Re: [flashrom] [PATCH/RFC] finish jedec converstion

[Carl-Daniel Hailfinger]

On 24.01.2010 05:32, Sean Nelson wrote:
> Complete the addition of Feature Bits for all Jedec based chips.
> Add FEATURE_SHORT_RESET, FEATURE_LONG_RESET, and FEATURE_EITHER_RESET
> rewrite jedec functions to use getaddrmask
> [...]
> Signed-off-by: Sean Nelson
>
> On 1/23/2010 6:28 PM, Carl-Daniel Hailfinger wrote:
> > Umm. Does this mean we have 555/2AA/555 and 555/AAA/555 chips?
>
> Yes, the ST M29F002T/NT/B and the Eon EN29F002A(N) uses the
> 555/AAA/555 address order scheme. There is one chipdriver that uses a
> AAA/555/AAA order scheme which is why some chips has FEATURE_ADDR_555
> set for them.

Good.

As I wrote before, I didn't have time to check the changed entries in
flashchips.c, but a random sampling turned up some reset sequence bugs,
so maybe there are more of them:
> {
>    	.vendor		= "SST",
>    	.name		= "SST39SF010A",
>    	.bustype	= CHIP_BUSTYPE_PARALLEL,
>    	.manufacture_id	= SST_ID,
>    	.model_id	= SST_39SF010,
>    	.total_size	= 128,
>    	.page_size	= 4096,
> .feature_bits	= FEATURE_LONG_RESET,

Should be FEATURE_EITHER_RESET according to my datasheet.
> {
>    	.vendor		= "SST",
>    	.name		= "SST39SF020A",
>    	.bustype	= CHIP_BUSTYPE_PARALLEL,
>    	.manufacture_id	= SST_ID,
>    	.model_id	= SST_39SF020,
>    	.total_size	= 256,
>    	.page_size	= 4096,
> .feature_bits	= FEATURE_LONG_RESET,

Should be FEATURE_EITHER_RESET according to my datasheet.
> {
>    	.vendor		= "SST",
>    	.name		= "SST39SF040",
>    	.bustype	= CHIP_BUSTYPE_PARALLEL,
>    	.manufacture_id	= SST_ID,
>    	.model_id	= SST_39SF040,
>    	.total_size	= 512,
>    	.page_size	= 4096,
> .feature_bits	= FEATURE_LONG_RESET,

Should be FEATURE_EITHER_RESET according to my datasheet.

SST39SF010A / SST39SF020A / SST39SF040 all support both reset sequences
according to this datasheet: http://www.sst.com/dotAsset/40746.pdf
> Software ID Exit6  XXH  F0H
> Software ID Exit6 5555H AAH 2AAAH 55H 5555H F0H
> 6. Both Software ID Exit operations are equivalent
> We can delete the unused chip-drivers when we are sure that everything
> we can do generically can be done.

Yes.

For all stuff outside flashchips.c (I assume you tested compilation):
Acked-by: Carl-Daniel Hailfinger

For flashchips.c I'd like someone to cross-check your
write_jedec_1/write_jedec and FEATURE_*_RESET values with datasheets. I
think macavity or someone else on IRC may be very interested in doing
this because it is a good way to get familiar with datasheets.

Please don't get me wrong, the work you're doing is very valuable. It's
just that the changes are so big that I wouldn't trust myself to send a
100% correct patch either.

Regards,
Carl-Daniel