netdev - Re: [PATCH bpf-next v3 08/21] bpf: implement common macros/helpers for target iterators

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: <a847e0ef-2308-ebf1-ca21-f30372fa2678@fb.com>
Date:   Fri, 8 May 2020 20:18:29 -0700
From:   Yonghong Song <yhs@...com>
To:     Andrii Nakryiko <andrii.nakryiko@...il.com>
CC:     Andrii Nakryiko <andriin@...com>, bpf <bpf@...r.kernel.org>,
        Martin KaFai Lau <kafai@...com>,
        Networking <netdev@...r.kernel.org>,
        Alexei Starovoitov <ast@...com>,
        Daniel Borkmann <daniel@...earbox.net>,
        Kernel Team <kernel-team@...com>
Subject: Re: [PATCH bpf-next v3 08/21] bpf: implement common macros/helpers
 for target iterators



On 5/8/20 12:07 PM, Andrii Nakryiko wrote:
> On Wed, May 6, 2020 at 10:40 PM Yonghong Song <yhs@...com> wrote:
>>
>> Macro DEFINE_BPF_ITER_FUNC is implemented so target
>> can define an init function to capture the BTF type
>> which represents the target.
>>
>> The bpf_iter_meta is a structure holding meta data, common
>> to all targets in the bpf program.
>>
>> Additional marker functions are called before or after
>> bpf_seq_read() show()/next()/stop() callback functions
>> to help calculate precise seq_num and whether call bpf_prog
>> inside stop().
>>
>> Two functions, bpf_iter_get_info() and bpf_iter_run_prog(),
>> are implemented so target can get needed information from
>> bpf_iter infrastructure and can run the program.
>>
>> Signed-off-by: Yonghong Song <yhs@...com>
>> ---
>>   include/linux/bpf.h   | 11 ++++++
>>   kernel/bpf/bpf_iter.c | 86 ++++++++++++++++++++++++++++++++++++++++---
>>   2 files changed, 92 insertions(+), 5 deletions(-)
>>
> 
> Looks good. I was worried about re-using seq_num when element is
> skipped, but this could already happen that same seq_num is associated
> with different objects: overflow + retry returns different object
> (because iteration is not a snapshot, so the element could be gone on
> retry). Both cases will have to be handled in about the same fashion,
> so it's fine.

This is what I thought as well.

> 
> Hm... Could this be a problem for start() implementation? E.g., if
> object is still there, but iterator wants to skip it permanently.
> Re-using seq_num will mean that start() will keep trying to fetch same
> to-be-skipped element? Not sure, please think about it, but we can fix
> it up later, if necessary.

The seq_num is for bpf_program context. It does not affect how start()
behaves. The start() MAY retry the same element over and over again
if show() overflows or returns <0, but in which case, user space
should check the return error code to decide to retry or give up.

> 
> Acked-by: Andrii Nakryiko <andriin@...com>
> 
> [...]
> 
>> @@ -112,11 +143,16 @@ static ssize_t bpf_seq_read(struct file *file, char __user *buf, size_t size,
>>                          err = PTR_ERR(p);
>>                          break;
>>                  }
>> +
>> +               /* get a valid next object, increase seq_num */
> 
> typo: get -> got

Ack.

> 
>> +               bpf_iter_inc_seq_num(seq);
>> +
>>                  if (seq->count >= size)
>>                          break;
>>
>>                  err = seq->op->show(seq, p);
>>                  if (err > 0) {
>> +                       bpf_iter_dec_seq_num(seq);
>>                          seq->count = offs;
>>                  } else if (err < 0 || seq_has_overflowed(seq)) {
>>                          seq->count = offs;
> 
> [...]
>