linux-kernel - Re: [PATCH v2 03/14] mtd: spinand: Patch spi_mem_op for the SPI IO protocol using reg

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: <20211012084047.6381ddce@collabora.com>
Date:   Tue, 12 Oct 2021 08:40:47 +0200
From:   Boris Brezillon <boris.brezillon@...labora.com>
To:     Apurva Nandan <a-nandan@...com>
Cc:     Miquel Raynal <miquel.raynal@...tlin.com>,
        Richard Weinberger <richard@....at>,
        Vignesh Raghavendra <vigneshr@...com>,
        Mark Brown <broonie@...nel.org>,
        Patrice Chotard <patrice.chotard@...s.st.com>,
        Christophe Kerello <christophe.kerello@...s.st.com>,
        <linux-mtd@...ts.infradead.org>, <linux-kernel@...r.kernel.org>,
        <linux-spi@...r.kernel.org>, <p.yadav@...com>
Subject: Re: [PATCH v2 03/14] mtd: spinand: Patch spi_mem_op for the SPI IO
 protocol using reg_proto

On Tue, 12 Oct 2021 02:16:08 +0530
Apurva Nandan <a-nandan@...com> wrote:

> Currently, the op macros in spinand.h don't give the option to setup
> any non-array access instructions for Dual/Quad/Octal DTR SPI bus.
> Having a function that patches the op based on reg_proto would be
> better than trying to write all the setup logic in op macros.
> 
> Create a spimem_patch_op() that would patch cmd, addr, dummy and data
> phase of the spi_mem op, for the given spinand->reg_proto. And hence,
> call the spimem_patch_op() before executing any spi_mem op.

I'm honestly not sure this rule will stand over time, and patching
operations at submission time has a cost (ok, it's negligible compared
to the IO duration, but still), so I'd rather see this done statically
with all relevant macros taking an extra dtr argument.

> 
> Note: In this commit, spimem_patch_op() isn't called in the
> read_reg_op(), write_reg_op() and wait() functions, as they need
> modifications in address value and data nbytes when in Octal DTR mode.
> This will be fixed in a later commit.

See, that's what I'm talking about.

> 
> Signed-off-by: Apurva Nandan <a-nandan@...com>
> ---
>  drivers/mtd/nand/spi/core.c | 49 +++++++++++++++++++++++++++++++++++++
>  1 file changed, 49 insertions(+)
> 
> diff --git a/drivers/mtd/nand/spi/core.c b/drivers/mtd/nand/spi/core.c
> index d82a3e6d9bb5..11746d858f87 100644
> --- a/drivers/mtd/nand/spi/core.c
> +++ b/drivers/mtd/nand/spi/core.c
> @@ -20,6 +20,49 @@
>  #include <linux/spi/spi.h>
>  #include <linux/spi/spi-mem.h>
>  
> +/**
> + * spinand_patch_op() - Helper function to patch the spi_mem op based on the
> + *			spinand->reg_proto
> + * @spinand: the spinand device
> + * @op: the spi_mem op to patch
> + *
> + * Set up buswidth and dtr fields for cmd, addr, dummy and data phase. Also
> + * adjust cmd opcode and dummy nbytes. This function doesn't make any changes
> + * to addr val or data buf.
> + */
> +static void spinand_patch_op(const struct spinand_device *spinand,
> +			     struct spi_mem_op *op)
> +{
> +	u8 op_buswidth = SPINAND_PROTO_BUSWIDTH(spinand->reg_proto);
> +	u8 op_is_dtr = SPINAND_PROTO_IS_DTR(spinand->reg_proto);
> +
> +	if (spinand->reg_proto == SPINAND_SINGLE_STR)
> +		return;
> +
> +	op->cmd.buswidth = op_buswidth;
> +	op->cmd.dtr = op_is_dtr;
> +	if (spinand->reg_proto == SPINAND_OCTAL_DTR) {
> +		op->cmd.opcode = (op->cmd.opcode << 8) | op->cmd.opcode;
> +		op->cmd.nbytes = 2;
> +	}
> +
> +	if (op->addr.nbytes) {
> +		op->addr.buswidth = op_buswidth;
> +		op->addr.dtr = op_is_dtr;
> +	}
> +
> +	if (op->dummy.nbytes) {
> +		op->dummy.buswidth = op_buswidth;
> +		op->dummy.dtr = op_is_dtr;
> +		op->dummy.nbytes <<= op_is_dtr;
> +	}
> +
> +	if (op->data.nbytes) {
> +		op->data.buswidth = op_buswidth;
> +		op->data.dtr = op_is_dtr;
> +	}
> +}
> +
>  static int spinand_read_reg_op(struct spinand_device *spinand, u8 reg, u8 *val)
>  {
>  	struct spi_mem_op op = SPINAND_GET_FEATURE_OP(reg,
> @@ -343,6 +386,7 @@ static int spinand_write_enable_op(struct spinand_device *spinand)
>  {
>  	struct spi_mem_op op = SPINAND_WR_EN_DIS_OP(true);
>  
> +	spinand_patch_op(spinand, &op);
>  	return spi_mem_exec_op(spinand->spimem, &op);
>  }
>  
> @@ -353,6 +397,7 @@ static int spinand_load_page_op(struct spinand_device *spinand,
>  	unsigned int row = nanddev_pos_to_row(nand, &req->pos);
>  	struct spi_mem_op op = SPINAND_PAGE_READ_OP(row);
>  
> +	spinand_patch_op(spinand, &op);
>  	return spi_mem_exec_op(spinand->spimem, &op);
>  }
>  
> @@ -477,6 +522,7 @@ static int spinand_program_op(struct spinand_device *spinand,
>  	unsigned int row = nanddev_pos_to_row(nand, &req->pos);
>  	struct spi_mem_op op = SPINAND_PROG_EXEC_OP(row);
>  
> +	spinand_patch_op(spinand, &op);
>  	return spi_mem_exec_op(spinand->spimem, &op);
>  }
>  
> @@ -487,6 +533,7 @@ static int spinand_erase_op(struct spinand_device *spinand,
>  	unsigned int row = nanddev_pos_to_row(nand, pos);
>  	struct spi_mem_op op = SPINAND_BLK_ERASE_OP(row);
>  
> +	spinand_patch_op(spinand, &op);
>  	return spi_mem_exec_op(spinand->spimem, &op);
>  }
>  
> @@ -533,6 +580,7 @@ static int spinand_read_id_op(struct spinand_device *spinand, u8 naddr,
>  		naddr, ndummy, spinand->scratchbuf, SPINAND_MAX_ID_LEN);
>  	int ret;
>  
> +	spinand_patch_op(spinand, &op);
>  	ret = spi_mem_exec_op(spinand->spimem, &op);
>  	if (!ret)
>  		memcpy(buf, spinand->scratchbuf, SPINAND_MAX_ID_LEN);
> @@ -545,6 +593,7 @@ static int spinand_reset_op(struct spinand_device *spinand)
>  	struct spi_mem_op op = SPINAND_RESET_OP;
>  	int ret;
>  
> +	spinand_patch_op(spinand, &op);
>  	ret = spi_mem_exec_op(spinand->spimem, &op);
>  	if (ret)
>  		return ret;