linux-kernel - Re: [PATCH v3 2/4] lib/string_helpers.c: protect string_get_size() against blk

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:	Fri, 30 Oct 2015 03:34:04 +0000
From:	James Bottomley <jbottomley@...n.com>
To:	"andy.shevchenko@...il.com" <andy.shevchenko@...il.com>
CC:	"ulf.hansson@...aro.org" <ulf.hansson@...aro.org>,
	"linux@...musvillemoes.dk" <linux@...musvillemoes.dk>,
	"andriy.shevchenko@...ux.intel.com" 
	<andriy.shevchenko@...ux.intel.com>,
	"vkuznets@...hat.com" <vkuznets@...hat.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"akpm@...ux-foundation.org" <akpm@...ux-foundation.org>,
	"keescook@...omium.org" <keescook@...omium.org>
Subject: Re: [PATCH v3 2/4] lib/string_helpers.c: protect string_get_size()
 against blk_size=0

On Fri, 2015-10-30 at 01:32 +0200, Andy Shevchenko wrote:
> On Fri, Oct 30, 2015 at 1:00 AM, James Bottomley <jbottomley@...n.com> wrote:
> > On Thu, 2015-10-29 at 17:30 +0100, Vitaly Kuznetsov wrote:
> >> Division by zero happens if blk_size=0 is supplied to string_get_size().
> >> Add WARN_ON() and set size to 0 to report '0 B'.
> >>
> >> Signed-off-by: Vitaly Kuznetsov <vkuznets@...hat.com>
> >> ---
> >>  lib/string_helpers.c | 5 +++++
> >>  1 file changed, 5 insertions(+)
> >>
> >> diff --git a/lib/string_helpers.c b/lib/string_helpers.c
> >> index f6c27dc..ff3575b 100644
> >> --- a/lib/string_helpers.c
> >> +++ b/lib/string_helpers.c
> >> @@ -50,6 +50,11 @@ void string_get_size(u64 size, u32 blk_size, const enum string_size_units units,
> >>
> >>       tmp[0] = '\0';
> >>       i = 0;
> >> +
> >> +     /* Calling string_get_size() with blk_size=0 is wrong! */
> >> +     if (WARN_ON(!blk_size))
> >
> > Get rid of the WARN_ON; it's the standard thing to do for a partially
> > connected device.  Seeing zero is standard in a whole variety of
> > situations.  SCSI shims the zero but most other drivers don't.
> 
> For *block* size? It will crash the kernel. I've checked, it wasn't
> changed from the beginning (b9f28d863594).

The standard signal for a drive error in capacity is zero size and zero
block size.  We have to take that case as standard without emitting
scary warnings.

James