linux-kernel - Re: [PATCH] PM / MTD: Fix regressions related to mtd

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 for Android: free password hash cracker in your pocket

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

Message-ID: <1326637306.2201.22.camel@koala>
Date:	Sun, 15 Jan 2012 16:21:45 +0200
From:	Artem Bityutskiy <dedekind1@...il.com>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
Cc:	Linux PM list <linux-pm@...r.kernel.org>,
	David Woodhouse <David.Woodhouse@...el.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Artem Bityutskiy <artem.bityutskiy@...ux.intel.com>,
	linux-mtd@...ts.infradead.org
Subject: Re: [PATCH] PM / MTD: Fix regressions related to mtd_suspend()

On Sun, 2012-01-15 at 14:43 +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rjw@...k.pl>
> 
> Commit 3fe4bae88460869a8e553397cd9057a4ee7ca341 (mtd: introduce
> mtd_suspend interface) introduced the mtd_suspend() routine
> which didn't check whether or not mtd and mtd->suspend were
> both valid pointers.  Later commit 079c985e7a6f4ce60f931cebfdd5ee3c3
> (mtd: do not use mtd->suspend and mtd->resume directly) added the
> check for mtd->suspend, but still it failed to check mtd.  Moreover,
> it caused mtd_suspend() to return an error code for NULL
> mtd->suspend, which makes system suspend fail on one of my test
> systems on every attempt and which is a regression from v3.2.
> 
> Fix mtd_suspend() by making it check if mtd is not NULL and
> return 0 if mtd or mtd->suspend is NULL, which shouldn't prevent
> the system from suspending.
> 
> Signed-off-by: Rafael J. Wysocki <rjw@...k.pl>
> ---
>  include/linux/mtd/mtd.h |    4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> Index: linux/include/linux/mtd/mtd.h
> ===================================================================
> --- linux.orig/include/linux/mtd/mtd.h
> +++ linux/include/linux/mtd/mtd.h
> @@ -427,9 +427,7 @@ static inline int mtd_is_locked(struct m
>  
>  static inline int mtd_suspend(struct mtd_info *mtd)
>  {
> -	if (!mtd->suspend)
> -		return -EOPNOTSUPP;
> -	return mtd->suspend(mtd);
> +	return mtd && mtd->suspend ? mtd->suspend(mtd) : 0;
>  }

Thanks! But All the MTD API functions assume that the "mtd" pointer is
non-NULL, and this function should to the same for consistency.

Would you please give this patch a test?

diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c
index b265188..de96865 100644
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -119,7 +119,7 @@ static int mtd_cls_suspend(struct device *dev, pm_message_t state)
 {
        struct mtd_info *mtd = dev_get_drvdata(dev);
 
-       return mtd_suspend(mtd);
+       return mtd ? mtd_suspend(mtd) : 0;
 }
 
 static int mtd_cls_resume(struct device *dev)
diff --git a/include/linux/mtd/mtd.h b/include/linux/mtd/mtd.h
index 1a81fde..d8c7aad 100644
--- a/include/linux/mtd/mtd.h
+++ b/include/linux/mtd/mtd.h
@@ -427,9 +427,7 @@ static inline int mtd_is_locked(struct mtd_info *mtd, loff_t ofs, uint64_t len)
 
 static inline int mtd_suspend(struct mtd_info *mtd)
 {
-       if (!mtd->suspend)
-               return -EOPNOTSUPP;
-       return mtd->suspend(mtd);
+       return mtd->suspend ? mtd->suspend(mtd) : 0;
 }

-- 
Best Regards,
Artem Bityutskiy

Download attachment "signature.asc" of type "application/pgp-signature" (491 bytes)