Re: [PATCH] mm: Change return type to vm_fault_t

[Matthew Wilcox <willy@infradead.org>]

On Thu, Mar 08, 2018 at 03:22:44PM -0800, Andrew Morton wrote:
> On Thu, 8 Mar 2018 14:22:01 -0800 Matthew Wilcox <willy@infradead.org> wrote:
> > On Thu, Mar 08, 2018 at 06:35:23PM +0530, Souptick Joarder wrote:
> > > Use new return type vm_fault_t for fault handler
> > > in struct vm_operations_struct.
> > > 
> > > vmf_insert_mixed(), vmf_insert_pfn() and vmf_insert_page()
> > > are newly added inline wrapper functions.
> > > 
> > > Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
> > 
> > Reviewed-by: Matthew Wilcox <mawilcox@microsoft.com>
> > 
> > Andrew, the plan for these patches is to introduce the typedef, initially
> > just as documentation ("This function should return a VM_FAULT_ status").
> > We'll trickle the patches to individual drivers/filesystems in through
> > the maintainers, as far as possible.  In a few months, we'll probably
> > dump a pile of patches to unloved drivers on you for merging.  Then we'll
> > change the typedef to an unsigned int and break the compilation of any
> > unconverted driver.
> > 
> > Souptick has done a few dozen drivers already, and I've been doing my best
> > to keep up with reviewing the patches submitted.  There's some interesting
> > patterns and commonalities between drivers (not to mention a few outright
> > bugs) that we've noticed, and this'll be a good time to clean them up.
> 
> OK.  All of this should be in the changelog, please.  Along with a full
> explanation of the reasons for adding the new functions.

Agreed.  Souptick, can you take care of doing that and resubmitting
the patch?

> > It'd be great to get this into Linus' tree sooner so we can start
> > submitting the patches to the driver maintainers.
> 
> Sure.  I assume that vm_fault_t is `int', so this bare patch won't
> cause a ton of type mismatch warnings?

Exactly so.