The output format of unsigned int should be %u, and the output
format of int should be %d, so fix it.

Signed-off-by: liujing <liujing@cmss.chinamobile.com>

diff --git a/tools/bpf/bpftool/netlink_dumper.c b/tools/bpf/bpftool/netlink_dumper.c
index 5f65140b003b..97e1e1dbc842 100644
--- a/tools/bpf/bpftool/netlink_dumper.c
+++ b/tools/bpf/bpftool/netlink_dumper.c
@@ -45,7 +45,7 @@ static int do_xdp_dump_one(struct nlattr *attr, unsigned int ifindex,
 	NET_START_OBJECT;
 	if (name)
 		NET_DUMP_STR("devname", "%s", name);
-	NET_DUMP_UINT("ifindex", "(%d)", ifindex);
+	NET_DUMP_UINT("ifindex", "(%u)", ifindex);
 
 	if (mode == XDP_ATTACHED_MULTI) {
 		if (json_output) {
@@ -168,7 +168,7 @@ int do_filter_dump(struct tcmsg *info, struct nlattr **tb, const char *kind,
 		NET_START_OBJECT;
 		if (devname[0] != '\0')
 			NET_DUMP_STR("devname", "%s", devname);
-		NET_DUMP_UINT("ifindex", "(%u)", ifindex);
+		NET_DUMP_UINT("ifindex", "(%d)", ifindex);
 		NET_DUMP_STR("kind", " %s", kind);
 		ret = do_bpf_filter_dump(tb[TCA_OPTIONS]);
 		NET_END_OBJECT_FINAL;
-- 
2.27.0

Please drop the "tools/bpf: " part of the prefix in your patch title,
it's enough to keep "bpftool: ".


2024-11-28 10:55 UTC+0800 ~ liujing <liujing@cmss.chinamobile.com>
> The output format of unsigned int should be %u, and the output
> format of int should be %d, so fix it.
> 
> Signed-off-by: liujing <liujing@cmss.chinamobile.com>
> 
> diff --git a/tools/bpf/bpftool/netlink_dumper.c b/tools/bpf/bpftool/netlink_dumper.c
> index 5f65140b003b..97e1e1dbc842 100644
> --- a/tools/bpf/bpftool/netlink_dumper.c
> +++ b/tools/bpf/bpftool/netlink_dumper.c
> @@ -45,7 +45,7 @@ static int do_xdp_dump_one(struct nlattr *attr, unsigned int ifindex,
>  	NET_START_OBJECT;
>  	if (name)
>  		NET_DUMP_STR("devname", "%s", name);
> -	NET_DUMP_UINT("ifindex", "(%d)", ifindex);
> +	NET_DUMP_UINT("ifindex", "(%u)", ifindex);
>  
>  	if (mode == XDP_ATTACHED_MULTI) {
>  		if (json_output) {
> @@ -168,7 +168,7 @@ int do_filter_dump(struct tcmsg *info, struct nlattr **tb, const char *kind,
>  		NET_START_OBJECT;
>  		if (devname[0] != '\0')
>  			NET_DUMP_STR("devname", "%s", devname);
> -		NET_DUMP_UINT("ifindex", "(%u)", ifindex);
> +		NET_DUMP_UINT("ifindex", "(%d)", ifindex);
>  		NET_DUMP_STR("kind", " %s", kind);
>  		ret = do_bpf_filter_dump(tb[TCA_OPTIONS]);
>  		NET_END_OBJECT_FINAL;


Thanks for this. The second chunk is not enough to fix the format
specifier cleanly, because NET_DUMP_UINT() may end up calling:

	jsonw_printf(self, "%"PRIu64, num);

So you probably need to add a NET_DUMP_INT() wrapper and call it here.

There's also another occurrence of the macro called on a signed
"ifindex" in net.c, in __show_dev_tc_bpf(), using "(%u)". Let's fix it
in the same patch, please?

Thanks,
Quentin

On Wed, Nov 27, 2024 at 6:56 PM liujing <liujing@cmss.chinamobile.com> wrote:
>
> The output format of unsigned int should be %u, and the output
> format of int should be %d, so fix it.
>
> Signed-off-by: liujing <liujing@cmss.chinamobile.com>

same issue. full name pls.

pw-bot: cr