mtd: spinand: add spi nand mtd resume handler

Miquel Raynal miquel.raynal at bootlin.com
Thu May 27 03:04:23 PDT 2021


Hi Pratyush,

Pratyush Yadav <p.yadav at ti.com> wrote on Thu, 27 May 2021 15:30:17
+0530:

> On 26/05/21 05:30PM, patrice.chotard at foss.st.com wrote:
> > From: Christophe Kerello <christophe.kerello at foss.st.com>
> > 
> > After power up, all SPI NAND's blocks are locked. Only read operations
> > are allowed, write and erase operations are forbidden.
> > The SPI NAND framework unlocks all the blocks during its initialization.
> > 
> > During a standby low power, the memory is powered down, losing its
> > configuration.
> > During the resume, the QSPI driver state is restored but the SPI NAND
> > framework does not reconfigured the memory.
> > 
> > This patch adds spi nand mtd PM handlers for resume ops.
> > SPI NAND resume op re-initializes SPI NAND flash to its probed state.
> > 
> > Signed-off-by: Christophe Kerello <christophe.kerello at foss.st.com>
> > Signed-off-by: Patrice Chotard <patrice.chotard at foss.st.com>
> > ---
> >  drivers/mtd/nand/spi/core.c | 56 +++++++++++++++++++++++++++++++++++++
> >  1 file changed, 56 insertions(+)
> > 
> > diff --git a/drivers/mtd/nand/spi/core.c b/drivers/mtd/nand/spi/core.c
> > index 17f63f95f4a2..6abaf874eb3f 100644
> > --- a/drivers/mtd/nand/spi/core.c
> > +++ b/drivers/mtd/nand/spi/core.c
> > @@ -1074,6 +1074,61 @@ static int spinand_detect(struct spinand_device *spinand)
> >  	return 0;
> >  }
> >  
> > +static void spinand_mtd_resume(struct mtd_info *mtd)
> > +{
> > +	struct spinand_device *spinand = mtd_to_spinand(mtd);
> > +	struct nand_device *nand = mtd_to_nanddev(mtd);
> > +	struct device *dev = &spinand->spimem->spi->dev;
> > +	int ret, i;
> > +
> > +	ret = spinand_reset_op(spinand);
> > +	if (ret)
> > +		return;
> > +
> > +	ret = spinand_init_quad_enable(spinand);
> > +	if (ret) {
> > +		dev_err(dev,
> > +			"Failed to initialize the quad part (err = %d)\n",
> > +			ret);
> > +		return;
> > +	}
> > +
> > +	ret = spinand_upd_cfg(spinand, CFG_OTP_ENABLE, 0);
> > +	if (ret) {
> > +		dev_err(dev,
> > +			"Failed to updtae the OTP (err = %d)\n",
> > +			ret);
> > +		return;
> > +	}  
> 
> Since you have reset the flash, this cache is invalid. You should reset 
> the cache and re-populate it before using it in any way.
> 
> > +
> > +	ret = spinand_manufacturer_init(spinand);
> > +	if (ret) {
> > +		dev_err(dev,
> > +			"Failed to initialize the SPI NAND chip (err = %d)\n",
> > +			ret);
> > +		return;
> > +	}
> > +
> > +	/* After power up, all blocks are locked, so unlock them here. */
> > +	for (i = 0; i < nand->memorg.ntargets; i++) {
> > +		ret = spinand_select_target(spinand, i);
> > +		if (ret) {
> > +			dev_err(dev,
> > +				"Failed to select the target (err = %d)\n",
> > +				ret);
> > +			return;
> > +		}
> > +
> > +		ret = spinand_lock_block(spinand, BL_ALL_UNLOCKED);
> > +		if (ret) {
> > +			dev_err(dev,
> > +				"Failed to unlock block (err = %d)\n",
> > +				ret);
> > +			return;
> > +		}
> > +	}
> > +}
> > +  
> 
> Most of these seem to be copied from spinand_init(). I think it is 
> better to create a common function that can be called from both 
> spinand_init() and spinand_mtd_resume(). This way when someone adds 
> something new to the init procedure, like support for some other modes, 
> they won't have to remember to update it in two places.

Agreed, let's write a common helper for more than just the unlocking
sequence (still in a separate patch).

> 
> >  static int spinand_init(struct spinand_device *spinand)
> >  {
> >  	struct device *dev = &spinand->spimem->spi->dev;
> > @@ -1167,6 +1222,7 @@ static int spinand_init(struct spinand_device *spinand)
> >  	mtd->_block_isreserved = spinand_mtd_block_isreserved;
> >  	mtd->_erase = spinand_mtd_erase;
> >  	mtd->_max_bad_blocks = nanddev_mtd_max_bad_blocks;
> > +	mtd->_resume = spinand_mtd_resume;  
> 
> Is it possible that the userspace can use this mtd device before the 
> resume is finished? Is there a way to temporarily "pause" or unregister 
> an mtd device?

I don't expect this to happen, I would expect the kernel to resume
entirely before giving the hand to userspace, but I am not 100% sure of
that neither.

> 
> >  
> >  	if (nand->ecc.engine) {
> >  		ret = mtd_ooblayout_count_freebytes(mtd);  
> 

Thanks,
Miquèl



More information about the linux-arm-kernel mailing list