[RFC] MTD: export more NAND-related informations to syfs

Ricard Wanderlof ricard.wanderlof at axis.com
Thu May 12 06:38:34 EDT 2011


> On Thu, 12 May 2011, Artem Bityutskiy wrote:
> 
> > On Mon, 2011-05-09 at 16:21 +0200, Florian Fainelli wrote:
> > Hello Artem,
> > 
> > On Friday 06 May 2011 20:39:33 Artem Bityutskiy wrote:
> > > On Fri, 2011-05-06 at 17:28 +0200, Florian Fainelli wrote:
> > > > +       /* NAND related attributes */
> > > > +       const char *nand_type;
> > > > +       const char *nand_manufacturer;
> > > 
> > > Why other flashes cannot have type and manufacturer ?
> > 
> > You are right, I will add this for NOR and SPI Flashes. My RFC was more about 
> > whether it is granted to add new members to struct mtd_info to get 
> > informations 
> 
> struct mtd_info describes an mtd device. Things, which are common for
> all mtd devices should be there, I think.
> 
> > Alternatively, would not it be better to have something like this:
> > 
> > /sys/class/mtd/...
> >    nand0/
> >        manufacturer
> >        type
> >        onfi/
> >           version
> >           jedec_id
> >           ...
> >     nor0/
> >        cfi/
> >           manufacturer
> >           id
> > 
> > This would allow us not to abuse the sysfs directory of a MTD partition to 
> > contain chip specific informations.
> 
> Why do you call it "abuse" and why is it better to "abuse" the directory
> common to all MTD devices comparing to the directory of the specific MTD
> device? :-)
> 
> I think that everything related to mtdX has to be in the mtdX
> subdirectory.

Couldn't one have a link to some form of chip directory structure, that 
seems to be a common way of dealing with multiple sys subdirectories that 
have something else in common than the directory path.

/Ricard
-- 
Ricard Wolf Wanderlöf                           ricardw(at)axis.com
Axis Communications AB, Lund, Sweden            www.axis.com
Phone +46 46 272 2016                           Fax +46 46 13 61 30



More information about the linux-mtd mailing list