On 07/24, Maciej Fijalkowski wrote: > On Fri, Jul 12, 2024 at 06:52:53PM -0700, Stanislav Fomichev wrote: > > Add a couple of things: > > 1. Remove xdp_umem_reg_v2 since its sizeof is the same as xdp_umem_reg > > So thing here is that adding __attribute__((packed)) on kernel side > wouldn't help because we wouldn't fix old uapi with this, correct? old > uapi would still yield 32 bytes for xdp_umem_reg without tx_metadata_len. > > Just explaining here to myself. Yea :-( > > 2. Add BUILD_BUG_ON that checks that the size of xdp_umem_reg_v1 is less > > than xdp_umem_reg; presumably, when we get to v2, there is gonna > > be a similar line to enforce that sizeof(v2) > sizeof(v1) > > 3. Add BUILD_BUG_ON to make sure the last field plus its size matches > > the overall struct size. The intent is to demonstrate that we don't > > have any lingering padding. > > This is good stuff but I wonder wouldn't it be more feasible to squash > this with 1/3 ? And have it backported. Regarding the patch logistics, you > did not provide fixes tag here for some reason, but still include the > patch routed via bpf tree. SG, will resend this against bpf-next.
