Re: [PATCH] MIPS: BPF: Add support for SKF_AD_HATYPE

[David Daney]

On 03/13/2017 03:56 AM, James Hogan wrote:
> On Fri, Mar 10, 2017 at 02:14:05PM -0800, David Daney wrote:
> > This let's us pass some additional "modprobe test-bpf" tests.
> >
> > Reuse the code for SKF_AD_IFINDEX, but substitute the offset and size
> > of the "type" field.
> >
> > Signed-off-by: David Daney <david.daney@xxxxxxxxxx>
>
> I think the BPF maintainers should probably be Cc'd on this patch.
> Cc'ing now.

Good point.

Since there are some corrections needed, I will send another version and 
CC the proper people.


> > ---
> >  arch/mips/net/bpf_jit.c | 14 ++++++++++----
> >  1 file changed, 10 insertions(+), 4 deletions(-)
> >
> > diff --git a/arch/mips/net/bpf_jit.c b/arch/mips/net/bpf_jit.c
> > index 49a2e22..f613708 100644
> > --- a/arch/mips/net/bpf_jit.c
> > +++ b/arch/mips/net/bpf_jit.c
> > @@ -1111,6 +1111,7 @@ static int build_body(struct jit_ctx *ctx)
> >  			emit_load(r_A, 28, off, ctx);
> >  			break;
> >  		case BPF_ANC | SKF_AD_IFINDEX:
> > +		case BPF_ANC | SKF_AD_HATYPE:
> >  			/* A = skb->dev->ifindex */
>
> this comment should probably be updated.

Right.

> >  			ctx->flags |= SEEN_SKB | SEEN_A;
> >  			off = offsetof(struct sk_buff, dev);
> > @@ -1120,10 +1121,15 @@ static int build_body(struct jit_ctx *ctx)
> >  			emit_bcond(MIPS_COND_EQ, r_s0, r_zero,
> >  				   b_imm(prog->len, ctx), ctx);
> >  			emit_reg_move(r_ret, r_zero, ctx);
> > -			BUILD_BUG_ON(FIELD_SIZEOF(struct net_device,
> > -						  ifindex) != 4);
> > -			off = offsetof(struct net_device, ifindex);
> > -			emit_load(r_A, r_s0, off, ctx);
> > +			if (code == (BPF_ANC | SKF_AD_IFINDEX)) {
> > +				BUILD_BUG_ON(FIELD_SIZEOF(struct net_device, ifindex) != 4);
> > +				off = offsetof(struct net_device, ifindex);
> > +				emit_load(r_A, r_s0, off, ctx);
> > +			} else { /* (code == (BPF_ANC | SKF_AD_HATYPE) */
> > +				BUILD_BUG_ON(FIELD_SIZEOF(struct net_device, type) != 2);
> > +				off = offsetof(struct net_device, type);
> > +				emit_half_load(r_A, r_s0, off, ctx);
>
> Technically net_device::type is unsigned, and emit_half_load uses LH
> which sign extends. Does that matter in practice.

The next version will use LHU.


> Cheers
> James