[PATCH] crash in 2.6.22-git2 sysctl_set_parent()

This is a patch (& bug report) for a crash in sysctl_set_parent() 
in 2.6.22-git2. 

Problem: 2.6.22-git2 crashes with a stack trace 
[c000000001d0fb00] c000000000067b4c .sysctl_set_parent+0x48/0x7c
[c000000001d0fb90] c000000000069b40 .register_sysctl_table+0x7c/0xf4
[c000000001d0fc30] c00000000065e710 .devinet_init+0x88/0xb0
[c000000001d0fcc0] c00000000065db74 .ip_rt_init+0x17c/0x32c
[c000000001d0fd70] c00000000065deec .ip_init+0x10/0x34
[c000000001d0fdf0] c00000000065e898 .inet_init+0x160/0x3dc
[c000000001d0fea0] c000000000630bc4 .kernel_init+0x204/0x3c8

A bit of poking around makes it clear what the problem is:
In sysctl_set_parent(), the for loop 

   for (; table->ctl_name || table->procname; table++) {

walks off the end of the table, and into garbage.  Basically,
this for-loop iterator expects all table arrays to be 
"null terminated".  However, net/ipv4/devinet.c statically 
declares an array that is not null-terminated.  The patch 
below fixes that; it works for me.  Its somewhat conservative;
if one wishes to assume that the compiler will always zero out
the empty parts of the structure, then this pach can be shrunk 
to one line: +	ctl_table		devinet_root_dir[3];

Signed-off-by: Linas Vepstas <redacted>

----
I tried to audit some of the code to see where else there 
might be similar badly-formed static declarations.  This is hard,
as there's a lot of code. Most seems fine.


 net/core/neighbour.c |    4 ++++
 net/ipv4/devinet.c   |    7 ++++++-
 2 files changed, 10 insertions(+), 1 deletion(-)

Index: linux-2.6.22-git2/net/ipv4/devinet.c
===================================================================

--- linux-2.6.22-git2.orig/net/ipv4/devinet.c	2007-07-13 14:23:21.000000000 -0500
+++ linux-2.6.22-git2/net/ipv4/devinet.c	2007-07-13 14:24:15.000000000 -0500

@@ -1424,7 +1424,7 @@ static struct devinet_sysctl_table {
 	ctl_table		devinet_dev[2];
 	ctl_table		devinet_conf_dir[2];
 	ctl_table		devinet_proto_dir[2];
-	ctl_table		devinet_root_dir[2];
+	ctl_table		devinet_root_dir[3];
 } devinet_sysctl = {
 	.devinet_vars = {
 		DEVINET_SYSCTL_COMPLEX_ENTRY(FORWARDING, "forwarding",

@@ -1493,8 +1493,13 @@ static struct devinet_sysctl_table {
 			.data		= &ipv4_devconf.loop,
 			.maxlen		= sizeof(int),
 			.mode		= 0644,
+			.child	= 0x0,
 			.proc_handler	= &proc_dointvec,
 		},
+		{
+			.ctl_name	= 0,
+			.procname	= 0,
+		},
 	},
 };
 

From: linas@austin.ibm.com (Linas Vepstas)
Date: Fri, 13 Jul 2007 15:05:15 -0500

This is a patch (& bug report) for a crash in sysctl_set_parent() 
in 2.6.22-git2. 

Problem: 2.6.22-git2 crashes with a stack trace 
[c000000001d0fb00] c000000000067b4c .sysctl_set_parent+0x48/0x7c
[c000000001d0fb90] c000000000069b40 .register_sysctl_table+0x7c/0xf4
[c000000001d0fc30] c00000000065e710 .devinet_init+0x88/0xb0
[c000000001d0fcc0] c00000000065db74 .ip_rt_init+0x17c/0x32c
[c000000001d0fd70] c00000000065deec .ip_init+0x10/0x34
[c000000001d0fdf0] c00000000065e898 .inet_init+0x160/0x3dc
[c000000001d0fea0] c000000000630bc4 .kernel_init+0x204/0x3c8

A bit of poking around makes it clear what the problem is:
In sysctl_set_parent(), the for loop 

   for (; table->ctl_name || table->procname; table++) {

walks off the end of the table, and into garbage.  Basically,
this for-loop iterator expects all table arrays to be 
"null terminated".  However, net/ipv4/devinet.c statically 
declares an array that is not null-terminated.  The patch 
below fixes that; it works for me.  Its somewhat conservative;
if one wishes to assume that the compiler will always zero out
the empty parts of the structure, then this pach can be shrunk 
to one line: +	ctl_table		devinet_root_dir[3];

Signed-off-by: Linas Vepstas <redacted>

Thanks for tracking this down, I'll apply your patch.

