* [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
@ 2023-08-17 5:03 kernel test robot
2023-08-17 6:09 ` Kees Cook
0 siblings, 1 reply; 6+ messages in thread
From: kernel test robot @ 2023-08-17 5:03 UTC (permalink / raw)
To: GONG, Ruiqi; +Cc: llvm, oe-kbuild-all, Florian Westphal, Kees Cook
tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
All warnings (new ones prefixed by >>):
In file included from <built-in>:1:
>> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
struct ebt_entry_target target;
^
1 warning generated.
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
2023-08-17 5:03 [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension kernel test robot
@ 2023-08-17 6:09 ` Kees Cook
2023-08-17 8:26 ` Florian Westphal
0 siblings, 1 reply; 6+ messages in thread
From: Kees Cook @ 2023-08-17 6:09 UTC (permalink / raw)
To: GONG, Ruiqi
Cc: kernel test robot, llvm, oe-kbuild-all, Florian Westphal,
linux-hardening
On Thu, Aug 17, 2023 at 01:03:20PM +0800, kernel test robot wrote:
> tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
> head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
> commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
> config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
> compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
> reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@intel.com>
> | Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
>
> All warnings (new ones prefixed by >>):
>
> In file included from <built-in>:1:
> >> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
> struct ebt_entry_target target;
> ^
> 1 warning generated.
Eww, it looks like "struct ebt_entry_target" is used _within_ another
struct:
struct ebt_standard_target {
struct ebt_entry_target target;
int verdict;
};
So "verdict" overlaps with the "data" FAM:
struct ebt_entry_target {
union {
struct {
char name[EBT_EXTENSION_MAXNAMELEN];
__u8 revision;
};
struct xt_target *target;
} u;
/* size of data */
unsigned int target_size;
unsigned char data[0] __attribute__ ((aligned (__alignof__(struct ebt_replace))));
};
These have been fixed in the past in a variety of ways -- it all depends
on how userspace is using them. In looking at Debian Code Search:
https://codesearch.debian.net/search?q=struct+ebt_standard_target&literal=1
It is exclusively doing casts and looking at the "verdict" member. So
the easiest conversion might be this:
struct ebt_standard_target {
- struct ebt_entry_target target;
+ unsigned char hdr[sizeof(struct ebt_entry_target)];
int verdict;
};
Or this might work (not tested):
diff --git a/include/uapi/linux/netfilter_bridge/ebtables.h b/include/uapi/linux/netfilter_bridge/ebtables.h
index a494cf43a755..d6f10163b14a 100644
--- a/include/uapi/linux/netfilter_bridge/ebtables.h
+++ b/include/uapi/linux/netfilter_bridge/ebtables.h
@@ -146,23 +146,25 @@ struct ebt_entry_watcher {
};
struct ebt_entry_target {
- union {
- struct {
- char name[EBT_EXTENSION_MAXNAMELEN];
- __u8 revision;
- };
- struct xt_target *target;
- } u;
- /* size of data */
- unsigned int target_size;
- unsigned char data[0] __attribute__ ((aligned (__alignof__(struct ebt_replace))));
+ struct ebt_entry_target_hdr {
+ union {
+ struct {
+ char name[EBT_EXTENSION_MAXNAMELEN];
+ __u8 revision;
+ };
+ struct xt_target *target;
+ } u;
+ /* size of data */
+ unsigned int target_size;
+ };
+ unsigned char data[] __attribute__ ((aligned (__alignof__(struct ebt_replace))));
};
#define EBT_STANDARD_TARGET "standard"
struct ebt_standard_target {
- struct ebt_entry_target target;
- int verdict;
+ struct ebt_entry_target_hdr target;
+ int verdict __attribute__ ((aligned (__alignof__(struct ebt_replace))));
};
/* one entry */
--
Kees Cook
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
2023-08-17 6:09 ` Kees Cook
@ 2023-08-17 8:26 ` Florian Westphal
2023-08-17 17:02 ` Kees Cook
0 siblings, 1 reply; 6+ messages in thread
From: Florian Westphal @ 2023-08-17 8:26 UTC (permalink / raw)
To: Kees Cook
Cc: GONG, Ruiqi, kernel test robot, llvm, oe-kbuild-all,
Florian Westphal, linux-hardening
Kees Cook <keescook@chromium.org> wrote:
> On Thu, Aug 17, 2023 at 01:03:20PM +0800, kernel test robot wrote:
> > tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
> > head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
> > commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
> > config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
> > compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
> > reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
> >
> > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > the same patch/commit), kindly add following tags
> > | Reported-by: kernel test robot <lkp@intel.com>
> > | Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
> >
> > All warnings (new ones prefixed by >>):
> >
> > In file included from <built-in>:1:
> > >> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
> > struct ebt_entry_target target;
> > ^
> > 1 warning generated.
