbdGCGdJ (ORCPT ); Mon, 3 Jul 2017 02:33:09 -0400 Received: from smtp.codeaurora.org ([198.145.29.96]:49972 "EHLO smtp.codeaurora.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751902AbdGCGdH (ORCPT ); Mon, 3 Jul 2017 02:33:07 -0400 DMARC-Filter: OpenDMARC Filter v1.3.2 smtp.codeaurora.org 477456031A 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=sricharan@codeaurora.org Subject: Re: [PATCH 02/14] qcom: mtd: nand: add and initialize QPIC DMA resources To: Abhishek Sahu , dwmw2@infradead.org, computersforpeace@gmail.com, boris.brezillon@free-electrons.com, marek.vasut@gmail.com, richard@nod.at, cyrille.pitchen@wedev4u.fr, robh+dt@kernel.org, mark.rutland@arm.com Cc: linux-mtd@lists.infradead.org, devicetree@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org, andy.gross@linaro.org, architt@codeaurora.org References: <1498720566-20782-1-git-send-email-absahu@codeaurora.org> <1498720566-20782-3-git-send-email-absahu@codeaurora.org> From: Sricharan R Message-ID: Date: Mon, 3 Jul 2017 12:02:58 +0530 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.2.1 MIME-Version: 1.0 In-Reply-To: <1498720566-20782-3-git-send-email-absahu@codeaurora.org> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit X-Antivirus: Avast (VPS 170702-0, 07/02/2017), Outbound message X-Antivirus-Status: Clean Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5365 Lines: 152 Hi Abhishek, On 6/29/2017 12:45 PM, Abhishek Sahu wrote: > 1. The QPIC NAND uses 3 BAM channels: command, data tx and > data rx while EBI2 NAND uses only single ADM channel. > > 2. The EBI2 NAND uses normal register read buffer since this > buffer will be remapped with dma_map_sg. The QPIC NAND will give > register read buffer in command descriptor and the command > descriptor will be mapped with dma_map_sg so the register buffer > should be DMA coherent. > > Signed-off-by: Abhishek Sahu > --- > .../devicetree/bindings/mtd/qcom_nandc.txt | 25 +++-- > drivers/mtd/nand/qcom_nandc.c | 106 ++++++++++++++++----- > 2 files changed, 99 insertions(+), 32 deletions(-) > > diff --git a/Documentation/devicetree/bindings/mtd/qcom_nandc.txt b/Documentation/devicetree/bindings/mtd/qcom_nandc.txt > index 5d0f7ae..87b9a56 100644 > --- a/Documentation/devicetree/bindings/mtd/qcom_nandc.txt > +++ b/Documentation/devicetree/bindings/mtd/qcom_nandc.txt > @@ -9,15 +9,17 @@ Required properties: > - clock-names: must contain "core" for the core clock and "aon" for the > always on clock > - dmas: DMA specifier, consisting of a phandle to the ADM DMA > - controller node and the channel number to be used for > - NAND. Refer to dma.txt and qcom_adm.txt for more details > -- dma-names: must be "rxtx" > -- qcom,cmd-crci: must contain the ADM command type CRCI block instance > - number specified for the NAND controller on the given > - platform > -- qcom,data-crci: must contain the ADM data type CRCI block instance > - number specified for the NAND controller on the given > - platform > + or BAM DMA controller node and the channel number to > + be used for NAND. Refer to dma.txt, qcom_adm.txt(ADM) > + and qcom/bam_dma.txt(BAM) for more details > +- dma-names: "rxtx" - ADM > + "tx", "rx", "cmd" - BAM > +- qcom,cmd-crci: Only required for ADM DMA. must contain the ADM command > + type CRCI block instance number specified for the NAND > + controller on the given platform. > +- qcom,data-crci: Only required for ADM DMA. must contain the ADM data > + type CRCI block instance number specified for the NAND > + controller on the given platform. > - #address-cells: <1> - subnodes give the chip-select number > - #size-cells: <0> > > @@ -95,6 +97,11 @@ nand@79b0000 { > <&gcc GCC_QPIC_AHB_CLK>; > clock-names = "core", "aon"; > > + dmas = <&qpicbam 0>, > + <&qpicbam 1>, > + <&qpicbam 2>; > + dma-names = "tx", "rx", "cmd"; > + > #address-cells = <1>; > #size-cells = <0>; > > diff --git a/drivers/mtd/nand/qcom_nandc.c b/drivers/mtd/nand/qcom_nandc.c > index f55f728..520add9 100644 > --- a/drivers/mtd/nand/qcom_nandc.c > +++ b/drivers/mtd/nand/qcom_nandc.c > @@ -226,6 +226,7 @@ struct nandc_regs { > * by upper layers directly > * @buf_size/count/start: markers for chip->read_buf/write_buf functions > * @reg_read_buf: local buffer for reading back registers via DMA > + * @reg_read_buf_phys: contains dma address for register read buffer > * @reg_read_pos: marker for data read in reg_read_buf > * > * @regs: a contiguous chunk of memory for DMA register > @@ -249,9 +250,19 @@ struct qcom_nand_controller { > struct clk *core_clk; > struct clk *aon_clk; > > - struct dma_chan *chan; > - unsigned int cmd_crci; > - unsigned int data_crci; > + union { > + struct { > + struct dma_chan *tx_chan; > + struct dma_chan *rx_chan; > + struct dma_chan *cmd_chan; > + }; > + struct { > + struct dma_chan *chan; > + unsigned int cmd_crci; > + unsigned int data_crci; > + }; > + }; > + > struct list_head desc_list; > > u8 *data_buffer; > @@ -261,6 +272,7 @@ struct qcom_nand_controller { > int buf_start; > > __le32 *reg_read_buf; > + dma_addr_t reg_read_buf_phys; > int reg_read_pos; > > struct nandc_regs *regs; > @@ -1956,16 +1968,48 @@ static int qcom_nandc_alloc(struct qcom_nand_controller *nandc) > if (!nandc->regs) > return -ENOMEM; > > - nandc->reg_read_buf = devm_kzalloc(nandc->dev, > - MAX_REG_RD * sizeof(*nandc->reg_read_buf), > - GFP_KERNEL); > - if (!nandc->reg_read_buf) > - return -ENOMEM; > > - nandc->chan = dma_request_slave_channel(nandc->dev, "rxtx"); > - if (!nandc->chan) { > - dev_err(nandc->dev, "failed to request slave channel\n"); > - return -ENODEV; > + if (!nandc->dma_bam_enabled) { > + nandc->reg_read_buf = > + devm_kzalloc(nandc->dev, MAX_REG_RD * > + sizeof(*nandc->reg_read_buf), GFP_KERNEL); > + > + if (!nandc->reg_read_buf) > + return -ENOMEM; > + > + nandc->chan = dma_request_slave_channel(nandc->dev, "rxtx"); > + if (!nandc->chan) { > + dev_err(nandc->dev, > + "failed to request slave channel\n"); > + return -ENODEV; > + } > + } else { > + nandc->reg_read_buf = > + dmam_alloc_coherent(nandc->dev, MAX_REG_RD * > + sizeof(*nandc->reg_read_buf), > + &nandc->reg_read_buf_phys, > + GFP_KERNEL); > + I think as Archit has commented, do not see a reason for this to be alloc_coherent change here. Regards, Sricharan -- "QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus