| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Wed, 12 Jun 2019 23:33:17 +0200
From: Toke Høiland-Jørgensen <toke@...hat.com>
To: Maciej Fijalkowski <maciejromanfijalkowski@...il.com>
Cc: Jesper Dangaard Brouer <brouer@...hat.com>, netdev@...r.kernel.org,
Daniel Borkmann <daniel@...earbox.net>,
Alexei Starovoitov <ast@...nel.org>,
David Miller <davem@...emloft.net>,
Jonathan Lemon <jonathan.lemon@...il.com>
Subject: Re: [PATCH bpf-next v3 2/3] bpf_xdp_redirect_map: Perform map lookup in eBPF helper
Maciej Fijalkowski <maciejromanfijalkowski@...il.com> writes:
> On Tue, 11 Jun 2019 20:17:02 +0200
> Toke Høiland-Jørgensen <toke@...hat.com> wrote:
>
>> Jesper Dangaard Brouer <brouer@...hat.com> writes:
>>
>> > On Tue, 11 Jun 2019 17:44:00 +0200
>> > Toke Høiland-Jørgensen <toke@...hat.com> wrote:
>> >
>> >> From: Toke Høiland-Jørgensen <toke@...hat.com>
>> >>
>> >> The bpf_redirect_map() helper used by XDP programs doesn't return any
>> >> indication of whether it can successfully redirect to the map index it was
>> >> given. Instead, BPF programs have to track this themselves, leading to
>> >> programs using duplicate maps to track which entries are populated in the
>> >> devmap.
>> >>
>> >> This patch fixes this by moving the map lookup into the bpf_redirect_map()
>> >> helper, which makes it possible to return failure to the eBPF program. The
>> >> lower bits of the flags argument is used as the return code, which means
>> >> that existing users who pass a '0' flag argument will get XDP_ABORTED.
>> >>
>> >> With this, a BPF program can check the return code from the helper call and
>> >> react by, for instance, substituting a different redirect. This works for
>> >> any type of map used for redirect.
>> >>
>> >> Signed-off-by: Toke Høiland-Jørgensen <toke@...hat.com>
>> >> ---
>> >> include/linux/filter.h | 1 +
>> >> include/uapi/linux/bpf.h | 8 ++++++++
>> >> net/core/filter.c | 26 ++++++++++++--------------
>> >> 3 files changed, 21 insertions(+), 14 deletions(-)
>> >>
>> >> diff --git a/include/linux/filter.h b/include/linux/filter.h
>> >> index 43b45d6db36d..f31ae8b9035a 100644
>> >> --- a/include/linux/filter.h
>> >> +++ b/include/linux/filter.h
>> >> @@ -580,6 +580,7 @@ struct bpf_skb_data_end {
>> >> struct bpf_redirect_info {
>> >> u32 ifindex;
>> >> u32 flags;
>> >> + void *item;
>> >> struct bpf_map *map;
>> >> struct bpf_map *map_to_flush;
>> >> u32 kern_flags;
>> >> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
>> >> index 7c6aef253173..9931cf02de19 100644
>> >> --- a/include/uapi/linux/bpf.h
>> >> +++ b/include/uapi/linux/bpf.h
>> >> @@ -3098,6 +3098,14 @@ enum xdp_action {
>> >> XDP_REDIRECT,
>> >> };
>> >>
>> >> +/* Flags for bpf_xdp_redirect_map helper */
>> >> +
>> >> +/* The lower flag bits will be the return code of bpf_xdp_redirect_map() helper
>> >> + * if the map lookup fails.
>> >> + */
>> >> +#define XDP_REDIRECT_INVALID_MASK (XDP_ABORTED | XDP_DROP | XDP_PASS | XDP_TX)
>> >> +#define XDP_REDIRECT_ALL_FLAGS XDP_REDIRECT_INVALID_MASK
>> >> +
>> >
>> > Slightly confused about the naming of the define, see later.
