xen: Fix off-by-one error when parsing command line arguments

xen: Fix off-by-one error when parsing command line arguments

[Andrew Cooper]

[-- Attachment #1: Type: text/plain, Size: 261 bytes --]

This is a bug which is present in at least unstable and 4.1, so should
be considered for 4.2, as well as being considered for backport to older
versions.

-- 
Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer
T: +44 (0)1223 225 900, http://www.citrix.com


[-- Attachment #2: xen-fix-cmdline-parsing.patch --]
[-- Type: text/x-patch, Size: 1239 bytes --]

# HG changeset patch
# Parent 4f92bdf3370c4fe5ed0f00cdeaf8156e4818ecb5
xen: Fix off-by-one error when parsing command line arguments

As Xen currently stands, it will attempt to interpret the first few bytes of the
initcall section as a struct kernel_param.

This can be verified as

         for ( param = &__setup_start; param <= &__setup_end; param++ )
         {
+            if ( (unsigned long)param == (unsigned long)&__initcall_start )
+                BUG();

causes Xen to BUG() during early boot.

The reason that this not caused problems is because in the overflow case,
param->name is actually a function pointer to the first initcall, and
intepreting it as string is very unlikely to match an ASCII command line
parameter name.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>

diff -r 4f92bdf3370c xen/common/kernel.c
--- a/xen/common/kernel.c
+++ b/xen/common/kernel.c
@@ -90,7 +90,7 @@ void __init cmdline_parse(const char *cm
         if ( !bool_assert )
             optkey += 3;
 
-        for ( param = &__setup_start; param <= &__setup_end; param++ )
+        for ( param = &__setup_start; param < &__setup_end; param++ )
         {
             if ( strcmp(param->name, optkey) )
                 continue;

[-- Attachment #3: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: xen: Fix off-by-one error when parsing command line arguments

[Jan Beulich]

>>> On 02.07.12 at 14:52, Andrew Cooper <andrew.cooper3@citrix.com> wrote:
> This is a bug which is present in at least unstable and 4.1, so should
> be considered for 4.2, as well as being considered for backport to older
> versions.

Indeed, the problem goes back to at least 3.2.0.

Keir, if you would ack this I could push it into -unstable at least
(pulling over into 4.1 and 4.0 would still need to be done later
by you).

Jan