linux-kernel - Re: [PATCH 6/6 v5] regmap: Incorporate the regcache core into regmap

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]

Message-ID: <20110919161358.GA540@opensource.wolfsonmicro.com>
Date:	Mon, 19 Sep 2011 17:13:59 +0100
From:	Mark Brown <broonie@...nsource.wolfsonmicro.com>
To:	Lars-Peter Clausen <lars@...afoo.de>
Cc:	Dimitris Papastamos <dp@...nsource.wolfsonmicro.com>,
	linux-kernel@...r.kernel.org, Liam Girdwood <lrg@...com>,
	Graeme Gregory <gg@...mlogic.co.uk>,
	Samuel Oritz <sameo@...ux.intel.com>
Subject: Re: [PATCH 6/6 v5] regmap: Incorporate the regcache core into regmap

On Mon, Sep 19, 2011 at 05:57:49PM +0200, Lars-Peter Clausen wrote:
> On 09/19/2011 03:34 PM, Dimitris Papastamos wrote:

> > +	if (!map->cache_bypass) {
> > +		ret = regcache_write(map, reg, val);
> > +		if (!ret || map->cache_only)
> > +			return 0;

> The hw write shouldn't be skipped if the cache write is successful. We should
> only exit here if cache_only is set.

There's a couple of other issues too.  I've already got the following
patch for this one locally:

--- a/drivers/base/regmap/regmap.c
+++ b/drivers/base/regmap/regmap.c
@@ -304,7 +304,9 @@ static int _regmap_write(struct regmap *map,
unsigned int re
 
        if (!map->cache_bypass) {
                ret = regcache_write(map, reg, val);
-               if (!ret || map->cache_only)
+               if (ret != 0)
+                       return ret;
+               if (map->cache_only)
                        return 0;
        }
 
> I also wonder if we should pass the return value of regcache_write on to the
> caller if cache_only is set.

Yup.

> Btw. what should happen if both cache_bypass and cache_only are set? Or is that
> an invalid configuration?

That's not sensible.  Probably BUG_ON().

> > @@ -428,6 +446,14 @@ int regmap_read(struct regmap *map, unsigned int reg, unsigned int *val)

> >  	mutex_lock(&map->lock);

> > +	if (!map->cache_bypass) {
> > +		ret = regcache_read(map, reg, val);
> > +		if (!ret) {
> > +			mutex_unlock(&map->lock);
> > +			return 0;
> > +		}
> > +	}

> This should go into _regmap_read. Otherwise regmap_update_bits will always use
> a hw read.

Yup, got that too.

> Also if cache_only is set I guess we shouldn't fallback to a hw read.

Probably.  Cache only is *mostly* there for write side stuff, but it'd
be useful to add this.  I've got a patch to do this.
--
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/