linux-kernel - Re: [PATCH] pinctrl: msm: Add support for MSM TLMM pinmux

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 15 Aug 2013 22:47:00 +0200
From:	Linus Walleij <linus.walleij@...aro.org>
To:	Hanumant Singh <hanumant@...eaurora.org>
Cc:	Bjorn Andersson <bjorn@...o.se>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] pinctrl: msm: Add support for MSM TLMM pinmux

On Thu, Aug 15, 2013 at 7:44 PM, Hanumant Singh <hanumant@...eaurora.org> wrote:

> Ok i can switch to using pin groups defined in per soc files.
> But in our case we have one soc going into different types of boards.
> (atleast 3). In each of the boards the same external devices end up using
> different pins. For ex camera on board 1 uses different pin group
> then the same camera on board 2. Both having the same SOC.
> So in this case the design would be to have all possible pin groups
> for different boards enumerated in the same soc-pinctrl.c file?

Sorry I don't get this at all.

What pin groups and functions that exist on a SoC is what you put into
a SoC driver. Because this is a hardware characteristic.

How these are combined on a board into different states is what you put
into the device tree. (Or platform data.)

> Also in this implementation I will have.
> 1) pinctrl-msm.c => DT parsing and interface to framework.
> 2) pinctrl-msm-tlmm<version>.c => Register programming and pin types
> supported by a particular TLMM pinmux version.
> 3) pinctrl-<soc>.c => All the pins/pin groups supported by a given SOC.

Seems OK.

> As I
> mentioned we will have a bloat of these, since we have entire families of
> SOC using a given TLMM version but with unique pin groupings.

Bring 'em on. But is that really different groups you are talking about,
and not just combinations of groups with functions for a certain board
as I describe above?

If you have many SoC subdrivers, consider creating a subdir as some
drivers already have.

> I don't override the default values, since resistor values are not
> configurable. I only care about which config option is chosen to program it
> as pull up/down or disable.

That sounds correct.

>> Actually the data should be more carefully handled for each
>> config option I think.
>>
> Not sure I follow. Based on my use mentioned above, what do you suggest for
> the read? Should I return default config value, which is what I am doing ?

Here:

+       switch (id) {
+       case PIN_CONFIG_BIAS_DISABLE:
+               mask = TLMMV3_GP_PULL_MASK;
+               shft = TLMMV3_GP_PULL_SHFT;
+               data = TLMMV3_NO_PULL;

data should just be zero. (Maybe TLMMV3_NO_PULL is
zero? But anyway...)

+               if (!write) {
+                       val >>= shft;
+                       val &= mask;
+                       data = rval_to_pull(val);

Dito.

Because it has no meaning in the framework.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/