Re: [PATCH] incorrect error handling inside generic_file_direct_write

["'Christoph Hellwig'" <hch@xxxxxxxxxxxxx>]

> +ssize_t
> +__generic_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
> +                             unsigned long nr_segs, loff_t pos)

I'd still call this generic_file_aio_write_nolock.

> +     loff_t          *ppos = &iocb->ki_pos;

I'd rather use iocb->ki_pos directly in the few places ppos is referenced
currently.

>       if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
> -             ssize_t err;
> -
>               err = sync_page_range_nolock(inode, mapping, pos, ret);
>               if (err < 0)
>                       ret = err;
>       }

So we're doing the sync_page_range once in __generic_file_aio_write
with i_mutex held.


>       mutex_lock(&inode->i_mutex);
> -     ret = __generic_file_aio_write_nolock(iocb, iov, nr_segs,
> -                     &iocb->ki_pos);
> +     ret = __generic_file_aio_write(iocb, iov, nr_segs, pos);
>       mutex_unlock(&inode->i_mutex);
>  
>       if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {

And then another time after it's unlocked, this seems wrong.