On Fri, Mar 24, 2023 at 05:24:26AM +0000, Matthew Wilcox wrote:
> On Fri, Mar 24, 2023 at 05:15:26AM +0000, aloktiagi wrote:
> > diff --git a/include/linux/eventpoll.h b/include/linux/eventpoll.h
> > index 3337745d81bd..38904fce3840 100644
> > --- a/include/linux/eventpoll.h
> > +++ b/include/linux/eventpoll.h
> > @@ -25,6 +25,8 @@ struct file *get_epoll_tfile_raw_ptr(struct file *file, int tfd, unsigned long t
> > /* Used to release the epoll bits inside the "struct file" */
> > void eventpoll_release_file(struct file *file);
> >
> > +void eventpoll_replace_file(struct file *toreplace, struct file *file);
> > +
> > /*
> > * This is called from inside fs/file_table.c:__fput() to unlink files
> > * from the eventpoll interface. We need to have this facility to cleanup
> > @@ -53,6 +55,22 @@ static inline void eventpoll_release(struct file *file)
> > eventpoll_release_file(file);
> > }
> >
> > +
> > +/*
> > + * This is called from fs/file.c:do_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
> > + */
> > +static inline void eventpoll_replace(struct file *toreplace, struct file *file)
> > +{
> > + /*
> > + * toreplace is the file being replaced. Install the new file for the
> > + * existing one that is linked in the epoll interface
> > + */
> > + eventpoll_replace_file(toreplace, file);
> > +}
>
> Why do we have both eventpoll_replace() and eventpoll_replace_file()?
> They seem identical?
I was following the eventpoll_release_file() convention but yes seems
unnecessary. I'll retain eventpoll_replace_file() and remove
eventpoll_replace() in v4.
>
> > diff --git a/include/linux/file.h b/include/linux/file.h
> > index 39704eae83e2..80e56b2b44fb 100644
> > --- a/include/linux/file.h
> > +++ b/include/linux/file.h
> > @@ -36,6 +36,7 @@ struct fd {
> > struct file *file;
> > unsigned int flags;
> > };
> > +
> > #define FDPUT_FPUT 1
> > #define FDPUT_POS_UNLOCK 2
> >
>
> You should drop this hunk of the patch.
I overlooked this. Will fix it in v4.
