Re: [PATCH v5 1/3] mtd: spi-nor: add support for is25wp256

[Sagar Kadam]

Hello Vignesh,

On Fri, Jun 21, 2019 at 11:33 AM Vignesh Raghavendra <vigneshr@xxxxxx> wrote:
>
> Hi,
>
> On 17/06/19 8:48 PM, Sagar Kadam wrote:
> > Hello Vignesh,
> >
> > Thanks for your review comments.
> >
> > On Sun, Jun 16, 2019 at 6:14 PM Vignesh Raghavendra <vigneshr@xxxxxx> wrote:
> >>
> >> Hi,
> >>
> >> On 12-Jun-19 4:17 PM, Sagar Shrikant Kadam wrote:
> >> [...]
> >>
> >>> @@ -4129,7 +4137,7 @@ int spi_nor_scan(struct spi_nor *nor, const char *name,
> >>>       if (ret)
> >>>               return ret;
> >>>
> >>> -     if (nor->addr_width) {
> >>> +     if (nor->addr_width && JEDEC_MFR(info) != SNOR_MFR_ISSI) {
> >>>               /* already configured from SFDP */
> >>
> >> Hmm, why would you want to ignore addr_width that's read from SFDP table?
> >
> > The SFDP table for ISSI device considered here, has addr_width set to
> > 3 byte, and the flash considered
> > here is 32MB. With 3 byte address width we won't be able to access
> > flash memories higher address range.
>
> Is it specific to a particular ISSI part as indicated here[1]? If so,
> please submit solution agreed there i.e. use spi_nor_fixups callback
>
> [1]https://patchwork.ozlabs.org/patch/1056049/
>

Thanks for sharing the link.
>From what I understand here, it seems that "Address Bytes" of SFDP
table for the device under
consideration (is25lp256) supports 3 byte only Addressing mode
(DWORD1[18:17] = 0b00.
where as that of ISSI device (is25LP/WP 256Mb/512/Mb/1Gb) support 3 or
4 byte Addressing mode DWORD1[18:17] = 0b01.

> > Hence I have ignored the addr width from SFDP.  I have verified that
> > with 3 byte address width, the
> > flascp util fails while verifying the written data.  Please let me
> > know your views on this?
> >
> If this affects multiple ISSI parts then:
> Instead of checking for mfr code, look for SNOR_F_4B_OPCODES flag in
> flash_info struct of the device and let it take precedence over SFDP in
> case size is over 16MB
>

So as per your suggestion I think second approach is a better one.
I will send this in V6.

Thanks & Regards,
Sagar

> Regards
> Vignesh
>
> > BR,
> > Sagar Kadam
> >
> >> Regards
> >> Vignesh
> >>
> >>
> >>>       } else if (info->addr_width) {
> >>>               nor->addr_width = info->addr_width;
> >>> diff --git a/include/linux/mtd/spi-nor.h b/include/linux/mtd/spi-nor.h
> >>> index b3d360b..ff13297 100644
> >>> --- a/include/linux/mtd/spi-nor.h
> >>> +++ b/include/linux/mtd/spi-nor.h
> >>> @@ -19,6 +19,7 @@
> >>>  #define SNOR_MFR_ATMEL               CFI_MFR_ATMEL
> >>>  #define SNOR_MFR_GIGADEVICE  0xc8
> >>>  #define SNOR_MFR_INTEL               CFI_MFR_INTEL
> >>> +#define SNOR_MFR_ISSI                0x9d            /* ISSI */
> >>>  #define SNOR_MFR_ST          CFI_MFR_ST      /* ST Micro */
> >>>  #define SNOR_MFR_MICRON              CFI_MFR_MICRON  /* Micron */
> >>>  #define SNOR_MFR_MACRONIX    CFI_MFR_MACRONIX
> >>>
>
> --
> Regards
> Vignesh