Hi,

I'm totally confuzed by this patch ...

On 7/14/07, Linas Vepstas [off-list ref] wrote:

This is a patch (& bug report) for a crash in sysctl_set_parent()
in 2.6.22-git2.

Are you sure you saw this crash on 22-git2 (Linus' tree) ???

Problem: 2.6.22-git2 crashes with a stack trace
[c000000001d0fb00] c000000000067b4c .sysctl_set_parent+0x48/0x7c
[c000000001d0fb90] c000000000069b40 .register_sysctl_table+0x7c/0xf4
[c000000001d0fc30] c00000000065e710 .devinet_init+0x88/0xb0
[c000000001d0fcc0] c00000000065db74 .ip_rt_init+0x17c/0x32c
[c000000001d0fd70] c00000000065deec .ip_init+0x10/0x34
[c000000001d0fdf0] c00000000065e898 .inet_init+0x160/0x3dc
[c000000001d0fea0] c000000000630bc4 .kernel_init+0x204/0x3c8

Please post the full dmesg output, if possible.

A bit of poking around makes it clear what the problem is:
In sysctl_set_parent(), the for loop

   for (; table->ctl_name || table->procname; table++) {

Yup, but note that the "table" there is a struct ctl_table * ...

walks off the end of the table, and into garbage.  Basically,
this for-loop iterator expects all table arrays to be
"null terminated".  However, net/ipv4/devinet.c statically
declares an array that is not null-terminated.

Whereas the "not null-terminated" array you're talking about
in devinet.c is a struct devinet_sysctl_table, whose *members*
are ctl_table structs.

Also note that all those members have already been declared
as arrays: struct ctl_table xxx[2]; *precisely* because the
second element (which is left un-initialized, hence will get
zeroed-out being static) is what will protect us from running
out into garbage in sysctl_set_parent().

The patch
below fixes that; it works for me.

Now I'm even more confuzed ...

Its somewhat conservative;
if one wishes to assume that the compiler will always zero out
the empty parts of the structure,

You can always safely assume that.

Off-topic, but note that unintialized static variables being automagically
initialized to zero / NULL is guaranteed by C (the language) itself, so
nothing left in the hands of compiler implementations here. [6.7.8:10]

then this pach can be shrunk
to one line: +  ctl_table               devinet_root_dir[3];

But that's pointless. It was already [2] for the reason I mentioned.

Signed-off-by: Linas Vepstas <redacted>
[...]
 net/core/neighbour.c |    4 ++++
 net/ipv4/devinet.c   |    7 ++++++-
 2 files changed, 10 insertions(+), 1 deletion(-)

Weirder. diffstat does not match the patch ...

And Linus' latest -git doesn't even have this stuff at line 1493.

                        .data           = &ipv4_devconf.loop,
                        .maxlen         = sizeof(int),
                        .mode           = 0644,
+                       .child  = 0x0,
                        .proc_handler   = &proc_dointvec,
                },
+               {
+                       .ctl_name       = 0,
+                       .procname       = 0,
+               },
        },
 };

Well, as I said, I'm totally confuzed ...

Satyam

On Fri, Jul 13, 2007 at 03:47:02PM -0700, David Miller wrote:

From: linas@austin.ibm.com (Linas Vepstas)
Date: Fri, 13 Jul 2007 15:05:15 -0500

quoted

This is a patch (& bug report) for a crash in sysctl_set_parent() 
in 2.6.22-git2. 

Problem: 2.6.22-git2 crashes with a stack trace 

Signed-off-by: Linas Vepstas <redacted>

Thanks for tracking this down, I'll apply your patch.

NAK. As I just explained in another email, this bug
was introduced by the "send-to-self" patch I habitually
apply -- so habitually, that I forgot I was not working 
with a "clean" tree. So it goes ... 

--linas

From: linas@austin.ibm.com (Linas Vepstas)
Date: Mon, 16 Jul 2007 12:25:35 -0500

On Fri, Jul 13, 2007 at 03:47:02PM -0700, David Miller wrote:

quoted

From: linas@austin.ibm.com (Linas Vepstas)
Date: Fri, 13 Jul 2007 15:05:15 -0500

quoted

This is a patch (& bug report) for a crash in sysctl_set_parent() 
in 2.6.22-git2. 

Problem: 2.6.22-git2 crashes with a stack trace 

Signed-off-by: Linas Vepstas <redacted>

Thanks for tracking this down, I'll apply your patch.

NAK. As I just explained in another email, this bug
was introduced by the "send-to-self" patch I habitually
apply -- so habitually, that I forgot I was not working 
with a "clean" tree. So it goes ... 

I saw all of this, thanks.