Re: [RFC v6 1/2] epoll: Implement eventpoll_replace_file()

[Alok Tiagi]

On Tue, May 23, 2023 at 02:32:06PM +0200, Christian Brauner wrote:
> On Tue, May 23, 2023 at 06:58:01AM +0000, aloktiagi wrote:
> > Introduce a mechanism to replace a file linked in the epoll interface with a new
> > file.
> > 
> > eventpoll_replace() finds all instances of the file to be replaced and replaces
> > them with the new file and the interested events.
> > 
> > Signed-off-by: aloktiagi <aloktiagi@xxxxxxxxx>
> > ---
> > Changes in v6:
> >   - incorporate latest changes that get rid of the global epmutex lock.
> > 
> > Changes in v5:
> >   - address review comments and move the call to replace old file in each
> >     subsystem (epoll, io_uring, etc.) outside the fdtable helpers like
> >     replace_fd().
> > 
> > Changes in v4:
> >   - address review comment to remove the redundant eventpoll_replace() function.
> >   - removed an extra empty line introduced in include/linux/file.h
> > 
> > Changes in v3:
> >   - address review comment and iterate over the file table while holding the
> >     spin_lock(&files->file_lock).
> >   - address review comment and call filp_close() outside the
> >     spin_lock(&files->file_lock).
> > ---
> >  fs/eventpoll.c            | 76 +++++++++++++++++++++++++++++++++++++++
> >  include/linux/eventpoll.h |  8 +++++
> >  2 files changed, 84 insertions(+)
> > 
> > diff --git a/fs/eventpoll.c b/fs/eventpoll.c
> > index 980483455cc0..9c7bffa8401b 100644
> > --- a/fs/eventpoll.c
> > +++ b/fs/eventpoll.c
> > @@ -973,6 +973,82 @@ void eventpoll_release_file(struct file *file)
> >  	spin_unlock(&file->f_lock);
> >  }
> >  
> > +static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
> > +			struct file *tfile, int fd, int full_check);
> > +
> > +/*
> > + * This is called from eventpoll_replace() to replace a linked file in the epoll
> > + * interface with a new file received from another process. This is useful in
> > + * cases where a process is trying to install a new file for an existing one
> > + * that is linked in the epoll interface
> > + */
> > +int eventpoll_replace_file(struct file *toreplace, struct file *file, int tfd)
> > +{
> > +	struct file *to_remove = toreplace;
> > +	struct epoll_event event;
> > +	struct hlist_node *next;
> > +	struct eventpoll *ep;
> > +	struct epitem *epi;
> > +	int error = 0;
> > +	bool dispose;
> > +	int fd;
> > +
> > +	if (!file_can_poll(file))
> > +		return 0;
> > +
> > +	spin_lock(&toreplace->f_lock);
> > +	if (unlikely(!toreplace->f_ep)) {
> > +		spin_unlock(&toreplace->f_lock);
> > +		return 0;
> > +	}
> > +	hlist_for_each_entry_safe(epi, next, toreplace->f_ep, fllink) {
> > +		ep = epi->ep;
> > +		mutex_lock(&ep->mtx);
> 
> Afaict, you're under a spinlock and you're acquiring a mutex. The
> spinlock can't sleep (on non-rt kernels at least) but the mutex can.
> 

thank you. I'll address this in another way in the next patch series. Please
let me know of your opinion on how it can be achieved differently.

> > +		fd = epi->ffd.fd;
> > +		if (fd != tfd) {
> > +			mutex_unlock(&ep->mtx);
> > +			continue;
> > +		}
> > +		event = epi->event;
> > +		error = ep_insert(ep, &event, file, fd, 1);
> > +		mutex_unlock(&ep->mtx);
> > +		if (error != 0) {
> > +			break;
> > +		}
> 
> nit: we don't do { } around single lines.

will fix this in the next series.