[PATCH] ssb: sprom: add dev_id field for value overriding standard ID

[PATCH] ssb: sprom: add dev_id field for value overriding standard ID

[Rafał Miłecki]

Some devices may have different features despite sharing the same ID
(e.g. PCI ID). For example 14e4:4331 is usually a dual band, but this
can be "limited". Device with "pci/x/y/devid=0x4332" supports 2.4 GHz
only. Similarly 0x4333 will mean support for 5 GHz only.
Add entry in SPROM so info described above can be extracted and stored.

Signed-off-by: Rafa? Mi?ecki <zajec5@gmail.com>
---
I managed to find "devid" on every common SPROM rev:
pci/1/1/sromrev=2
pci/1/1/devid=0x4318

pci/1/1/sromrev=4
pci/1/1/devid=0x4329

pci/1/1/sromrev=8
pci/1/1/devid=0x432d

pci/1/1/sromrev=9
pci/1/1/devid=0x4332

pci/1/1/sromrev=11
pci/1/1/devid=0x43a1

So it makes the most sense to always try to extract it.
---
 arch/mips/bcm47xx/sprom.c | 1 +
 include/linux/ssb/ssb.h   | 1 +
 2 files changed, 2 insertions(+)

diff --git a/arch/mips/bcm47xx/sprom.c b/arch/mips/bcm47xx/sprom.c
index a8b5408..da4cdb1 100644
--- a/arch/mips/bcm47xx/sprom.c
+++ b/arch/mips/bcm47xx/sprom.c
@@ -168,6 +168,7 @@ static void nvram_read_alpha2(const char *prefix, const char *name,
 static void bcm47xx_fill_sprom_r1234589(struct ssb_sprom *sprom,
 					const char *prefix, bool fallback)
 {
+	nvram_read_u16(prefix, NULL, "devid", &sprom->dev_id, 0, fallback);
 	nvram_read_u8(prefix, NULL, "ledbh0", &sprom->gpio0, 0xff, fallback);
 	nvram_read_u8(prefix, NULL, "ledbh1", &sprom->gpio1, 0xff, fallback);
 	nvram_read_u8(prefix, NULL, "ledbh2", &sprom->gpio2, 0xff, fallback);
diff --git a/include/linux/ssb/ssb.h b/include/linux/ssb/ssb.h
index 07ef9b8..4568a5c 100644
--- a/include/linux/ssb/ssb.h
+++ b/include/linux/ssb/ssb.h
@@ -33,6 +33,7 @@ struct ssb_sprom {
 	u8 et1phyaddr;		/* MII address for enet1 */
 	u8 et0mdcport;		/* MDIO for enet0 */
 	u8 et1mdcport;		/* MDIO for enet1 */
+	u16 dev_id;		/* Device ID overriding e.g. PCI ID */
 	u16 board_rev;		/* Board revision number from SPROM. */
 	u16 board_num;		/* Board number from SPROM. */
 	u16 board_type;		/* Board type from SPROM. */
-- 
1.8.4.5

[PATCH] ssb: sprom: add dev_id field for value overriding standard ID

[Hauke Mehrtens]

On 05/18/2014 12:22 AM, Rafa? Mi?ecki wrote:
> Some devices may have different features despite sharing the same ID
> (e.g. PCI ID). For example 14e4:4331 is usually a dual band, but this
> can be "limited". Device with "pci/x/y/devid=0x4332" supports 2.4 GHz
> only. Similarly 0x4333 will mean support for 5 GHz only.
> Add entry in SPROM so info described above can be extracted and stored.
> 
> Signed-off-by: Rafa? Mi?ecki <zajec5@gmail.com>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>

This patch is ok, I missed that when adding BCM4716 support to brcmsmac.
There is also a vendid var which should be used. See this code from
brcmsmac:
http://lxr.free-electrons.com/source/drivers/staging/brcm80211/brcmsmac/wlc_bmac.c?v=3.0#L660

Hauke

> ---
> I managed to find "devid" on every common SPROM rev:
> pci/1/1/sromrev=2
> pci/1/1/devid=0x4318
> 
> pci/1/1/sromrev=4
> pci/1/1/devid=0x4329
> 
> pci/1/1/sromrev=8
> pci/1/1/devid=0x432d
> 
> pci/1/1/sromrev=9
> pci/1/1/devid=0x4332
> 
> pci/1/1/sromrev=11
> pci/1/1/devid=0x43a1
> 
> So it makes the most sense to always try to extract it.
> ---
>  arch/mips/bcm47xx/sprom.c | 1 +
>  include/linux/ssb/ssb.h   | 1 +
>  2 files changed, 2 insertions(+)
> 
> diff --git a/arch/mips/bcm47xx/sprom.c b/arch/mips/bcm47xx/sprom.c
> index a8b5408..da4cdb1 100644
> --- a/arch/mips/bcm47xx/sprom.c
> +++ b/arch/mips/bcm47xx/sprom.c
> @@ -168,6 +168,7 @@ static void nvram_read_alpha2(const char *prefix, const char *name,
>  static void bcm47xx_fill_sprom_r1234589(struct ssb_sprom *sprom,
>  					const char *prefix, bool fallback)
>  {
> +	nvram_read_u16(prefix, NULL, "devid", &sprom->dev_id, 0, fallback);
>  	nvram_read_u8(prefix, NULL, "ledbh0", &sprom->gpio0, 0xff, fallback);
>  	nvram_read_u8(prefix, NULL, "ledbh1", &sprom->gpio1, 0xff, fallback);
>  	nvram_read_u8(prefix, NULL, "ledbh2", &sprom->gpio2, 0xff, fallback);
> diff --git a/include/linux/ssb/ssb.h b/include/linux/ssb/ssb.h
> index 07ef9b8..4568a5c 100644
> --- a/include/linux/ssb/ssb.h
> +++ b/include/linux/ssb/ssb.h
> @@ -33,6 +33,7 @@ struct ssb_sprom {
>  	u8 et1phyaddr;		/* MII address for enet1 */
>  	u8 et0mdcport;		/* MDIO for enet0 */
>  	u8 et1mdcport;		/* MDIO for enet1 */
> +	u16 dev_id;		/* Device ID overriding e.g. PCI ID */
>  	u16 board_rev;		/* Board revision number from SPROM. */
>  	u16 board_num;		/* Board number from SPROM. */
>  	u16 board_type;		/* Board type from SPROM. */
> 

[PATCH] ssb: sprom: add dev_id field for value overriding standard ID

[Rafał Miłecki]

On 18 May 2014 11:43, Hauke Mehrtens <hauke@hauke-m.de> wrote:
> There is also a vendid var which should be used. See this code from
> brcmsmac:
> http://lxr.free-electrons.com/source/drivers/staging/brcm80211/brcmsmac/wlc_bmac.c?v=3.0#L660

You're right, but is this used anywhere?

I can't reliably follow brcmsmac code, it copies the same data over
and over. Simple grepping for "vendorid" shows copying from "vendorid"
to "vendorid" of various structs, no real usage.
The one exception is "sha_params.vid = wlc_hw->vendorid;. However grepping with:
grep -R "\.vid" ./*
grep -R "\->vid" ./*
doesn't reveal any usage neither.

Seems like a dead code.