Re: [PATCH bpf-next 2/4] selftests/bpf: convert test_cgroup_storage to test_progs

[Alan Maguire]

On 01/08/2024 10:21, Alexis Lothoré wrote:
> On 8/1/24 10:27, Alan Maguire wrote:
>> On 31/07/2024 11:38, Alexis Lothoré (eBPF Foundation) wrote:
>>> test_cgroup_storage is currently a standalone program which is not run
>>> when executing test_progs.
>>>
>>> Convert it to the test_progs framework so it can be automatically executed
>>> in CI. The conversion led to the following changes:
>>> - converted the raw bpf program in the userspace test file into a dedicated
>>>   test program in progs/ dir
>>> - reduced the scope of cgroup_storage test: the content from this test
>>>   overlaps with some other tests already present in test_progs, most
>>>   notably netcnt and cgroup_storage_multi*. Those tests already check
>>>   extensively local storage, per-cpu local storage, cgroups interaction,
>>>   etc. So the new test only keep the part testing that the program return
>>>   code (based on map content) properly leads to packet being passed or
>>>   dropped.
>>>
>>> Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@xxxxxxxxxxx>
>>
>> Two small things below, but
>>
>> Reviewed-by: Alan Maguire <alan.maguire@xxxxxxxxxx>
> 
> [...]
> 
>>> +#define PING_CMD "ping localhost -c 1 -W 1 -q"
>>
>> other tests seem to redirect ping stdout output to /dev/null ; might be
>> worth doing that too.
> 
> That's in fact performed automatically by SYS_NOFAIL :)
> > #define SYS_NOFAIL(fmt, ...)						\
> 	({								\
> 		char cmd[1024];						\
> 		int n;							\
> 		n = snprintf(cmd, sizeof(cmd), fmt, ##__VA_ARGS__);	\
> 		if (n < sizeof(cmd) && sizeof(cmd) - n >= sizeof(ALL_TO_DEV_NULL)) \
> 			strcat(cmd, ALL_TO_DEV_NULL);			\
> 		system(cmd);						\
> 	})
> 
> [...]
> 

Perfect, I missed that.

>>> +{
>>> +	__u64 *counter;
>>> +
>>> +	counter = bpf_get_local_storage(&cgroup_storage, 0);
>>
>> don't we need a NULL check for counter here? Or does the verifier know
>> bpf_get_local_storage never fails?
> 
> Good question. Since the verifier accepted the prog during my tests, I indeed
> assume that the returned pointer is always valid. Amongst all calls to this
> function in progs involved in selftests, I found only one performing a check
> before using the value (lsm_cgroup.c). So I guess it is fine ?
> 

Looks like the prototype for the helper specifies a return type of
RET_PTR_TO_MAP_VALUE ; if it was RET_PTR_TO_MAP_VALUE_OR_NULL we'd need
the NULL check, but because it's a guaranteed map ptr we are good here
without a NULL check.

Thanks!

Alan