(no subject)

[Murphy Zhou]

Hi Christoph,

The linux-next tree, since the next-20230522 tag, LTP/writev07[1]
starts to fail on xfs, not on other fs. It was pass on the previous
tag next-20230519.

After those 2 commits reverted on the top of 0522 tree, it passed.

    iomap: update ki_pos in iomap_file_buffered_write
    iomap: assign current->backing_dev_info in iomap_file_buffered_write

(the second one was reverted because the first one depends on it)

The test case writev07 forms an iovec with a bad address in the middle,
then writes to the file, expecting a fault return and file not being written.
Now it fails with a fault return but the file has been written.

Looks like it is related to the pos update, could you help to take a look ?

Thanks,
Murphy


[1] https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/syscalls/writev/writev07.c

Re: your mail

[Christoph Hellwig]

On Thu, May 25, 2023 at 08:01:22PM +0800, Murphy Zhou wrote:
> Hi Christoph,
> 
> The linux-next tree, since the next-20230522 tag, LTP/writev07[1]
> starts to fail on xfs, not on other fs. It was pass on the previous
> tag next-20230519.
> 
> After those 2 commits reverted on the top of 0522 tree, it passed.
> 
>     iomap: update ki_pos in iomap_file_buffered_write
>     iomap: assign current->backing_dev_info in iomap_file_buffered_write
> 
> (the second one was reverted because the first one depends on it)

Yes, they are known broken.  There has been a v2 on the list already,
which still has issues for fuse, so there will be a v3.

Re: your mail

[Murphy Zhou]

On Thu, May 25, 2023 at 8:04 PM Christoph Hellwig <hch@lst.de> wrote:
>
> On Thu, May 25, 2023 at 08:01:22PM +0800, Murphy Zhou wrote:
> > Hi Christoph,
> >
> > The linux-next tree, since the next-20230522 tag, LTP/writev07[1]
> > starts to fail on xfs, not on other fs. It was pass on the previous
> > tag next-20230519.
> >
> > After those 2 commits reverted on the top of 0522 tree, it passed.
> >
> >     iomap: update ki_pos in iomap_file_buffered_write
> >     iomap: assign current->backing_dev_info in iomap_file_buffered_write
> >
> > (the second one was reverted because the first one depends on it)
>
> Yes, they are known broken.  There has been a v2 on the list already,
> which still has issues for fuse, so there will be a v3.

Great! Thank you.
>