On Wed, Nov 18, 2020 at 9:34 AM <mariusz.dudek@xxxxxxxxx> wrote:
>
> From: Mariusz Dudek <mariuszx.dudek@xxxxxxxxx>
>
> Add support for separation of eBPF program load and xsk socket
> creation.
>
> This is needed for use-case when you want to privide as little
> privileges as possible to the data plane application that will
> handle xsk socket creation and incoming traffic.
>
> With this patch the data entity container can be run with only
> CAP_NET_RAW capability to fulfill its purpose of creating xsk
> socket and handling packages. In case your umem is larger or
> equal process limit for MEMLOCK you need either increase the
> limit or CAP_IPC_LOCK capability.
>
> To resolve privileges issue two APIs are introduced:
>
> - xsk_setup_xdp_prog - loads the built in XDP program. It can
> also return xsks_map_fd which is needed by unprivileged process
> to update xsks_map with AF_XDP socket "fd"
>
> - xsk_socket__update_xskmap - inserts an AF_XDP socket into an xskmap
> for a particular xsk_socket
>
> Signed-off-by: Mariusz Dudek <mariuszx.dudek@xxxxxxxxx>
> ---
> tools/lib/bpf/libbpf.map | 2 +
> tools/lib/bpf/xsk.c | 97 ++++++++++++++++++++++++++++++++++++----
> tools/lib/bpf/xsk.h | 5 +++
> 3 files changed, 95 insertions(+), 9 deletions(-)
>
> diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map
> index 29ff4807b909..d939d5ac092e 100644
> --- a/tools/lib/bpf/libbpf.map
> +++ b/tools/lib/bpf/libbpf.map
> @@ -345,4 +345,6 @@ LIBBPF_0.3.0 {
> btf__parse_split;
> btf__new_empty_split;
> btf__new_split;
> + xsk_setup_xdp_prog;
> + xsk_socket__update_xskmap;
> } LIBBPF_0.2.0;
> diff --git a/tools/lib/bpf/xsk.c b/tools/lib/bpf/xsk.c
> index 9bc537d0b92d..e16f920d2ef9 100644
> --- a/tools/lib/bpf/xsk.c
> +++ b/tools/lib/bpf/xsk.c
> @@ -566,8 +566,42 @@ static int xsk_set_bpf_maps(struct xsk_socket *xsk)
> &xsk->fd, 0);
> }
>
> -static int xsk_setup_xdp_prog(struct xsk_socket *xsk)
> +static int xsk_create_xsk_struct(int ifindex, struct xsk_socket *xsk)
> {
> + char ifname[IFNAMSIZ];
> + struct xsk_ctx *ctx;
> + char *interface;
> + int res = -1;
No need to set it to -1 anymore, due to the below.
> +
> + ctx = calloc(1, sizeof(*ctx));
> + if (!ctx)
> + goto error_ctx;
return an -ENOMEM here directly.
> +
> + interface = if_indextoname(ifindex, &ifname[0]);
> + if (!interface) {
> + res = -errno;
> + goto error_ifindex;
> + }
> +
> + ctx->ifindex = ifindex;
> + strncpy(ctx->ifname, ifname, IFNAMSIZ - 1);
> + ctx->ifname[IFNAMSIZ - 1] = 0;
> +
> + xsk->ctx = ctx;
> +
> + return 0;
> +
> +error_ifindex:
> + free(ctx);
> +error_ctx:
And you can get rid of this label.
> + return res;
> +}
> +
> +static int __xsk_setup_xdp_prog(struct xsk_socket *_xdp,
> + bool force_set_map,
force_set_map always seems to be false now. Correct? If it is, then it
is not needed anymore. What was the original use case of this boolean?
