From: Toke Høiland-Jørgensen <toke@xxxxxxxxxx>
Date: Thu, 09 Feb 2023 21:04:38 +0100
> Alexander Lobakin <alexandr.lobakin@xxxxxxxxx> writes:
>
>> &xdp_buff and &xdp_frame are bound in a way that
>>
>> xdp_buff->data_hard_start == xdp_frame
>>
>> It's always the case and e.g. xdp_convert_buff_to_frame() relies on
>> this.
>> IOW, the following:
>>
>> for (u32 i = 0; i < 0xdead; i++) {
>> xdpf = xdp_convert_buff_to_frame(&xdp);
>> xdp_convert_frame_to_buff(xdpf, &xdp);
>> }
>>
>> shouldn't ever modify @xdpf's contents or the pointer itself.
>> However, "live packet" code wrongly treats &xdp_frame as part of its
>> context placed *before* the data_hard_start. With such flow,
>> data_hard_start is sizeof(*xdpf) off to the right and no longer points
>> to the XDP frame.
>
> Oh, nice find!
>
>> Instead of replacing `sizeof(ctx)` with `offsetof(ctx, xdpf)` in several
>> places and praying that there are no more miscalcs left somewhere in the
>> code, unionize ::frm with ::data in a flex array, so that both starts
>> pointing to the actual data_hard_start and the XDP frame actually starts
>> being a part of it, i.e. a part of the headroom, not the context.
>> A nice side effect is that the maximum frame size for this mode gets
>> increased by 40 bytes, as xdp_buff::frame_sz includes everything from
>> data_hard_start (-> includes xdpf already) to the end of XDP/skb shared
>> info.
>
> I like the union approach, however...
>
>> (was found while testing XDP traffic generator on ice, which calls
>> xdp_convert_frame_to_buff() for each XDP frame)
>>
>> Fixes: b530e9e1063e ("bpf: Add "live packet" mode for XDP in BPF_PROG_RUN")
>> Signed-off-by: Alexander Lobakin <alexandr.lobakin@xxxxxxxxx>
>> ---
>> net/bpf/test_run.c | 13 ++++++++-----
>> 1 file changed, 8 insertions(+), 5 deletions(-)
>>
>> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
>> index 2723623429ac..c3cce7a8d47d 100644
>> --- a/net/bpf/test_run.c
>> +++ b/net/bpf/test_run.c
>> @@ -97,8 +97,11 @@ static bool bpf_test_timer_continue(struct bpf_test_timer *t, int iterations,
>> struct xdp_page_head {
>> struct xdp_buff orig_ctx;
>> struct xdp_buff ctx;
>> - struct xdp_frame frm;
>> - u8 data[];
>> + union {
>> + /* ::data_hard_start starts here */
>> + DECLARE_FLEX_ARRAY(struct xdp_frame, frm);
>> + DECLARE_FLEX_ARRAY(u8, data);
>> + };
>
> ...why does the xdp_frame need to be a flex array? Shouldn't this just be:
>
> + union {
> + /* ::data_hard_start starts here */
> + struct xdp_frame frm;
> + DECLARE_FLEX_ARRAY(u8, data);
> + };
>
> which would also get rid of the other three hunks of the patch?
That was my first thought. However, as I mentioned in between the lines
in the commitmsg, this doesn't decrease the sizeof(ctx), so we'd have to
replace those sizeofs with offsetof() in a couple places (-> the patch
length would be the same). So I went this way to declare that frm
doesn't belong to ctx but to the headroom.
I'm fine either way tho, so up to you guys.
>
> -Toke
>
Thanks,
Olek
