linux-kernel - Re: [PATCH 2/3] mtd: rawnand: qcom: Add sg list to handle status pipe request

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210928160315.GC12183@thinkpad>
Date:   Tue, 28 Sep 2021 21:33:15 +0530
From:   Manivannan Sadhasivam <mani@...nel.org>
To:     Md Sadre Alam <mdalam@...eaurora.org>
Cc:     miquel.raynal@...tlin.com, linux-mtd@...ts.infradead.org,
        linux-kernel@...r.kernel.org, sricharan@...eaurora.org
Subject: Re: [PATCH 2/3] mtd: rawnand: qcom: Add sg list to handle status
 pipe request

On Wed, Sep 15, 2021 at 03:27:30PM +0530, Md Sadre Alam wrote:
> From QPIC V2.0 onwards there is separate pipe to read status
> for each code word while reading in enhanced mode. page scope
> read and multi page read.
> 
> This sgl list will be use to handle the request via status pipe
> during page scope and multi page read.
> 
> Signed-off-by: Md Sadre Alam <mdalam@...eaurora.org>
> ---
>  drivers/mtd/nand/raw/qcom_nandc.c | 34 ++++++++++++++++++++++++++++++++++
>  1 file changed, 34 insertions(+)
> 
> diff --git a/drivers/mtd/nand/raw/qcom_nandc.c b/drivers/mtd/nand/raw/qcom_nandc.c
> index 42c6291..07448c4 100644
> --- a/drivers/mtd/nand/raw/qcom_nandc.c
> +++ b/drivers/mtd/nand/raw/qcom_nandc.c
> @@ -213,6 +213,7 @@ nandc_set_reg(chip, reg,			\
>  #define QPIC_PER_CW_CMD_ELEMENTS	32
>  #define QPIC_PER_CW_CMD_SGL		32
>  #define QPIC_PER_CW_DATA_SGL		8
> +#define	QPIC_PER_CW_STS_SGL		8

Use space after #define

>  
>  #define QPIC_NAND_COMPLETION_TIMEOUT	msecs_to_jiffies(2000)
>  
> @@ -258,6 +259,7 @@ struct bam_transaction {
>  	struct bam_cmd_element *bam_ce;
>  	struct scatterlist *cmd_sgl;
>  	struct scatterlist *data_sgl;
> +	struct scatterlist *sts_sgl;
>  	u32 bam_ce_pos;
>  	u32 bam_ce_start;
>  	u32 cmd_sgl_pos;
> @@ -266,6 +268,8 @@ struct bam_transaction {
>  	u32 tx_sgl_start;
>  	u32 rx_sgl_pos;
>  	u32 rx_sgl_start;
> +	u32 sts_sgl_pos;
> +	u32 sts_sgl_start;
>  	bool wait_second_completion;
>  	struct completion txn_done;
>  	struct dma_async_tx_descriptor *last_data_desc;
> @@ -508,6 +512,8 @@ alloc_bam_transaction(struct qcom_nand_controller *nandc)
>  		((sizeof(*bam_txn->bam_ce) * QPIC_PER_CW_CMD_ELEMENTS) +
>  		(sizeof(*bam_txn->cmd_sgl) * QPIC_PER_CW_CMD_SGL) +
>  		(sizeof(*bam_txn->data_sgl) * QPIC_PER_CW_DATA_SGL));
> +	if (nandc->props->qpic_v2)
> +		bam_txn_size += (sizeof(*bam_txn->sts_sgl) * QPIC_PER_CW_STS_SGL);
>  
>  	bam_txn_buf = devm_kzalloc(nandc->dev, bam_txn_size, GFP_KERNEL);
>  	if (!bam_txn_buf)
> @@ -526,6 +532,12 @@ alloc_bam_transaction(struct qcom_nand_controller *nandc)
>  
>  	bam_txn->data_sgl = bam_txn_buf;
>  
> +	if (nandc->props->qpic_v2) {
> +		bam_txn_buf +=
> +			sizeof(*bam_txn->sts_sgl) * QPIC_PER_CW_STS_SGL * num_cw;
> +		bam_txn->sts_sgl = bam_txn_buf;
> +	}
> +
>  	init_completion(&bam_txn->txn_done);
>  
>  	return bam_txn;
> @@ -554,6 +566,12 @@ static void clear_bam_transaction(struct qcom_nand_controller *nandc)
>  		      QPIC_PER_CW_CMD_SGL);
>  	sg_init_table(bam_txn->data_sgl, nandc->max_cwperpage *
>  		      QPIC_PER_CW_DATA_SGL);
> +	if (nandc->props->qpic_v2) {
> +		bam_txn->sts_sgl_pos = 0;
> +		bam_txn->sts_sgl_start = 0;
> +		sg_init_table(bam_txn->sts_sgl, nandc->max_cwperpage *
> +				QPIC_PER_CW_STS_SGL);
> +	}
>  
>  	reinit_completion(&bam_txn->txn_done);
>  }
> @@ -808,6 +826,12 @@ static int prepare_bam_async_desc(struct qcom_nand_controller *nandc,
>  		bam_txn->tx_sgl_start = bam_txn->tx_sgl_pos;
>  		dir_eng = DMA_MEM_TO_DEV;
>  		desc->dir = DMA_TO_DEVICE;
> +	} else if (nandc->props->qpic_v2 && chan == nandc->sts_chan) {

These two are mutually inclusive, so why can't you simply check for the
existence of "nanc->sts_chan"?

> +		sgl = &bam_txn->sts_sgl[bam_txn->sts_sgl_start];
> +		sgl_cnt = bam_txn->sts_sgl_pos - bam_txn->sts_sgl_start;
> +		bam_txn->sts_sgl_start = bam_txn->sts_sgl_pos;
> +		dir_eng = DMA_DEV_TO_MEM;
> +		desc->dir = DMA_FROM_DEVICE;
>  	} else {
>  		sgl = &bam_txn->data_sgl[bam_txn->rx_sgl_start];
>  		sgl_cnt = bam_txn->rx_sgl_pos - bam_txn->rx_sgl_start;
> @@ -1394,6 +1418,14 @@ static int submit_descs(struct qcom_nand_controller *nandc)
>  			if (r)
>  				return r;
>  		}
> +
> +		if (nandc->props->qpic_v2) {

I feel like you should just check for "nandc->sts_chan" instead of qpic_v2. This
will make the driver work with future revisions of the IP as well.

> +			if (bam_txn->sts_sgl_pos > bam_txn->sts_sgl_start) {
> +				r = prepare_bam_async_desc(nandc, nandc->sts_chan, 0);
> +				if (r)
> +					return r;
> +			}
> +		}
>  	}
>  
>  	list_for_each_entry(desc, &nandc->desc_list, node)
> @@ -1411,6 +1443,8 @@ static int submit_descs(struct qcom_nand_controller *nandc)
>  		dma_async_issue_pending(nandc->tx_chan);
>  		dma_async_issue_pending(nandc->rx_chan);
>  		dma_async_issue_pending(nandc->cmd_chan);
> +		if (nandc->props->qpic_v2)
> +			dma_async_issue_pending(nandc->sts_chan);

Same here.

Thanks,
Mani

>  
>  		if (!wait_for_completion_timeout(&bam_txn->txn_done,
>  						 QPIC_NAND_COMPLETION_TIMEOUT))
> -- 
> 2.7.4
>