From: Pavel Begunkov <[email protected]>
To: Stefan Metzmacher <[email protected]>, Jens Axboe <[email protected]>
Cc: io-uring <[email protected]>,
Linux API Mailing List <[email protected]>
Subject: Re: IORING_REGISTER_CREDS[_UPDATE]() and credfd_create()?
Date: Wed, 29 Jan 2020 17:23:25 +0300 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
On 1/29/2020 4:56 PM, Stefan Metzmacher wrote:
>>> However I think there're a few things to improve/simplify.
>> Since 5.6 is already semi-open, it'd be great to have an incremental
>> patch for that. I'll retoss things as usual, if nobody do it before.
>
> I'll wait for comments from Jens first:-)
> I guess we'll have things changed in his branch, when I wake up
> tomorrow. Otherwise I can also create patches and submit them.
Sure, I won't get there any time soon.
>
> But I currently don't have an environment where I can do runtime tests
> with it.
>
>>>> https://git.kernel.dk/cgit/linux-block/commit/?h=for-5.6/io_uring-vfs&id=a26d26412e1e1783473f9dc8f030c3af3d54b1a6
>>>
>>> In fs/io_uring.c mmgrab() and get_current_cred() are used together in
>>> two places, why is put_cred() called in __io_req_aux_free while
>>> mmdrop() is called from io_put_work(). I think both should be called
>>> in io_put_work(), that makes the code much easier to understand.
>>>
>>> My guess is that you choose __io_req_aux_free() for put_cred() because
>>> of the following patches, but I'll explain on the other commit
>>> why it's not needed.
>>>
>>>> https://git.kernel.dk/cgit/linux-block/commit/?h=for-5.6/io_uring-vfs&id=d9db233adf034bd7855ba06190525e10a05868be
>>>
>>> A minor one would be starting with 1 instead of 0 and using
>>> idr_alloc_cyclic() in order to avoid immediate reuse of ids.
>>> That way we could include the id in the tracing message and
>>> 0 would mean the current creds were used.
>>>
>>>> +static int io_remove_personalities(int id, void *p, void *data)
>>>> +{
>>>> + struct io_ring_ctx *ctx = data;
>>>> +
>>>> + idr_remove(&ctx->personality_idr, id);
>>>
>>> Here we need something like:
>>> put_creds((const struct cred *)p);
>>
>> Good catch
>>
>>>
>>>> + return 0;
>>>> +}
>>>
>>>
>>> The io_uring_register() calles would look like this, correct?
>>>
>>> id = io_uring_register(ring_fd, IORING_REGISTER_PERSONALITY, NULL, 0);
>>> io_uring_register(ring_fd, IORING_UNREGISTER_PERSONALITY, NULL, id);
>>>
>>>> https://git.kernel.dk/cgit/linux-block/commit/?h=for-5.6/io_uring-vfs&id=eec9e69e0ad9ad364e1b6a5dfc52ad576afee235
>>>> +
>>>> + if (sqe_flags & IOSQE_PERSONALITY) {
>>>> + int id = READ_ONCE(sqe->personality);
>>>> +
>>>> + req->work.creds = idr_find(&ctx->personality_idr, id);
>>>> + if (unlikely(!req->work.creds)) {
>>>> + ret = -EINVAL;
>>>> + goto err_req;
>>>> + }
>>>> + get_cred(req->work.creds);> + old_creds = override_creds(req->work.creds);
>>>> + }
>>>> +
>>>
>>> Here we could use a helper variable
>>> const struct cred *personality_creds;
>>> and leave req->work.creds as NULL.
>>> It means we can avoid the explicit get_cred() call
>>> and can skip the following hunk too:
>>>
>>>> @@ -3977,7 +3977,8 @@ static int io_req_defer_prep(struct io_kiocb *req,
>>>> mmgrab(current->mm);
>>>> req->work.mm = current->mm;
>>>> }
>>>> - req->work.creds = get_current_cred();
>>>> + if (!req->work.creds)
>>>> + req->work.creds = get_current_cred();
>>>>
>>>> switch (req->opcode) {
>>>> case IORING_OP_NOP:
>>>
>>> The override_creds(personality_creds) has changed current->cred
>>> and get_current_cred() will just pick it up as in the default case.