> + int *xsks_map_fd)
> +{
> + struct xsk_socket *xsk = _xdp;
> struct xsk_ctx *ctx = xsk->ctx;
> __u32 prog_id = 0;
> int err;
> @@ -584,8 +618,7 @@ static int xsk_setup_xdp_prog(struct xsk_socket *xsk)
>
> err = xsk_load_xdp_prog(xsk);
> if (err) {
> - xsk_delete_bpf_maps(xsk);
> - return err;
> + goto err_load_xdp_prog;
> }
> } else {
> ctx->prog_fd = bpf_prog_get_fd_by_id(prog_id);
> @@ -598,15 +631,29 @@ static int xsk_setup_xdp_prog(struct xsk_socket *xsk)
> }
> }
>
> - if (xsk->rx)
> + if (xsk->rx || force_set_map) {
> err = xsk_set_bpf_maps(xsk);
> - if (err) {
> - xsk_delete_bpf_maps(xsk);
> - close(ctx->prog_fd);
> - return err;
> + if (err) {
> + if (!prog_id) {
> + goto err_set_bpf_maps;
> + } else {
> + close(ctx->prog_fd);
> + return err;
> + }
> + }
> }
> + if (xsks_map_fd)
> + *xsks_map_fd = ctx->xsks_map_fd;
>
> return 0;
> +
> +err_set_bpf_maps:
> + close(ctx->prog_fd);
> + bpf_set_link_xdp_fd(ctx->ifindex, -1, 0);
> +err_load_xdp_prog:
> + xsk_delete_bpf_maps(xsk);
> +
> + return err;
> }
>
> static struct xsk_ctx *xsk_get_ctx(struct xsk_umem *umem, int ifindex,
> @@ -689,6 +736,38 @@ static struct xsk_ctx *xsk_create_ctx(struct xsk_socket *xsk,
> return ctx;
> }
>
> +static void xsk_destroy_xsk_struct(struct xsk_socket *xsk)
> +{
> + free(xsk->ctx);
> + free(xsk);
> +}
> +
> +int xsk_socket__update_xskmap(struct xsk_socket *xsk, int fd)
> +{
> + xsk->ctx->xsks_map_fd = fd;
> + return xsk_set_bpf_maps(xsk);
> +}
> +
> +int xsk_setup_xdp_prog(int ifindex, int *xsks_map_fd)
> +{
> + struct xsk_socket *xsk;
> + int res = -1;
> +
> + xsk = calloc(1, sizeof(*xsk));
> + if (!xsk)
> + return res;
> +
> + res = xsk_create_xsk_struct(ifindex, xsk);
> + if (res)
> + return -EINVAL;
Here you can now return the error from the function, i.e. return res,
as we returned -ENOMEM in that function. You are however leaking the
xsk struct you just allocated in case of error. Needs to be
deallocated.
> +
> + res = __xsk_setup_xdp_prog(xsk, false, xsks_map_fd);
> +
> + xsk_destroy_xsk_struct(xsk);
> +
> + return res;
> +}
> +
> int xsk_socket__create_shared(struct xsk_socket **xsk_ptr,
> const char *ifname,
> __u32 queue_id, struct xsk_umem *umem,
> @@ -838,7 +917,7 @@ int xsk_socket__create_shared(struct xsk_socket **xsk_ptr,
> ctx->prog_fd = -1;
>
> if (!(xsk->config.libbpf_flags & XSK_LIBBPF_FLAGS__INHIBIT_PROG_LOAD)) {
> - err = xsk_setup_xdp_prog(xsk);
> + err = __xsk_setup_xdp_prog(xsk, false, NULL);
> if (err)
> goto out_mmap_tx;
> }
> diff --git a/tools/lib/bpf/xsk.h b/tools/lib/bpf/xsk.h
> index 1069c46364ff..5b74c17ed3d4 100644
> --- a/tools/lib/bpf/xsk.h
> +++ b/tools/lib/bpf/xsk.h
> @@ -201,6 +201,11 @@ struct xsk_umem_config {
> __u32 flags;
> };
>
> +LIBBPF_API int xsk_setup_xdp_prog(int ifindex,
> + int *xsks_map_fd);
> +LIBBPF_API int xsk_socket__update_xskmap(struct xsk_socket *xsk,
> + int xsks_map_fd);
> +
> /* Flags for the libbpf_flags field. */
> #define XSK_LIBBPF_FLAGS__INHIBIT_PROG_LOAD (1 << 0)
>
> --
> 2.20.1
>
