[PATCH v2] dfu: dfu_sf: Add support for multiple flashes
patrick.delaunay at foss.st.com
patrick.delaunay at foss.st.com
Tue Oct 5 15:41:55 CEST 2021
Hi,
ST Restricted
> -----Original Message-----
> From: Marek Vasut <marex at denx.de>
> Sent: mardi 5 octobre 2021 15:33
> To: patrick.delaunay at foss.st.com; u-boot at lists.denx.de
> Cc: 'Lukasz Majewski' <lukma at denx.de>
> Subject: Re: [PATCH v2] dfu: dfu_sf: Add support for multiple flashes
>
> On 10/5/21 2:50 PM, patrick.delaunay at foss.st.com wrote:
> > Hi Marek,
> >
> >
> > ST Restricted
> >
> >> -----Original Message-----
> >> From: U-Boot <u-boot-bounces at lists.denx.de> On Behalf Of Marek Vasut
> >> Sent: mardi 5 octobre 2021 14:40
> >> To: u-boot at lists.denx.de
> >> Cc: Marek Vasut <marex at denx.de>; Lukasz Majewski <lukma at denx.de>
> >> Subject: [PATCH v2] dfu: dfu_sf: Add support for multiple flashes
> >>
> >> Add dfu_alt_info option which allows specifying multiple SPI flashes
> >> as an
> > alt info.
> >> The syntax is as follows:
> >>
> >> altname sf bus:cs[:speed[:mode]]
> >>
> >> Example:
> >> dfu_alt_info=qspi0 sf 0:0;qspi1 sf 0:1
> >>
> >> Signed-off-by: Marek Vasut <marex at denx.de>
> >> Cc: Lukasz Majewski <lukma at denx.de>
> >> ---
> >> V2: Fix device string parsing in all cases
> >> ---
> >> drivers/dfu/dfu_sf.c | 31 ++++++++++++++++++++++++-------
> >> 1 file changed, 24 insertions(+), 7 deletions(-)
> >>
> >> diff --git a/drivers/dfu/dfu_sf.c b/drivers/dfu/dfu_sf.c index
> >> b72493ced86..6d33497990a 100644
> >> --- a/drivers/dfu/dfu_sf.c
> >> +++ b/drivers/dfu/dfu_sf.c
> >> @@ -168,30 +168,46 @@ static struct spi_flash *parse_dev(char
> >> *devstr)
> > int
> >> dfu_fill_entity_sf(struct dfu_entity *dfu, char *devstr, char *s) {
> >> char *st;
> >> - char *devstr_bkup = strdup(devstr);
> >> -
> >> - dfu->data.sf.dev = parse_dev(devstr_bkup);
> >> - free(devstr_bkup);
> >> - if (!dfu->data.sf.dev)
> >> - return -ENODEV;
> >>
> >> dfu->dev_type = DFU_DEV_SF;
> >> - dfu->max_buf_size = dfu->data.sf.dev->sector_size;
> >>
> >> st = strsep(&s, " ");
> >> if (!strcmp(st, "raw")) {
> >> + char *devstr_bkup = strdup(devstr);
> >> + dfu->data.sf.dev = parse_dev(devstr_bkup);
> >> + free(devstr_bkup);
> >> + if (!dfu->data.sf.dev)
> >> + return -ENODEV;
> >> +
> >> dfu->layout = DFU_RAW_ADDR;
> >> + dfu->max_buf_size = dfu->data.sf.dev->sector_size;
> >> dfu->data.sf.start = hextoul(s, &s);
> >> s++;
> >> dfu->data.sf.size = hextoul(s, &s);
> >> + } else if (!strcmp(st, "sf")) {
> >> + st = strsep(&s, " ");
> >> + dfu->data.sf.dev = parse_dev(st);
> >> + if (!dfu->data.sf.dev)
> >> + return -ENODEV;
> >> +
> >> + dfu->layout = DFU_RAW_ADDR;
> >> + dfu->max_buf_size = dfu->data.sf.dev->sector_size;
> >> + dfu->data.sf.start = 0;
> >> + dfu->data.sf.size = dfu->data.sf.dev->size;
> >> } else if (CONFIG_IS_ENABLED(DFU_SF_PART) &&
> >> (!strcmp(st, "part") || !strcmp(st, "partubi"))) {
> >> + char *devstr_bkup = strdup(devstr);
> >> char mtd_id[32];
> >> struct mtd_device *mtd_dev;
> >> u8 part_num;
> >> struct part_info *pi;
> >> int ret, dev, part;
> >>
> >> + dfu->data.sf.dev = parse_dev(devstr_bkup);
> >> + free(devstr_bkup);
> >> + if (!dfu->data.sf.dev)
> >> + return -ENODEV;
> >> +
> >> dfu->layout = DFU_RAW_ADDR;
> >>
> >> dev = dectoul(s, &s);
> >> @@ -208,6 +224,7 @@ int dfu_fill_entity_sf(struct dfu_entity *dfu,
> >> char
> > *devstr,
> >> char *s)
> >> printf("Could not locate '%s'\n", mtd_id);
> >> return -1;
> >> }
> >> + dfu->max_buf_size = dfu->data.sf.dev->sector_size;
> >> dfu->data.sf.start = pi->offset;
> >> dfu->data.sf.size = pi->size;
> >> if (!strcmp(st, "partubi"))
> >> --
> >> 2.33.0
> >
> > The same use case can be supported with mtd backend
> >
> > https://elixir.bootlin.com/u-boot/latest/source/doc/usage/dfu.rst
> >
> >
> > U-Boot> env set dfu_alt_info \
> > "mtd nor0=spl part 1;u-boot part 2;u-boot-env part 3&"\
> > "mtd nand0=UBI partubi 1"
>
> Does it work for multiple SPI NORs at the same time ?
If it is support on mtd side, yes it is supported in dfu...
And I think it is supported now (with u-class UCLASS_SPI_FLASH).
I need to check.deeper
Patrick
More information about the U-Boot
mailing list