netdev - Re: [PATCH bpf 1/2] bpftool: support queues and stacks in update command

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190103114311.22fc80e6@cakuba.hsd1.ca.comcast.net>
Date:   Thu, 3 Jan 2019 11:43:11 -0800
From:   Jakub Kicinski <jakub.kicinski@...ronome.com>
To:     Stanislav Fomichev <sdf@...gle.com>
Cc:     netdev@...r.kernel.org, davem@...emloft.net, ast@...nel.org,
        daniel@...earbox.net, quentin.monnet@...ronome.com
Subject: Re: [PATCH bpf 1/2] bpftool: support queues and stacks in update
 command

On Thu,  3 Jan 2019 10:33:05 -0800, Stanislav Fomichev wrote:
> Bpftool expects both key and value for 'update' operations. For some
> map types, key should not be specified. Support updating those map types.
> 
> Before:
> bpftool map create /sys/fs/bpf/q type queue value 4 entries 10 name q
> bpftool map update pinned /sys/fs/bpf/q value 0 1 2 3
> Error: did not find key
> 
> After:
> bpftool map create /sys/fs/bpf/q type queue value 4 entries 10 name q
> bpftool map update pinned /sys/fs/bpf/q value 0 1 2 3
> 
> Signed-off-by: Stanislav Fomichev <sdf@...gle.com>

I guess it doesn't hurt to fix update/lookup, but I'd prefer to see new
separate subcommands to be honest :(

bpftool map push/pop/peek

Could you add those as well?  I think most users will be more familiar
with the helpers than the fact that the syscall reuses the old commands.

> diff --git a/tools/bpf/bpftool/map.c b/tools/bpf/bpftool/map.c
> index 2037e3dc864b..30b92715248d 100644
> --- a/tools/bpf/bpftool/map.c
> +++ b/tools/bpf/bpftool/map.c
> @@ -781,11 +781,11 @@ static int do_dump(int argc, char **argv)
>  
>  static int do_update(int argc, char **argv)
>  {
> +	void *key = NULL, *value = NULL;

nit: it seems tiny bit more readable to init these in the place you'd
     otherwise set key to malloc (in an else clause)

>  	struct bpf_map_info info = {};
>  	__u32 len = sizeof(info);
>  	__u32 *value_fd = NULL;
>  	__u32 flags = BPF_ANY;
> -	void *key, *value;
>  	int fd, err;
>  
>  	if (argc < 2)
> @@ -795,9 +795,16 @@ static int do_update(int argc, char **argv)
>  	if (fd < 0)
>  		return -1;
>  
> -	key = malloc(info.key_size);
> +	if (info.key_size) {
> +		key = malloc(info.key_size);
> +		if (!key) {
> +			p_err("mem alloc failed");
> +			err = -1;
> +			goto exit_free;
> +		}
> +	}
>  	value = alloc_value(&info);

Would you mind taking care of the value as well?  So we are ready if
sets are ever added?

> -	if (!key || !value) {
> +	if (!value) {
>  		p_err("mem alloc failed");
>  		err = -1;
>  		goto exit_free;

I'd consider this -next material TBH, but not strongly.