linux-kernel - Add NULL pointer checks to regulator

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

Message-ID: <CADkCAutiOaeJitJ-8xCWRJGaxD=m7a+2oNc-xqKJu9t+JdZYQw@mail.gmail.com>
Date:	Wed, 25 Jan 2012 08:39:35 +0100
From:	Bill Gatliff <bgat@...lgatliff.com>
To:	linux-kernel@...r.kernel.org
Cc:	lrg@...mlogic.co.uk
Subject: Add NULL pointer checks to regulator_enable() et. al?

Guys:

(Please CC me on replies).

I have this pattern coming up a LOT in my code:

   struct regulator *reg;

   /* probe() */
   reg = regulator_get(dev, "VLOGIC");

   /* elsewhere() */
   if (!IS_ERR_OR_NULL(reg))
     ret = regulator_enable(reg);

   /* remove() */
   if (!IS_ERR_OR_NULL(reg))
      regulator_put(reg);

It comes up particularly often in drivers that are doing aggressive
runtime power management.

Unlike other kernel API, a NULL pointer from regulator_get() isn't a
true "error": it just means that there is no regulator associated with
the requested pin and device.  It seems natural, therefore, for
regulator_enable() and friends to not consider a NULL regulator
pointer to be an error--- and definitely not something OOPS-worthy.
But if I forget even ONE check in the above code, that's exactly what
I get: an OOPS.  But all those checks seem to be just tedium and noise
in this case.

Is there something truly objectionable in the following patch?  If
not, then I'm happy to post a more complete one.

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 5b2328d..b6e303a 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -1482,9 +1482,16 @@ static int _regulator_enable(struct regulator_dev *rdev)
  */
 int regulator_enable(struct regulator *regulator)
 {
-	struct regulator_dev *rdev = regulator->rdev;
+	struct regulator_dev *rdev;
 	int ret = 0;

+	if (unlikely(!regulator))
+		return 0;
+	if (unlikely(IS_ERR(regulator)))
+		return PTR_ERR(regulator);
+
+	rdev = regulator->rdev;
+
 	mutex_lock(&rdev->mutex);

 	if (!regulator_check_voltage_update(rdev)) {

Note that I'm not arguing for such treatment all over the kernel---
just certain places in the regulator framework!

Kindest regards,

b.g.
-- 
Bill Gatliff
bgat@...lgatliff.com
--
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/