>
> Eww, it looks like "struct ebt_entry_target" is used _within_ another
> struct:
>
> struct ebt_standard_target {
> struct ebt_entry_target target;
> int verdict;
> };
Yes, same as xt_standard_target.
> These have been fixed in the past in a variety of ways -- it all depends
> on how userspace is using them. In looking at Debian Code Search:
> https://codesearch.debian.net/search?q=struct+ebt_standard_target&literal=1
>
> It is exclusively doing casts and looking at the "verdict" member. So
> the easiest conversion might be this:
>
> struct ebt_standard_target {
> - struct ebt_entry_target target;
> + unsigned char hdr[sizeof(struct ebt_entry_target)];
> int verdict;
I don't think its worth doing all of this.
Can't we just keep it as-is and drop the relevant hunk from the patch?
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
2023-08-17 8:26 ` Florian Westphal
@ 2023-08-17 17:02 ` Kees Cook
2023-08-18 3:24 ` GONG, Ruiqi
0 siblings, 1 reply; 6+ messages in thread
From: Kees Cook @ 2023-08-17 17:02 UTC (permalink / raw)
To: Florian Westphal
Cc: GONG, Ruiqi, kernel test robot, llvm, oe-kbuild-all,
linux-hardening
On Thu, Aug 17, 2023 at 10:26:18AM +0200, Florian Westphal wrote:
> Kees Cook <keescook@chromium.org> wrote:
> > On Thu, Aug 17, 2023 at 01:03:20PM +0800, kernel test robot wrote:
> > > tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
> > > head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
> > > commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
> > > config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
> > > compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
> > > reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
> > >
> > > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > > the same patch/commit), kindly add following tags
> > > | Reported-by: kernel test robot <lkp@intel.com>
> > > | Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
> > >
> > > All warnings (new ones prefixed by >>):
> > >
> > > In file included from <built-in>:1:
> > > >> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
> > > struct ebt_entry_target target;
> > > ^
> > > 1 warning generated.
> >
> > Eww, it looks like "struct ebt_entry_target" is used _within_ another
> > struct:
> >
> > struct ebt_standard_target {
> > struct ebt_entry_target target;
> > int verdict;
> > };
>
> Yes, same as xt_standard_target.
>
> > These have been fixed in the past in a variety of ways -- it all depends
> > on how userspace is using them. In looking at Debian Code Search:
> > https://codesearch.debian.net/search?q=struct+ebt_standard_target&literal=1
> >
> > It is exclusively doing casts and looking at the "verdict" member. So
> > the easiest conversion might be this:
> >
> > struct ebt_standard_target {
> > - struct ebt_entry_target target;
> > + unsigned char hdr[sizeof(struct ebt_entry_target)];
> > int verdict;
>
> I don't think its worth doing all of this.
>
> Can't we just keep it as-is and drop the relevant hunk from the patch?
For now, yeah. But we'll need to find a solution as flex-array
structures overlapping variables is considered deprecated (it can lead
to ambiguous sizing results). But as long as the kernel itself doesn't
use struct ebt_standard_target, we can kick the can down the road a bit
more. :)
--
Kees Cook
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
2023-08-17 17:02 ` Kees Cook
@ 2023-08-18 3:24 ` GONG, Ruiqi
2023-08-18 5:23 ` GONG, Ruiqi
0 siblings, 1 reply; 6+ messages in thread
From: GONG, Ruiqi @ 2023-08-18 3:24 UTC (permalink / raw)
To: Kees Cook, Florian Westphal
Cc: kernel test robot, llvm, oe-kbuild-all, linux-hardening
On 2023/08/18 1:02, Kees Cook wrote:
> On Thu, Aug 17, 2023 at 10:26:18AM +0200, Florian Westphal wrote:
>> Kees Cook <keescook@chromium.org> wrote:
>>> On Thu, Aug 17, 2023 at 01:03:20PM +0800, kernel test robot wrote:
>>>> tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
>>>> head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
>>>> commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
>>>> config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
>>>> compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
>>>> reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
>>>>
>>>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>>>> the same patch/commit), kindly add following tags
>>>> | Reported-by: kernel test robot <lkp@intel.com>
>>>> | Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
>>>>
>>>> All warnings (new ones prefixed by >>):
>>>>
>>>> In file included from <built-in>:1:
>>>>>> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
>>>> struct ebt_entry_target target;
>>>> ^
>>>> 1 warning generated.
>>>
>>> Eww, it looks like "struct ebt_entry_target" is used _within_ another
>>> struct:
>>>
>>> struct ebt_standard_target {
>>> struct ebt_entry_target target;
>>> int verdict;
>>> };
>>
>> Yes, same as xt_standard_target.
