| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Tue, 25 May 2021 10:24:42 +0930
From: "Andrew Jeffery" <andrew@...id.au>
To: "Steven Lee" <steven_lee@...eedtech.com>,
"Linus Walleij" <linus.walleij@...aro.org>,
"Rob Herring" <robh+dt@...nel.org>,
"Joel Stanley" <joel@....id.au>,
"moderated list:ASPEED PINCTRL DRIVERS"
<linux-aspeed@...ts.ozlabs.org>,
"moderated list:ASPEED PINCTRL DRIVERS" <openbmc@...ts.ozlabs.org>,
"open list:ASPEED PINCTRL DRIVERS" <linux-gpio@...r.kernel.org>,
"open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
<devicetree@...r.kernel.org>,
"moderated list:ARM/ASPEED MACHINE SUPPORT"
<linux-arm-kernel@...ts.infradead.org>,
"open list" <linux-kernel@...r.kernel.org>
Cc: "Hongwei Zhang" <Hongweiz@....com>,
"Ryan Chen" <ryan_chen@...eedtech.com>,
"Billy Tsai" <billy_tsai@...eedtech.com>
Subject: Re: [PATCH v1 3/3] pinctrl: pinctrl-aspeed-g6: Add sgpio pinctrl settings
Hi Steven,
On Mon, 24 May 2021, at 20:43, Steven Lee wrote:
> AST2600 supports 2 SGPIO master interfaces and 2 SGPIO slave interfaces.
> Current pinctrl driver only define the first sgpio master and slave
> interfaces.
> The sencond SGPIO master and slave interfaces should be added in
> pinctrl driver as well.
>
> Signed-off-by: Steven Lee <steven_lee@...eedtech.com>
> ---
> drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c | 24 ++++++++++++++++++----
> drivers/pinctrl/aspeed/pinmux-aspeed.h | 9 ++++++++
> 2 files changed, 29 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c
> b/drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c
> index 5c1a109842a7..d0e9ab9d1a9c 100644
> --- a/drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c
> +++ b/drivers/pinctrl/aspeed/pinctrl-aspeed-g6.c
> @@ -46,8 +46,10 @@
> #define SCU620 0x620 /* Disable GPIO Internal Pull-Down #4 */
> #define SCU634 0x634 /* Disable GPIO Internal Pull-Down #5 */
> #define SCU638 0x638 /* Disable GPIO Internal Pull-Down #6 */
> +#define SCU690 0x690 /* Multi-function Pin Control #24 */
> #define SCU694 0x694 /* Multi-function Pin Control #25 */
> #define SCU69C 0x69C /* Multi-function Pin Control #27 */
> +#define SCU6D0 0x6D0 /* Multi-function Pin Control #29 */
> #define SCUC20 0xC20 /* PCIE configuration Setting Control */
>
> #define ASPEED_G6_NR_PINS 256
> @@ -81,13 +83,17 @@ FUNC_GROUP_DECL(I2C12, L26, K24);
> #define K26 4
> SIG_EXPR_LIST_DECL_SESG(K26, MACLINK1, MACLINK1, SIG_DESC_SET(SCU410, 4));
> SIG_EXPR_LIST_DECL_SESG(K26, SCL13, I2C13, SIG_DESC_SET(SCU4B0, 4));
> -PIN_DECL_2(K26, GPIOA4, MACLINK1, SCL13);
> +SIG_EXPR_LIST_DECL_SESG(K26, SGPS2CK, SGPS2, SIG_DESC_SET(SCU690, 4));
> +SIG_EXPR_LIST_DECL_SESG(K26, SGPM2CLK, SGPM2, SIG_DESC_SET(SCU6D0, 4));
> +PIN_DECL_4(K26, GPIOA4, SGPM2CLK, SGPS2CK, MACLINK1, SCL13);
Is this the right priority order? Looking at the Multi-Function Pin
Mapping and Control table, function 1 is MACLINK1,
function 2 is SCL13, function 3 is SGPS2CK, and I assume function 4
would be SGPM2CLK, except it's not documented in the table in v9 of the
datasheet (I hope it will be documented?).
If function 1 is the highest priority (which is what all the Aspeed
pinctrl drivers assume), then this should be:
PIN_DECL_4(K26, GPIOA4, MACLINK1, SCL13, SGPS2CK, SGPM2CLK);
Anyway, one of several things could be at fault here:
1. I've made a wrong assumption about the priority order in how I've
implemented pinctrl support for Aspeed SoCs
2. The Multi-Function Pin Mapping and Control table is out of date and
needs to be fixed (which it already does as it doesn't list SGPM2CLK).
3. The patch needs to align with the assumptions of the Aspeed pinctrl
support.
I don't think it's 1 as I haven't heard of any issues where we are
getting incorrect behaviour because of pinmux. I don't think it's 2 as
this patch makes a non-linear change to the ordering. So my hunch is
the issue is 3, that the patch is putting the signals in the wrong order.
In this case, you want the PIN_DECL_4(...) I outlined above.
