Re: [PATCH bpf-next v2 05/20] bpf: implement bpf_seq_read() for bpf iterator

Subject: Re: [PATCH bpf-next v2 05/20] bpf: implement bpf_seq_read() for bpf iterator
From: Alexei Starovoitov <ast@xxxxxx>
Date: Tue, 5 May 2020 12:57:46 -0700
Cc: Andrii Nakryiko <andriin@xxxxxx>, bpf <bpf@xxxxxxxxxxxxxxx>, Martin KaFai Lau <kafai@xxxxxx>, Networking <netdev@xxxxxxxxxxxxxxx>, Daniel Borkmann <daniel@xxxxxxxxxxxxx>, Kernel Team <kernel-team@xxxxxx>
In-reply-to: <CAEf4BzYKACiOB+cAC+g-LdJNJbnz9yrGyw7VsBoW1b2pHjUghw@mail.gmail.com>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:68.0) Gecko/20100101 Thunderbird/68.7.0

On 5/5/20 12:56 PM, Andrii Nakryiko wrote:

+               seq->buf = kmalloc(seq->size, GFP_KERNEL);
+               if (!seq->buf)
+                       goto Enomem;

Why not just mutex_unlock and exit with -ENOMEM? Less goto'ing, more
straightforward.


no. please keep kernel coding style. goto is appropriate here.

References:
- [PATCH bpf-next v2 00/20] bpf: implement bpf iterator for kernel data
  - From: Yonghong Song
- [PATCH bpf-next v2 05/20] bpf: implement bpf_seq_read() for bpf iterator
  - From: Yonghong Song
- Re: [PATCH bpf-next v2 05/20] bpf: implement bpf_seq_read() for bpf iterator
  - From: Andrii Nakryiko