Re: [PATCH v6 30/30] mtd: rawnand: Allocate the best interface configuration dynamically

[Miquel Raynal <miquel.raynal@xxxxxxxxxxx>]

Boris Brezillon <boris.brezillon@xxxxxxxxxxxxx> wrote on Fri, 29 May
2020 10:54:31 +0200:

> On Fri, 29 May 2020 01:16:12 +0200
> Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote:
> 
> >  
> > @@ -947,7 +949,8 @@ static int nand_reset_interface(struct nand_chip *chip, int chipnr)
> >   */
> >  static int nand_setup_interface(struct nand_chip *chip, int chipnr)
> >  {
> > -	u8 mode = chip->interface_config.timings.mode;
> > +	const struct nand_controller_ops *ops = chip->controller->ops;
> > +	u8 mode = chip->best_interface_config->timings.mode;  
> 
> I didn't check, but I assume it's safe to dereference
> ->best_interface_config here (IOW, ->best_interface_config is  
> guaranteed to be != NULL).

Yes it's safe. ->best_interface_config is always set, even with mode 0
configuration.

> 
> >  	u8 tmode_param[ONFI_SUBFEATURE_PARAM_LEN] = { mode, };
> >  	int ret;
> >    
> 
> >  
> >  /**
> > @@ -2523,8 +2542,8 @@ int nand_reset(struct nand_chip *chip, int chipnr)
> >  	 * nand_setup_interface() uses ->set/get_features() which would
> >  	 * fail anyway as the parameter page is not available yet.
> >  	 */
> > -	if (!memcmp(&chip->interface_config, nand_get_reset_interface(),
> > -		    sizeof(chip->interface_config)))
> > +	if (!chip->best_interface_config ||
> > +	    chip->best_interface_config == nand_get_reset_interface())  
> 
> I think 'chip->best_interface_config == nand_get_reset_interface()' is
> never true.

Oh right. I'll drop this, and we'll accept to sometimes call
nand_setup_interface() to apply timings mode 0 while they are already
applied. I don't think it's a big deal.


______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/