>>>
>>> This would make the patch much simpler and allows put_cred() to be
>>> in io_put_work() instead of __io_req_aux_free() as explained above.
>>>
>>
>> It's one extra get_current_cred(). I'd prefer to find another way to
>> clean this up.
>
> As far as I can see it avoids a get_cred() in the IOSQE_PERSONALITY case
> and the if (!req->work.creds) for both cases.
Great, that you turned attention to that! override_creds() is already
grabbing a ref, so it shouldn't call get_cred() there.
So, that's a bug.
It could be I'm wrong with the statement above, need to recheck all this
code to be sure.
BTW, io_req_defer_prep() may be called twice for a req, so you will
reassign it without putting a ref. It's safer to leave NULL checks. At
least, until I've done reworking and fixing preparation paths.
>
> What do you mean exactly with one extra get_current_cred()?
> Is that any worse than calling get_cred() and having an if check?
>
> It also seems to avoid req->work.creds from being filled at all
> for the non-blocking case.
>
> metze
>
--
Pavel Begunkov
next prev parent reply other threads:[~2020-01-29 14:23 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-28 10:18 IORING_REGISTER_CREDS[_UPDATE]() and credfd_create()? Stefan Metzmacher
2020-01-28 16:10 ` Jens Axboe
2020-01-28 16:17 ` Stefan Metzmacher
2020-01-28 16:19 ` Jens Axboe
2020-01-28 17:19 ` Jens Axboe
2020-01-28 18:04 ` Jens Axboe
2020-01-28 19:42 ` Jens Axboe
2020-01-28 20:16 ` Pavel Begunkov
2020-01-28 20:19 ` Jens Axboe
2020-01-28 20:50 ` Pavel Begunkov
2020-01-28 20:56 ` Jens Axboe
2020-01-28 21:25 ` Christian Brauner
2020-01-28 22:38 ` Pavel Begunkov
2020-01-28 23:36 ` Pavel Begunkov
2020-01-28 23:40 ` Jens Axboe
2020-01-28 23:51 ` Jens Axboe
2020-01-29 0:10 ` Pavel Begunkov
2020-01-29 0:15 ` Jens Axboe
2020-01-29 0:18 ` Jens Axboe
2020-01-29 0:20 ` Jens Axboe
2020-01-29 0:21 ` Pavel Begunkov
2020-01-29 0:24 ` Jens Axboe
2020-01-29 0:54 ` Jens Axboe
2020-01-29 10:17 ` Pavel Begunkov
2020-01-29 13:11 ` Stefan Metzmacher
2020-01-29 13:41 ` Pavel Begunkov
2020-01-29 13:56 ` Stefan Metzmacher
2020-01-29 14:23 ` Pavel Begunkov [this message]
2020-01-29 14:27 ` Stefan Metzmacher
2020-01-29 14:34 ` Pavel Begunkov
2020-01-29 17:34 ` Jens Axboe
2020-01-29 17:42 ` Jens Axboe
2020-01-29 20:09 ` Stefan Metzmacher
2020-01-29 20:48 ` Jens Axboe
2020-01-29 17:46 ` Pavel Begunkov
2020-01-29 14:59 ` Jann Horn
2020-01-29 17:34 ` Jens Axboe
2020-01-30 1:08 ` Jens Axboe
2020-01-30 2:20 ` Jens Axboe
2020-01-30 3:18 ` Jens Axboe
2020-01-30 6:53 ` Stefan Metzmacher
2020-01-30 10:11 ` Jann Horn
2020-01-30 10:26 ` Christian Brauner
2020-01-30 14:11 ` Jens Axboe
2020-01-30 14:47 ` Stefan Metzmacher
2020-01-30 15:34 ` Jens Axboe
2020-01-30 15:13 ` Christian Brauner
2020-01-30 15:29 ` Jens Axboe
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 \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/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