Re: [net-next PATCH 2/2] bpf: sock_map fixes for !CONFIG_BPF_SYSCALL and !STREAM_PARSER

[John Fastabend <john.fastabend@gmail.com>]

On 08/16/2017 03:06 PM, David Ahern wrote:
> On 8/16/17 4:02 PM, John Fastabend wrote:
>> diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile
>> index aa24287..897daa0 100644
>> --- a/kernel/bpf/Makefile
>> +++ b/kernel/bpf/Makefile
>> @@ -3,7 +3,10 @@ obj-y := core.o
>>  obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o
>>  obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o
>>  ifeq ($(CONFIG_NET),y)
>> -obj-$(CONFIG_BPF_SYSCALL) += devmap.o sockmap.o
>> +obj-$(CONFIG_BPF_SYSCALL) += devmap.o
>> +ifeq ($(CONFIG_STREAM_PARSER),y)
>> +obj-$(CONFIG_BPF_SYSCALL) += sockmap.o
>> +endif
>>  endif
> 
> sockmap requires KCM? I thought it just needed STREAM_PARSER. Can't
> STREAM_PARSER be used outside of KCM? If so why not make STREAM_PARSER
> enabled if sockmap is wanted vs requiring KCM to be compiled in to get
> stream parser to get sockmap?
> 

KCM is not required all sockmap needs is STREAM_PARSER. STREAM_PARSER can
be used outside of KCM so no problem there.

I didn't want to require all users of BPF maps to automatically get
STREAM_PARSER however. I'll add a Kconfig option for STREAM_PARSER so we
can pull it in without KCM easily. This is very similar to how the cgroup
BPF build logic works as well with CONFIG_CGROUP_BPF.

I wanted to get the build fix out though as soon as possible rather than
wait for me to write the STREAM_PARSER patch and test it.

Thanks,
John