Re: writev error codes

From: Michal Hocko
Date: Thu Jan 19 2017 - 02:55:56 EST


On Thu 19-01-17 15:09:52, Michael Kerrisk wrote:
> Hello Michal,
>
> On 19 January 2017 at 04:44, Michal Hocko <mhocko@xxxxxxxxxx> wrote:
> > Hi,
> > we have noticed that one of the LTP tests started to fail after
> > 99526912c934 ("fix iov_iter_fault_in_readable()"). The code has expected
> > EINVAL while it gets EFAULT. I believe the new behavior is reasonable,
> > but checking the man 2 writev, there is no mention about EFAULT,
> > and other errnos for that matter, so it seems this is rather under
> > documented and it can confuse users. LTP has been fixed in the meantime
> > [1] but this might come unexpected to others.
> >
> > In principle writev as a write
> > "multiplier" should be allowed all the error codes that write(2) allows,
> > right? I am not sure how we should reflect that. Either c&p what we have
> > in man 2 write or put a reference to it and only describe writev
> > specific, if there are any (I haven't checked that).
> >
> > [1] https://github.com/linux-test-project/ltp/commit/db19194527060a962955a7db54a2f5710e69bec9
>
> Perhaps I am misunderstanding you, but in the writev(2) page there is the text:
>
> ERRORS
> The errors are as given for read(2) and write(2). Furthermore,
> preadv(), preadv2(), pwritev(), and pwritev2() can also fail for
> the same reasons as lseek(2).

This has hit right to the blind spot. Sorry about the noise!

>
> And in the write(2) page, we have:
>
> EFAULT buf is outside your accessible address space.
>
> Does this not cover the case you describe?

Yes

--
Michal Hocko
SUSE Labs