>>
>>> These have been fixed in the past in a variety of ways -- it all depends
>>> on how userspace is using them. In looking at Debian Code Search:
>>> https://codesearch.debian.net/search?q=struct+ebt_standard_target&literal=1
>>>
>>> It is exclusively doing casts and looking at the "verdict" member. So
>>> the easiest conversion might be this:
>>>
>>> struct ebt_standard_target {
>>> - struct ebt_entry_target target;
>>> + unsigned char hdr[sizeof(struct ebt_entry_target)];
>>> int verdict;
>>
>> I don't think its worth doing all of this.
>>
>> Can't we just keep it as-is and drop the relevant hunk from the patch?
>
> For now, yeah. But we'll need to find a solution as flex-array
> structures overlapping variables is considered deprecated (it can lead
> to ambiguous sizing results). But as long as the kernel itself doesn't
> use struct ebt_standard_target, we can kick the can down the road a bit
> more. :)
>
I agree that we can complete the whole transformation step-by-step. I
will make another patch without the problematic replacement.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension
2023-08-18 3:24 ` GONG, Ruiqi
@ 2023-08-18 5:23 ` GONG, Ruiqi
0 siblings, 0 replies; 6+ messages in thread
From: GONG, Ruiqi @ 2023-08-18 5:23 UTC (permalink / raw)
To: Kees Cook, Florian Westphal
Cc: kernel test robot, llvm, oe-kbuild-all, linux-hardening
On 2023/08/18 11:24, GONG, Ruiqi wrote:
>
>
> On 2023/08/18 1:02, Kees Cook wrote:
>> On Thu, Aug 17, 2023 at 10:26:18AM +0200, Florian Westphal wrote:
>>> Kees Cook <keescook@chromium.org> wrote:
>>>> On Thu, Aug 17, 2023 at 01:03:20PM +0800, kernel test robot wrote:
>>>>> tree: git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git testing
>>>>> head: 015e2d9101d3713c7bee16dccad171df04a3bbd5
>>>>> commit: 61b9e6bd48a6317c0a44ee4f3fecdec9de5baa9e [2/9] netfilter: ebtables: replace zero-length array members
>>>>> config: i386-buildonly-randconfig-r004-20230817 (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/config)
>>>>> compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
>>>>> reproduce: (https://download.01.org/0day-ci/archive/20230817/202308171249.g1ywxhII-lkp@intel.com/reproduce)
>>>>>
>>>>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>>>>> the same patch/commit), kindly add following tags
>>>>> | Reported-by: kernel test robot <lkp@intel.com>
>>>>> | Closes: https://lore.kernel.org/oe-kbuild-all/202308171249.g1ywxhII-lkp@intel.com/
>>>>>
>>>>> All warnings (new ones prefixed by >>):
>>>>>
>>>>> In file included from <built-in>:1:
>>>>>>> ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension [-Wgnu-variable-sized-type-not-at-end]
>>>>> struct ebt_entry_target target;
>>>>> ^
>>>>> 1 warning generated.
>>>>
>>>> Eww, it looks like "struct ebt_entry_target" is used _within_ another
>>>> struct:
>>>>
>>>> struct ebt_standard_target {
>>>> struct ebt_entry_target target;
>>>> int verdict;
>>>> };
>>>
>>> Yes, same as xt_standard_target.
>>>
>>>> These have been fixed in the past in a variety of ways -- it all depends
>>>> on how userspace is using them. In looking at Debian Code Search:
>>>> https://codesearch.debian.net/search?q=struct+ebt_standard_target&literal=1
>>>>
>>>> It is exclusively doing casts and looking at the "verdict" member. So
>>>> the easiest conversion might be this:
>>>>
>>>> struct ebt_standard_target {
>>>> - struct ebt_entry_target target;
>>>> + unsigned char hdr[sizeof(struct ebt_entry_target)];
>>>> int verdict;
>>>
>>> I don't think its worth doing all of this.
>>>
>>> Can't we just keep it as-is and drop the relevant hunk from the patch?
>>
>> For now, yeah. But we'll need to find a solution as flex-array
>> structures overlapping variables is considered deprecated (it can lead
>> to ambiguous sizing results). But as long as the kernel itself doesn't
>> use struct ebt_standard_target, we can kick the can down the road a bit
>> more. :)
>>
>
> I agree that we can complete the whole transformation step-by-step. I
> will make another patch without the problematic replacement.
I didn't notice that the patch had been modified and merged into
nf-next's testing branch. Let's use that and I won't send another one.
Thanks Florian! XD
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2023-08-18 5:23 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-08-17 5:03 [netfilter-nf-next:testing 2/9] ./usr/include/linux/netfilter_bridge/ebtables.h:163:26: warning: field 'target' with variable sized type 'struct ebt_entry_target' not at the end of a struct or class is a GNU extension kernel test robot
2023-08-17 6:09 ` Kees Cook
2023-08-17 8:26 ` Florian Westphal
2023-08-17 17:02 ` Kees Cook
2023-08-18 3:24 ` GONG, Ruiqi
2023-08-18 5:23 ` GONG, Ruiqi
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox