mm: ptep_deref() conversion

Convert all instances of direct pte_t* dereferencing to instead use
ptep_deref() helper. By default, the helper does a direct dereference as
before, but it can (and will) be overridden by the architecture to fully
encapsulate the contents of the pte. Arch code is deliberately not
converted, as the arch code knows best.

Conversion was done using Coccinelle:

----

// $ make coccicheck \
//          COCCI=ptepderef.cocci \
//          SPFLAGS="--include-headers" \
//          MODE=patch

virtual patch

@ depends on patch @
pte_t *v;
@@

- *v
+ ptep_deref(v)

----

Then reviewed and hand-edited to avoid multiple unnecessary calls to
ptep_deref(), instead opting to store the result of a single in a
variable, where it is correct to do so. This will benefit arch-overrides
that may be more complex than a simple (optimizable) pointer
dereference.

Included is a fix for an issue in an earlier version of this patch that
was pointed out by kernel test robot. The issue arose because config
MMU=n elides definition of the ptep helper functions, including
ptep_deref(). HUGETLB_PAGE=n configs still define a simple
huge_ptep_clear_flush() for linking purposes, which dereferences the
ptep. So when both configs are disabled, this caused a build error
because ptep_deref() is not defined. Fix by continuing to do a direct
dereference when MMU=n. This is safe because for this config the arch
code cannot be trying to virtualize the ptes because none of the ptep
helpers are defined.

Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202305120142.yXsNEo6H-lkp@intel.com/


Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>