Re: [PATCH v6 3/5] mm: introduce mmap3 for safely defining new mmap flags
From: Kirill A. Shutemov
Date: Sat Aug 26 2017 - 19:56:59 EST
On Sat, Aug 26, 2017 at 03:46:22PM -0700, Dan Williams wrote:
> On Sat, Aug 26, 2017 at 12:50 PM, Helge Deller <deller@xxxxxx> wrote:
> > On 26.08.2017 17:15, Dan Williams wrote:
> [..]
> >> I have not seen any patches for parisc pmem+dax enabling so it seems
> >> too early to worry about these "last mile" enabling features of
> >> MAP_DIRECT and MAP_SYNC. In particular parisc doesn't appear to have
> >> ARCH_ENABLE_MEMORY_HOTPLUG, so as far as I can see it can't yet
> >> support the ZONE_DEVICE scheme that is a pre-requisite for MAP_DIRECT.
> >
> > I see, but then it's probably best to not to define any MAP_DIRECT or
> > MAP_SYNC at all in the headers of those arches which don't support
> > pmem+dax (parisc, m68k, alpha, and probably quite some others).
> > That way applications can detect at configure time if the platform
> > supports that, and can leave out the functionality completely.
>
> Yes, that's a good idea we can handle this similar to
> CONFIG_MMAP_ALLOW_UNINITIALIZED. These patches will also modify
> 'struct file_operations' so that do_mmap() can validate whether a flag
> is supported on per architecture basis. Also the plan is to plumb the
> flags passed to the syscall all the way down to the individual mmap
> implementations. The ext4 and xfs ->mmap() operations will be able to
> return -EOPNOTSUP based on runtime variables.
BTW, we may be able to reuse the bit used for MAP_UNINITIALIZED -- it's
only used on !MMU machines.
--
Kirill A. Shutemov