| Message ID | 20200728190527.110830-1-andriin@fb.com |
|---|---|
| State | Accepted |
| Delegated to: | BPF Maintainers |
| Headers | show |
| Series | [bpf-next] bpf: fix build without CONFIG_NET when using BPF XDP link | expand |
On 7/28/20 12:05 PM, Andrii Nakryiko wrote: > Entire net/core subsystem is not built without CONFIG_NET. linux/netdevice.h > just assumes that it's always there, so the easiest way to fix this is to > conditionally compile out bpf_xdp_link_attach() use in bpf/syscall.c. > > Reported-by: Randy Dunlap <rdunlap@infradead.org> > Fixes: aa8d3a716b59 ("bpf, xdp: Add bpf_link-based XDP attachment API") > Signed-off-by: Andrii Nakryiko <andriin@fb.com> Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested Thanks. > --- > kernel/bpf/syscall.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c > index 0e8c88db7e7a..cd3d599e9e90 100644 > --- a/kernel/bpf/syscall.c > +++ b/kernel/bpf/syscall.c > @@ -3923,9 +3923,11 @@ static int link_create(union bpf_attr *attr) > case BPF_PROG_TYPE_SK_LOOKUP: > ret = netns_bpf_link_create(attr, prog); > break; > +#ifdef CONFIG_NET > case BPF_PROG_TYPE_XDP: > ret = bpf_xdp_link_attach(attr, prog); > break; > +#endif > default: > ret = -EINVAL; > } >
On Tue, Jul 28, 2020 at 1:08 PM Randy Dunlap <rdunlap@infradead.org> wrote: > > On 7/28/20 12:05 PM, Andrii Nakryiko wrote: > > Entire net/core subsystem is not built without CONFIG_NET. linux/netdevice.h > > just assumes that it's always there, so the easiest way to fix this is to > > conditionally compile out bpf_xdp_link_attach() use in bpf/syscall.c. > > > > Reported-by: Randy Dunlap <rdunlap@infradead.org> > > Fixes: aa8d3a716b59 ("bpf, xdp: Add bpf_link-based XDP attachment API") > > Signed-off-by: Andrii Nakryiko <andriin@fb.com> > > Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested Acked-by: Song Liu <songliubraving@fb.com>
On 7/28/20 9:05 PM, Andrii Nakryiko wrote: > Entire net/core subsystem is not built without CONFIG_NET. linux/netdevice.h > just assumes that it's always there, so the easiest way to fix this is to > conditionally compile out bpf_xdp_link_attach() use in bpf/syscall.c. > > Reported-by: Randy Dunlap <rdunlap@infradead.org> > Fixes: aa8d3a716b59 ("bpf, xdp: Add bpf_link-based XDP attachment API") > Signed-off-by: Andrii Nakryiko <andriin@fb.com> Applied, thanks!
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 0e8c88db7e7a..cd3d599e9e90 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -3923,9 +3923,11 @@ static int link_create(union bpf_attr *attr) case BPF_PROG_TYPE_SK_LOOKUP: ret = netns_bpf_link_create(attr, prog); break; +#ifdef CONFIG_NET case BPF_PROG_TYPE_XDP: ret = bpf_xdp_link_attach(attr, prog); break; +#endif default: ret = -EINVAL; }
Entire net/core subsystem is not built without CONFIG_NET. linux/netdevice.h just assumes that it's always there, so the easiest way to fix this is to conditionally compile out bpf_xdp_link_attach() use in bpf/syscall.c. Reported-by: Randy Dunlap <rdunlap@infradead.org> Fixes: aa8d3a716b59 ("bpf, xdp: Add bpf_link-based XDP attachment API") Signed-off-by: Andrii Nakryiko <andriin@fb.com> --- kernel/bpf/syscall.c | 2 ++ 1 file changed, 2 insertions(+)