Thank you, Jan.
> Can we maybe rephrase the whole comment like:
>
> /*
> * Before flushing the journal and switching inode's aops, we have
> * to flush all dirty data the inode has. There can be outstanding
> * delayed allocations, there can be unwritten extents created by
> * fallocate or buffered writes in dioread_nolock mode covered by
> * dirty data which can be converted only after flushing the dirty
> * data (and journalled aops don't know how to handle these cases).
> */
Your comment is more understandable, so I will change the comment in
the patch.
> BTW, the code is still racy wrt mmap write faults. Once Ted merges my
> patches for hole punching races, we need to grab i_mmap_sem for writing
> here as well to avoid write page fault from creating dirty page while we
> are switching aops... I'm mostly writing this here so that there's lower
> chance this gets lost.
Ok. I've recognized another racy condition when switching aops of an inode.
I will also consider the point in v2 of this patch.
��.n��������+%������w��{.n�����{�����ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f
