[PATCH] kconfig: Avoid buffer underrun in choice input

[PATCH] kconfig: Avoid buffer underrun in choice input

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1268 bytes --]

commit 40aee729b350672c2550640622416a855e27938f ('kconfig: fix default
value for choice input') fixed some cases where kconfig would select
the wrong option from a choice with a single valid option and thus
enter an infinite loop.

However, this broke the test for user input of the form 'N?', because
when kconfig selects the single valid option the input is zero-length
and the test will read the byte before the input buffer.  If this
happens to contain '?' (as it will in a mips build on Debian unstable
today) then kconfig again enters an infinite loop.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Cc: stable@kernel.org [2.6.17+]
---
Roman has failed to respond to this after 5 weeks and one reminder, so
please take it directly.

Ben.

 scripts/kconfig/conf.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/scripts/kconfig/conf.c b/scripts/kconfig/conf.c
index 659326c..006ad81 100644
--- a/scripts/kconfig/conf.c
+++ b/scripts/kconfig/conf.c
@@ -332,7 +332,7 @@ static int conf_choice(struct menu *menu)
 		}
 		if (!child)
 			continue;
-		if (line[strlen(line) - 1] == '?') {
+		if (line[0] && line[strlen(line) - 1] == '?') {
 			print_help(child);
 			continue;
 		}
-- 
1.7.4.1




[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Sam Ravnborg]

On Sat, Apr 23, 2011 at 06:42:56PM +0100, Ben Hutchings wrote:
> commit 40aee729b350672c2550640622416a855e27938f ('kconfig: fix default
> value for choice input') fixed some cases where kconfig would select
> the wrong option from a choice with a single valid option and thus
> enter an infinite loop.
> 
> However, this broke the test for user input of the form 'N?', because
> when kconfig selects the single valid option the input is zero-length
> and the test will read the byte before the input buffer.  If this
> happens to contain '?' (as it will in a mips build on Debian unstable
> today) then kconfig again enters an infinite loop.
> 
> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> Cc: stable@kernel.org [2.6.17+]
> ---
> Roman has failed to respond to this after 5 weeks and one reminder, so
> please take it directly.

No-one has heard from Roman Zippel in several years.
I have suggested Michal Marek to take over the maintainer
role - as he in practice is the kconfig maintainer.

	Sam

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1331 bytes --]

On Sat, 2011-04-23 at 20:04 +0200, Sam Ravnborg wrote:
> On Sat, Apr 23, 2011 at 06:42:56PM +0100, Ben Hutchings wrote:
> > commit 40aee729b350672c2550640622416a855e27938f ('kconfig: fix default
> > value for choice input') fixed some cases where kconfig would select
> > the wrong option from a choice with a single valid option and thus
> > enter an infinite loop.
> > 
> > However, this broke the test for user input of the form 'N?', because
> > when kconfig selects the single valid option the input is zero-length
> > and the test will read the byte before the input buffer.  If this
> > happens to contain '?' (as it will in a mips build on Debian unstable
> > today) then kconfig again enters an infinite loop.
> > 
> > Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> > Cc: stable@kernel.org [2.6.17+]
> > ---
> > Roman has failed to respond to this after 5 weeks and one reminder, so
> > please take it directly.
> 
> No-one has heard from Roman Zippel in several years.
> I have suggested Michal Marek to take over the maintainer
> role - as he in practice is the kconfig maintainer.

Can someone please update MAINTAINERS then?  What is the point of it if
people don't update it for years?

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Geert Uytterhoeven]

On Sun, Apr 24, 2011 at 06:02, Ben Hutchings <ben@decadent.org.uk> wrote:
> On Sat, 2011-04-23 at 20:04 +0200, Sam Ravnborg wrote:
>> On Sat, Apr 23, 2011 at 06:42:56PM +0100, Ben Hutchings wrote:
>> > Roman has failed to respond to this after 5 weeks and one reminder, so
>> > please take it directly.
>>
>> No-one has heard from Roman Zippel in several years.
>> I have suggested Michal Marek to take over the maintainer
>> role - as he in practice is the kconfig maintainer.
>
> Can someone please update MAINTAINERS then?  What is the point of it if
> people don't update it for years?

I have removed him from AFFS, HFS, and m68k.
I left the kbuild entry for the kbuild people.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Michal Marek]

On 23.4.2011 19:42, Ben Hutchings wrote:
> commit 40aee729b350672c2550640622416a855e27938f ('kconfig: fix default
> value for choice input') fixed some cases where kconfig would select
> the wrong option from a choice with a single valid option and thus
> enter an infinite loop.
>
> However, this broke the test for user input of the form 'N?', because
> when kconfig selects the single valid option the input is zero-length
> and the test will read the byte before the input buffer.  If this
> happens to contain '?' (as it will in a mips build on Debian unstable
> today) then kconfig again enters an infinite loop.
>
> Signed-off-by: Ben Hutchings<ben@decadent.org.uk>
> Cc: stable@kernel.org [2.6.17+]
> ---
> Roman has failed to respond to this after 5 weeks and one reminder, so
> please take it directly.

I applied this on 8th April, see
http://www.spinics.net/lists/linux-kbuild/msg04431.html. Please check 
linux-next before reposting patches next time, now I either have to 
rewind the kconfig branch or let Linus merge it with a duplicate commit :-(.

Michal

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1409 bytes --]

On Wed, 2011-04-27 at 13:19 +0200, Michal Marek wrote:
> On 23.4.2011 19:42, Ben Hutchings wrote:
> > commit 40aee729b350672c2550640622416a855e27938f ('kconfig: fix default
> > value for choice input') fixed some cases where kconfig would select
> > the wrong option from a choice with a single valid option and thus
> > enter an infinite loop.
> >
> > However, this broke the test for user input of the form 'N?', because
> > when kconfig selects the single valid option the input is zero-length
> > and the test will read the byte before the input buffer.  If this
> > happens to contain '?' (as it will in a mips build on Debian unstable
> > today) then kconfig again enters an infinite loop.
> >
> > Signed-off-by: Ben Hutchings<ben@decadent.org.uk>
> > Cc: stable@kernel.org [2.6.17+]
> > ---
> > Roman has failed to respond to this after 5 weeks and one reminder, so
> > please take it directly.
> 
> I applied this on 8th April, see
> http://www.spinics.net/lists/linux-kbuild/msg04431.html.

Sorry, I forgot that.

> Please check 
> linux-next before reposting patches next time, now I either have to 
> rewind the kconfig branch or let Linus merge it with a duplicate commit :-(.

But the fix belongs in this release (2.6.39), not the next.  So I looked
in Linus' tree.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

Re: [PATCH] kconfig: Avoid buffer underrun in choice input

[Linus Torvalds]

On Wed, Apr 27, 2011 at 4:19 AM, Michal Marek <mmarek@suse.cz> wrote:
>
> I applied this on 8th April, see
> http://www.spinics.net/lists/linux-kbuild/msg04431.html. Please check
> linux-next before reposting patches next time, now I either have to rewind
> the kconfig branch or let Linus merge it with a duplicate commit :-(.

Merging the occasional duplicate is normal flow, don't worry about it.
I'd much rather see the occasional smaller diffstat due to a diff that
got applied in both sides than I'd see people rebasing just to avoid
some silly duplicate.

It's if two branches consistently have many duplicates that I react,
because it tends to be some systematic error (often excessive
rebasing, but it could be some "ownership" argument too)

The "occasionally the same fix made it through two different trees" is
not just normal, it's _expected_ in distributed development. If it
never happens, that's a sign that people are serializing overmuch.

                                   Linus