> FUNC_GROUP_DECL(MACLINK1, K26);
>
> #define L24 5
> SIG_EXPR_LIST_DECL_SESG(L24, MACLINK2, MACLINK2, SIG_DESC_SET(SCU410, 5));
> SIG_EXPR_LIST_DECL_SESG(L24, SDA13, I2C13, SIG_DESC_SET(SCU4B0, 5));
> -PIN_DECL_2(L24, GPIOA5, MACLINK2, SDA13);
> +SIG_EXPR_LIST_DECL_SESG(L24, SGPS2LD, SGPS2, SIG_DESC_SET(SCU690, 5));
> +SIG_EXPR_LIST_DECL_SESG(L24, SGPM2LD, SGPM2, SIG_DESC_SET(SCU6D0, 5));
> +PIN_DECL_4(L24, GPIOA5, SGPM2LD, SGPS2LD, MACLINK2, SDA13);
See above.
> FUNC_GROUP_DECL(MACLINK2, L24);
>
> FUNC_GROUP_DECL(I2C13, K26, L24);
> @@ -95,16 +101,22 @@ FUNC_GROUP_DECL(I2C13, K26, L24);
> #define L23 6
> SIG_EXPR_LIST_DECL_SESG(L23, MACLINK3, MACLINK3, SIG_DESC_SET(SCU410, 6));
> SIG_EXPR_LIST_DECL_SESG(L23, SCL14, I2C14, SIG_DESC_SET(SCU4B0, 6));
> -PIN_DECL_2(L23, GPIOA6, MACLINK3, SCL14);
> +SIG_EXPR_LIST_DECL_SESG(L23, SGPS2O, SGPS2, SIG_DESC_SET(SCU690, 6));
> +SIG_EXPR_LIST_DECL_SESG(L23, SGPM2O, SGPM2, SIG_DESC_SET(SCU6D0, 6));
> +PIN_DECL_4(L23, GPIOA6, SGPM2O, SGPS2O, MACLINK3, SCL14);
See above.
> FUNC_GROUP_DECL(MACLINK3, L23);
>
> #define K25 7
> SIG_EXPR_LIST_DECL_SESG(K25, MACLINK4, MACLINK4, SIG_DESC_SET(SCU410, 7));
> SIG_EXPR_LIST_DECL_SESG(K25, SDA14, I2C14, SIG_DESC_SET(SCU4B0, 7));
> -PIN_DECL_2(K25, GPIOA7, MACLINK4, SDA14);
> +SIG_EXPR_LIST_DECL_SESG(K25, SGPS2I, SGPS2, SIG_DESC_SET(SCU690, 7));
> +SIG_EXPR_LIST_DECL_SESG(K25, SGPM2I, SGPM2, SIG_DESC_SET(SCU6D0, 7));
> +PIN_DECL_4(K25, GPIOA7, SGPM2I, SGPS2I, MACLINK4, SDA14);
See above.
> FUNC_GROUP_DECL(MACLINK4, K25);
>
> FUNC_GROUP_DECL(I2C14, L23, K25);
> +FUNC_GROUP_DECL(SGPM2, K26, L24, L23, K25);
> +FUNC_GROUP_DECL(SGPS2, K26, L24, L23, K25);
>
> #define J26 8
> SIG_EXPR_LIST_DECL_SESG(J26, SALT1, SALT1, SIG_DESC_SET(SCU410, 8));
> @@ -2060,7 +2072,9 @@ static const struct aspeed_pin_group
> aspeed_g6_groups[] = {
> ASPEED_PINCTRL_GROUP(EMMCG4),
> ASPEED_PINCTRL_GROUP(EMMCG8),
> ASPEED_PINCTRL_GROUP(SGPM1),
> + ASPEED_PINCTRL_GROUP(SGPM2),
> ASPEED_PINCTRL_GROUP(SGPS1),
> + ASPEED_PINCTRL_GROUP(SGPS2),
> ASPEED_PINCTRL_GROUP(SIOONCTRL),
> ASPEED_PINCTRL_GROUP(SIOPBI),
> ASPEED_PINCTRL_GROUP(SIOPBO),
> @@ -2276,7 +2290,9 @@ static const struct aspeed_pin_function
> aspeed_g6_functions[] = {
> ASPEED_PINCTRL_FUNC(SD1),
> ASPEED_PINCTRL_FUNC(SD2),
> ASPEED_PINCTRL_FUNC(SGPM1),
> + ASPEED_PINCTRL_FUNC(SGPM2),
> ASPEED_PINCTRL_FUNC(SGPS1),
> + ASPEED_PINCTRL_FUNC(SGPS2),
> ASPEED_PINCTRL_FUNC(SIOONCTRL),
> ASPEED_PINCTRL_FUNC(SIOPBI),
> ASPEED_PINCTRL_FUNC(SIOPBO),
> diff --git a/drivers/pinctrl/aspeed/pinmux-aspeed.h
> b/drivers/pinctrl/aspeed/pinmux-aspeed.h
> index dba5875ff276..125df796af36 100644
> --- a/drivers/pinctrl/aspeed/pinmux-aspeed.h
> +++ b/drivers/pinctrl/aspeed/pinmux-aspeed.h
> @@ -730,6 +730,15 @@ struct aspeed_pin_desc {
> SIG_EXPR_LIST_PTR(pin, low), \
> SIG_EXPR_LIST_PTR(pin, other))
>
> +#define PIN_DECL_4(pin, other, high, medium, low1, low2) \
Bit of a nit pick, but we might as well drop identifying the priorities
as high, medium and low given we now have low1 and low2. Maybe
something like this:
#define PIN_DECL_4(pin, other, prio1, prio2, prio3, prio4)
Cheers,
Andrew
Powered by blists - more mailing lists