Re: [PATCH] vfs: remove might_sleep() from clear_inode()

From: Al Viro
Date: Tue Nov 07 2017 - 21:25:01 EST


On Tue, Nov 07, 2017 at 04:43:54PM -0800, Shakeel Butt wrote:
> Commit 7994e6f72543 ("vfs: Move waiting for inode writeback from
> end_writeback() to evict_inode()") removed inode_sync_wait() from
> end_writeback() and commit dbd5768f87ff ("vfs: Rename end_writeback()
> to clear_inode()") renamed end_writeback() to clear_inode(). After
> these patches there is no sleeping operation in clear_inode(). So,
> remove might_sleep() from it.

Point, but... this is far from the worst annoyance in clear_inode().
Starting with "BUG_ON() under spin_lock_irq() is antisocial and
not in a good way", of course, but that's not all - the whole
cycling of ->tree_lock has already been done back in
truncate_inode_pages_final() and we'd better have called that
in all cases when ->i_data might have ever contained anything.

The whole thing looks bogus these days... I wonder if we should
simply move the remaining paranoia into destroy_inode() and get
rid of the I_CLEAR completely...