>> >
>> >> /* user accessible metadata for XDP packet hook
>> >> * new fields must be added to the end of this structure
>> >> */
>> >> diff --git a/net/core/filter.c b/net/core/filter.c
>> >> index 7a996887c500..dd43be497480 100644
>> >> --- a/net/core/filter.c
>> >> +++ b/net/core/filter.c
>> >> @@ -3608,17 +3608,13 @@ static int xdp_do_redirect_map(struct net_device *dev, struct xdp_buff *xdp,
>> >> struct bpf_redirect_info *ri)
>> >> {
>> >> u32 index = ri->ifindex;
>> >> - void *fwd = NULL;
>> >> + void *fwd = ri->item;
>> >> int err;
>> >>
>> >> ri->ifindex = 0;
>> >> + ri->item = NULL;
>> >> WRITE_ONCE(ri->map, NULL);
>> >>
>> >> - fwd = __xdp_map_lookup_elem(map, index);
>> >> - if (unlikely(!fwd)) {
>> >> - err = -EINVAL;
>> >> - goto err;
>> >> - }
>> >> if (ri->map_to_flush && unlikely(ri->map_to_flush != map))
>> >> xdp_do_flush_map();
>> >>
>> >> @@ -3655,18 +3651,13 @@ static int xdp_do_generic_redirect_map(struct net_device *dev,
>> >> {
>> >> struct bpf_redirect_info *ri = this_cpu_ptr(&bpf_redirect_info);
>> >> u32 index = ri->ifindex;
>> >> - void *fwd = NULL;
>> >> + void *fwd = ri->item;
>> >> int err = 0;
>> >>
>> >> ri->ifindex = 0;
>> >> + ri->item = NULL;
>> >> WRITE_ONCE(ri->map, NULL);
>> >>
>> >> - fwd = __xdp_map_lookup_elem(map, index);
>> >> - if (unlikely(!fwd)) {
>> >> - err = -EINVAL;
>> >> - goto err;
>> >> - }
>> >> -
>> >> if (map->map_type == BPF_MAP_TYPE_DEVMAP) {
>> >> struct bpf_dtab_netdev *dst = fwd;
>> >>
>> >> @@ -3735,6 +3726,7 @@ BPF_CALL_2(bpf_xdp_redirect, u32, ifindex, u64, flags)
>> >>
>> >> ri->ifindex = ifindex;
>> >> ri->flags = flags;
>> >> + ri->item = NULL;
>> >> WRITE_ONCE(ri->map, NULL);
>> >>
>> >> return XDP_REDIRECT;
>> >> @@ -3753,9 +3745,15 @@ BPF_CALL_3(bpf_xdp_redirect_map, struct bpf_map *, map, u32, ifindex,
>> >> {
>> >> struct bpf_redirect_info *ri = this_cpu_ptr(&bpf_redirect_info);
>> >>
>> >> - if (unlikely(flags))
>> >> + if (unlikely(flags & ~XDP_REDIRECT_ALL_FLAGS))
>> >> return XDP_ABORTED;
>> >>
>
> Here you don't allow the flags to get different value than
> XDP_REDIRECT_ALL_FLAGS.
>
>> >> + ri->item = __xdp_map_lookup_elem(map, ifindex);
>> >> + if (unlikely(!ri->item)) {
>> >> + WRITE_ONCE(ri->map, NULL);
>> >> + return (flags & XDP_REDIRECT_INVALID_MASK);
>
> So here you could just return flags? Don't we know that the flags
> value is legit here? Am I missing something?
No, you're right. I was just worried that we would forget to change this
when we add another flag later on, so I thought I'd mask it from the
get-go. Can't the compiler figure out that the second mask is unnecessary?
> TBH the v2 was more clear to me.
Clearer how? As in, you prefer just always returning XDP_PASS on error?
-Toke
>> > Maybe I'm reading it wrong, but shouldn't the mask be called the "valid" mask?
>>
>> It's the mask that is applied when the index looked up is invalid (i.e.,
>> the entry doesn't exist)? But yeah, can see how the name can be
>> confusing; maybe it should just be "RETURN_MASK" or something like that?
>
> Maybe something along ALLOWED_RETVAL_MASK?
Yeah, good idea! :)
-Toke
Powered by blists - more mailing lists