Re: [PATCH] fs: use __fput_sync in close(2)

[Matthew Wilcox <willy@xxxxxxxxxxxxx>]

On Mon, Aug 07, 2023 at 01:06:27AM +0200, Mateusz Guzik wrote:
> With the assumption this is not going to work, I wrote my own patch
> which adds close_fd_sync() and filp_close_sync().  They are shipped as
> dedicated func entry points, but perhaps inlines which internally add a
> flag to to the underlying routine would be preferred?

Yes, I think static inlines would be better here.  

> Also adding __ in
> front would be in line with __fput_sync, but having __filp_close_sync
> call  __filp_close looks weird to me.

I'd handle this as ...

int file_close_sync(struct file *, fl_owner_t, bool sync);
static inline filp_close(struct file *file, fl_owner_t owner)
{
	return file_close_sync(file, owner, false);
}