linux-kernel - Re: [PATCH 2/2] pinctrl: skip deferral of hogs

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] [day] [month] [year] [list]

Message-ID: <50C79C41.8070305@wwwdotorg.org>
Date:	Tue, 11 Dec 2012 13:49:05 -0700
From:	Stephen Warren <swarren@...dotorg.org>
To:	Linus Walleij <linus.walleij@...ricsson.com>
CC:	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	Stephen Warren <swarren@...dia.com>,
	Anmar Oueja <anmar.oueja@...aro.org>,
	Linus Walleij <linus.walleij@...aro.org>
Subject: Re: [PATCH 2/2] pinctrl: skip deferral of hogs

On 12/11/2012 06:36 AM, Linus Walleij wrote:
> From: Linus Walleij <linus.walleij@...aro.org>
> 
> Up until now, as hogs were always taken at the end of the
> pin control device registration, it didn't cause any problem.
> But when starting to hog pins from the device core it will
> cause deferral of the pin controller device itself since the
> default pin fetch is done *before* the device probes, so
> let's fix this annoyance (which is also aesthetically ugly).

I guess this is OK, although I can't quite get over the feeling of it
being a bit of a hack.

The issue is IIUC that the driver core is going to call pinctrl_get()
for any device before it's probed, and if it succeeds it will select the
default state, but this patch deliberately fails the pinctrl_get() for a
pin controller, and hence relies on pinctrl_register() making a second
attempt at pinctrl_get(), and also activating the default state there.
That's a bit of duplication and rather asymmetric.

That all said, I don't really see what we can do better, so I won't object!

Oh, one comment on the patch. It returns -ENODEV as soon as any hog is
seen. Just in case someone does something wierd, shouldn't it accumulate
the "worst-case" return value and not return early? Consider a pin
controller whose mapping table entries contain both hogs, /and/ entries
that use a different pin controller. That's a little contrived, but
trivial to actually do. As such, I think we want something more like:

ret = 0
for every mapping table entry:
    if can't find pinctrl device:
        if map entry is a hog:
            # Requiring defer is a worst case, so defer the return!
            ret = -ENODEV
        else:
            # This is worst-case, so can return now
            return -EPROBE_DEFER
if ret:
    cleanup
return ret

That way, if a mapping table entry requiring defer comes after a hog,
we'll still return the correct value.

--
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/