Re: [PATCH v7 4/5] nvmem: eeprom: at25: export FRAM serial num

[Greg Kroah-Hartman]

On Tue, Jun 08, 2021 at 11:45:56AM +0200, Jiří Prchal wrote:
> 
> 
> On 08. 06. 21 11:03, Greg Kroah-Hartman wrote:
> > On Mon, Jun 07, 2021 at 04:47:44PM +0200, Jiří Prchal wrote:
> > > 
> > > 
> > > On 07. 06. 21 14:36, Greg Kroah-Hartman wrote:
> > > > On Mon, Jun 07, 2021 at 02:26:39PM +0200, Jiri Prchal wrote:
> > > > > +	return sysfs_emit(buf, "%016llx\n", *(unsigned long long *)at25->sernum);
> > > > 
> > > > That's a horrid hack, why not use the %*phN modifier?
> > > 
> > > Prints as little endian, is that OK?
> > 
> > You tell me!  What tool is going to be reading this?  What do they
> > expect it to look like?
> 
> sh, php in my usecase as unique id.

I am sorry, I do not understand.

> So endianess does not matter to me too much. The question is what is usual
> (like mac address, uuid...?).

What does the device export?  Why not just export it as:
	0123456789ABCDEF
if it is 8 bytes long?

> > And it's a byte array, why would there be endian issues?
> 
> Now is printed as one big number. Not real issue. Just human readability?
> Should I turn back it to space separated bytes?

It's up to you, what do you want to do with it and what does a tool want
it to look like?

thanks,

greg k-h