diff mbox

[iproute2,-next,1/5] {f,m}_bpf: make tail calls working

Message ID 88d074dcdf1469ea381e1d732852cc87fb1aba2c.1448541097.git.daniel@iogearbox.net
State Superseded, archived
Delegated to: stephen hemminger
Headers show

Commit Message

Daniel Borkmann Nov. 26, 2015, 12:58 p.m. UTC 
Now that we have the possibility of sharing maps, it's time we get the
ELF loader fully working with regards to tail calls. Since program array
maps are pinned, we can keep them finally alive. I've noticed two bugs
that are being fixed in bpf_fill_prog_arrays() with this patch. Example
code comes as follow-up.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
---
 tc/tc_bpf.c | 31 +++++++++++++++++++++++--------
 1 file changed, 23 insertions(+), 8 deletions(-)

Comments

Sergei Shtylyov Nov. 26, 2015, 1:10 p.m. UTC | #1 
Hello.

On 11/26/2015 3:58 PM, Daniel Borkmann wrote:

> Now that we have the possibility of sharing maps, it's time we get the
> ELF loader fully working with regards to tail calls. Since program array
> maps are pinned, we can keep them finally alive. I've noticed two bugs
> that are being fixed in bpf_fill_prog_arrays() with this patch. Example
> code comes as follow-up.
>
> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
> Acked-by: Alexei Starovoitov <ast@kernel.org>
> ---
>   tc/tc_bpf.c | 31 +++++++++++++++++++++++--------
>   1 file changed, 23 insertions(+), 8 deletions(-)
>
> diff --git a/tc/tc_bpf.c b/tc/tc_bpf.c
> index bc7bc9f..73a0f41 100644
> --- a/tc/tc_bpf.c
> +++ b/tc/tc_bpf.c
> @@ -1139,11 +1139,26 @@ static int bpf_fetch_prog_sec(struct bpf_elf_ctx *ctx, const char *section)
>   	return ret;
>   }
>
> +static int bpf_find_map_by_id(struct bpf_elf_ctx *ctx, uint32_t id)
> +{
> +	int i, ret = -1;
> +
> +	for (i = 0; i < ARRAY_SIZE(ctx->map_fds); i++) {
> +		if (ctx->map_fds[i] && ctx->maps[i].id == id &&
> +		    ctx->maps[i].type == BPF_MAP_TYPE_PROG_ARRAY) {
> +			ret = i;
> +			break;

    Isn't just *return* i; simpler? That way, you don't need 'ret' at all.

> +		}
> +	}
> +
> +	return ret;
> +}
> +
[...]

MBR, Sergei

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Daniel Borkmann Nov. 26, 2015, 2:40 p.m. UTC | #2 
On 11/26/2015 02:10 PM, Sergei Shtylyov wrote:
[...]
>     Isn't just *return* i; simpler? That way, you don't need 'ret' at all.

Okay, sure, fixed.

Thanks,
Daniel
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/tc/tc_bpf.c b/tc/tc_bpf.c
index bc7bc9f..73a0f41 100644
--- a/tc/tc_bpf.c
+++ b/tc/tc_bpf.c
@@ -1139,11 +1139,26 @@  static int bpf_fetch_prog_sec(struct bpf_elf_ctx *ctx, const char *section)
 	return ret;
 }
 
+static int bpf_find_map_by_id(struct bpf_elf_ctx *ctx, uint32_t id)
+{
+	int i, ret = -1;
+
+	for (i = 0; i < ARRAY_SIZE(ctx->map_fds); i++) {
+		if (ctx->map_fds[i] && ctx->maps[i].id == id &&
+		    ctx->maps[i].type == BPF_MAP_TYPE_PROG_ARRAY) {
+			ret = i;
+			break;
+		}
+	}
+
+	return ret;
+}
+
 static int bpf_fill_prog_arrays(struct bpf_elf_ctx *ctx)
 {
 	struct bpf_elf_sec_data data;
 	uint32_t map_id, key_id;
-	int fd, i, ret;
+	int fd, i, ret, idx;
 
 	for (i = 1; i < ctx->elf_hdr.e_shnum; i++) {
 		if (ctx->sec_done[i])
@@ -1153,20 +1168,20 @@  static int bpf_fill_prog_arrays(struct bpf_elf_ctx *ctx)
 		if (ret < 0)
 			continue;
 
-		ret = sscanf(data.sec_name, "%u/%u", &map_id, &key_id);
-		if (ret != 2 || map_id >= ARRAY_SIZE(ctx->map_fds) ||
-		    !ctx->map_fds[map_id])
+		ret = sscanf(data.sec_name, "%i/%i", &map_id, &key_id);
+		if (ret != 2)
 			continue;
-		if (ctx->maps[map_id].type != BPF_MAP_TYPE_PROG_ARRAY ||
-		    ctx->maps[map_id].max_elem <= key_id)
+
+		idx = bpf_find_map_by_id(ctx, map_id);
+		if (idx < 0)
 			continue;
 
 		fd = bpf_fetch_prog_sec(ctx, data.sec_name);
 		if (fd < 0)
 			return -EIO;
 
-		ret = bpf_map_update(ctx->map_fds[map_id], &key_id,
-				     &fd, BPF_NOEXIST);
+		ret = bpf_map_update(ctx->map_fds[idx], &key_id,
+				     &fd, BPF_ANY);
 		if (ret < 0)
 			return -ENOENT;