Am 2020-02-10 12:27, schrieb Tudor.Ambarus@xxxxxxxxxxxxx:
On Monday, February 10, 2020 12:40:59 PM EET Michael Walle wrote:
EXTERNAL EMAIL: Do not click links or open attachments unless you know
the
content is safe
Am 2020-02-10 10:59, schrieb Tudor.Ambarus@xxxxxxxxxxxxx:
> On Monday, February 10, 2020 11:47:23 AM EET Tudor Ambarus wrote:
>> > btw. we should catch the two special cases:
>> > - lock none -> 0 (that was already the case)
>> > - lock all -> all BP bits
>> >
>> > The latter is important if "bp_slots_needed < bp_slots_available"
>> > because there
>> > are multiple settings for protect all. Most flashes will define any
>> > remaining
>> > setting for "protect all", but I've also seen flashes where the
>> > in-between ones
>> > were undefined (not mentioned) and only the "all bit set" was protect
>> > all.
>
> I re-read this. Do you have such an example of flash? Aren't the BP
> bits in
> this case marked with "X", i.e. don't care? If not, probably we can
> mask out
> those undefined slots.
There was definetly some datasheet where the remaining ones wasn't
described
_and_ there was no X (don't care) bits. Unfortunately, I don't find it
right
now. That datasheed made me wonder what the other "undefinded" cases
would be.
Probably it will also be "protect all"; I just mentioned it because it
would
be an easy special case to handle. I don't think we should mask out
anything,
either use the slot in between (ie the one which the formula gives us)
to
protect all or use the largest setting (ie [TB3 |] TB2 | TB1 | TB0).
And
given the reasons above, I'd prefer the latter.
I'm fine with using the largest setting, but we'll need the proof
before going
this way.
have a look at 7.1.11:
https://www.winbond.com/resource-files/w25q80dv_revg_07212015.pdf
BP[2:0]=0b101 and 0b110 is not defined.
-michael
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/
