[U-Boot] [PATCH 3/5] Add fuse API and commands

[Benoît Thébaudeau]

Hi Dirk,

On Wednesday, August 22, 2012 6:25:57 PM, Dirk Behme wrote:
> On 22.08.2012 13:11, Benoît Thébaudeau wrote:
> > Hi Dirk,
> >
> > On Wednesday, August 22, 2012 12:43:05 PM, Dirk Behme wrote:
> >> On 14.08.2012 14:52, Benoît Thébaudeau wrote:
> >>> This can be useful for fuse-like hardware, OTP SoC options, etc.
> >>
> >> For i.MX6, I have a port of the OTP support from Freescale's
> >> U-Boot
> >> to
> >> our mainline U-Boot in the queue [1].
> >>
> >> As I don't have the overview over the various i.MXxx SoCs and
> >> don't
> >> understand much of this patch below: Should this implement the
> >> same
> >> functionality like my patch [1] for i.MX6?
> >>
> >> Or shall I send my patch [1] to this mailing list for official
> >> review
> >> because the functionality here and there is orthogonal?
> >>
> >> Thanks
> >>
> >> Dirk
> >>
> >> [1]
> >> https://github.com/dirkbehme/u-boot-imx6/commit/da718b338a79af160f7b7e542fe97b24edbfc36a
> >
> > This OTP IP is different from the IIM IP of other i.MXs, so having
> > a different
> > driver for it is fine.
> >
> > What you can do is implement in your driver the fuse API
> 
> You mean
> 
> http://lists.denx.de/pipermail/u-boot/2012-August/130904.html
> 
> here, correct?

Yes.

> >that I've defined in
> > this series. In that way, you could drop your common/cmd_imxotp.c
> > and use the
> > commands from this series.
> 
> Let's see how this fits to i.MX6:
> 
> My understanding is that the i.MX6 has 128 fuse 'registers' (#define
> IMX_OTP_ADDR_MAX 0x7F), each containing 32 fuses/bits. These
> fuses/bits can be written from 0 -> 1 once.
> 
> Looking at your API [2] we could set 'bank' to 0 and interpret 'row'
> as the 'register' number. That might fit. Doing this,
> fuse_read_bit/row() and fuse_prog_bit/row() could be used for i.MX6.

Sounds good.

>  From i.MX6 point of view, as I don't know your use case, I'm not
>  sure
> what fuse_sense_bit/row() and fuse_override_bit/row() might be good
> for?

These are hardware features on the other i.MXs' IIM that may exist on other
hardware as well.

Override allows to change the value read from fuses without reprogramming the
fuses. This change is lost after a power cycle. This can be useful e.g. to
change some hardware behavior linked to values read from fuses.

Sense allows to get the real programmed value of a fuse without being disturbed
by a pending override on this fuse.

I don't know if i.MX6's OTP has such features. If your driver is complete, i.MX6
doesn't have these features, so the corresponding functions should simply return
an error.

> Anybody else: Does this make sense? Or should we keep the i.MX6
> specific common/cmd_imxotp.c (see [1] above)?.
> 
> Best regards
> 
> Dirk
> 
> [2]
> 
> + * Read/Sense/Program/Override interface:
> + *   bank:    Fuse bank
> + *   row:     Fuse row within the bank
> + *   bit:     Fuse bit within the row
> + *   val:     Value to read/write
> + *
> + *   Returns: 0 on success, not 0 on failure
> + */
> +int fuse_read_bit(u32 bank, u32 row, u32 bit, u32 *val);
> +int fuse_read_row(u32 bank, u32 row, u32 *val);
> +int fuse_sense_bit(u32 bank, u32 row, u32 bit, u32 *val);
> +int fuse_sense_row(u32 bank, u32 row, u32 *val);
> +int fuse_prog_bit(u32 bank, u32 row, u32 bit);
> +int fuse_prog_row(u32 bank, u32 row, u32 val);
> +int fuse_override_bit(u32 bank, u32 row, u32 bit, u32 val);
> +int fuse_override_row(u32 bank, u32 row, u32 val);

Best regards,
Benoît