From: Yonghong Song <yhs@fb.com>
To: Carlos Antonio Neira Bustos <cneirabustos@gmail.com>
Cc: "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
"ebiederm@xmission.com" <ebiederm@xmission.com>,
"brouer@redhat.com" <brouer@redhat.com>,
"bpf@vger.kernel.org" <bpf@vger.kernel.org>
Subject: Re: [PATCH bpf-next V9 1/3] bpf: new helper to obtain namespace data from current task
Date: Tue, 20 Aug 2019 17:29:54 +0000 [thread overview]
Message-ID: <e1265647-f934-1cb5-a31d-73608b7adad5@fb.com> (raw)
In-Reply-To: <20190820151040.GA53610@localhost>
On 8/20/19 8:10 AM, Carlos Antonio Neira Bustos wrote:
> Hi Yonghong,
>
> Thanks for taking the time to review this.
>
>
>>> + *
>>> + * **-EINVAL** if *size_of_pidns* is not valid or unable to get ns, pid
>>> + * or tgid of the current task.
>>> + *
>>> + * **-ECHILD** if /proc/self/ns/pid does not exists.
>>> + *
>>> + * **-ENOTDIR** if /proc/self/ns does not exists.
>>
>> Let us remove ECHILD and ENOTDIR and replace it with ENOENT as I
>> described below.
>>
>> Please *do verify* what happens when namespaces or pid_ns are not
>> configured.
>>
>
>
> I have tested kernel configurations without namespace support and with
> namespace support but without pid namespaces, the helper returns -EINVAL
> on both cases, now it should return -ENOENT.
Indeed. -ENOENT is better.
>
>
>>> +struct bpf_pidns_info {
>>> + __u32 dev;
>>
>> Please add a comment for dev for how device major and minor number are
>> derived. User space gets device major and minor number, they need to
>> compare to the corresponding major/minor numbers returned by this helper.
>>
>>> + __u32 nsid;
>>> + __u32 tgid;
>>> + __u32 pid;
>>> +};
>>
>
> What do you think of this comment ?
>
> struct bpf_pidns_info {
> __u32 dev; /* major/minor numbers from /proc/self/ns/pid.
> * User space gets device major and minor numbers from
> * the same device that need to be compared against the
> * major/minor numbers returned by this helper.
> */
> __u32 nsid;
> __u32 tgid;
> __u32 pid;
> };
>
To be more specific, I like a comment similar to below in uapi bpf.h
struct bpf_cgroup_dev_ctx {
/* access_type encoded as (BPF_DEVCG_ACC_* << 16) |
BPF_DEVCG_DEV_* */
__u32 access_type;
__u32 major;
__u32 minor;
};
Some like:
/* dev encoded as (major << 8 | (minor & 0xff)) */
>>
>> Please put an empty line. As a general rule for readability,
>> put an empty line if control flow is interrupted, e.g., by
>> "return", "break" or "continue". At least this is what
>> I saw most in bpf mailing list.
>>
> I'll fix it in version 10.
>
>>> + len = strlen(pidns_path) + 1;
>>> + memcpy((char *)tmp->name, pidns_path, len);
>>> + tmp->uptr = NULL;
>>> + tmp->aname = NULL;
>>> + tmp->refcnt = 1;
>>> + ret = filename_lookup(AT_FDCWD, tmp, 0, &kp, NULL);
>> Adding below to free kmem cache memory
>> kmem_cache_free(names_cachep, fname);
>>
>
> I think we don't need to call kmem_cache_free as filename_lookup
> calls putname that calls kmem_cache_free.
Oh, right. Thanks for checking this.
>
>
> Thanks a lot for your help.
>
> Bests
>
>> In the above, we checked task_active_pid_ns().
>> If not returning NULL, we have a valid pid ns. So the above
>> filename_lookup should not go wrong. We can still keep
>> the error checking though.
>>
>>> + if (ret) {
>>> + memset((void *)pidns_info, 0, (size_t) size);
>>> + return ret;
>>
>>
>
> I think we could get rid of this.
>
>
next prev parent reply other threads:[~2019-08-20 17:30 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-13 18:47 [PATCH bpf-next V9 0/3] BPF: New helper to obtain namespace data from current task Carlos Neira
2019-08-13 18:47 ` [PATCH bpf-next V9 1/3] bpf: new " Carlos Neira
2019-08-13 22:35 ` Yonghong Song
2019-08-20 15:10 ` Carlos Antonio Neira Bustos
2019-08-20 17:29 ` Yonghong Song [this message]
2019-08-13 23:11 ` Yonghong Song
2019-08-13 23:51 ` [Potential Spoof] " Yonghong Song
2019-08-14 0:56 ` Carlos Antonio Neira Bustos
2019-08-14 4:50 ` Yonghong Song
[not found] ` <CACiB22jyN9=0ATWWE+x=BoWD6u+8KO+MvBfsFQmcNfkmANb2_w@mail.gmail.com>
2019-08-28 20:39 ` Carlos Antonio Neira Bustos
2019-08-28 20:53 ` Yonghong Song
2019-08-28 21:03 ` Carlos Antonio Neira Bustos
2019-09-03 18:45 ` Carlos Antonio Neira Bustos
2019-09-03 20:36 ` Yonghong Song
2019-08-13 18:47 ` [PATCH bpf-next V9 2/3] samples/bpf: added sample code for bpf_get_current_pidns_info Carlos Neira
2019-08-13 18:47 ` [PATCH bpf-next V9 3/3] tools/testing/selftests/bpf: Add self-tests for new helper Carlos Neira
2019-08-13 23:19 ` Yonghong Song
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e1265647-f934-1cb5-a31d-73608b7adad5@fb.com \
--to=yhs@fb.com \
--cc=bpf@vger.kernel.org \
--cc=brouer@redhat.com \
--cc=cneirabustos@gmail.com \
--cc=ebiederm@xmission.com \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).