powerpc/mm: Move hash table ops to a separate structure
authorBenjamin Herrenschmidt <benh@kernel.crashing.org>
Tue, 5 Jul 2016 05:03:58 +0000 (15:03 +1000)
committerMichael Ellerman <mpe@ellerman.id.au>
Thu, 21 Jul 2016 08:59:09 +0000 (18:59 +1000)
commit7025776ed1ebdfa1959932e7a4662c2f88607df0
tree4b1b0e0be66d8a0b9d2a266a036416c01877ea41
parentb521f576df6c49fcc06fbc06a349f7590f223a38
powerpc/mm: Move hash table ops to a separate structure

Moving probe_machine() to after mmu init will cause the ppc_md
fields relative to the hash table management to be overwritten.

Since we have essentially disconnected the machine type from
the hash backend ops, finish the job by moving them to a different
structure.

The only callback that didn't quite fix is update_partition_table
since this is not specific to hash, so I moved it to a standalone
variable for now. We can revisit later if needed.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
[mpe: Fix ppc64e build failure in kexec]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
13 files changed:
arch/powerpc/include/asm/book3s/64/mmu-hash.h
arch/powerpc/include/asm/machdep.h
arch/powerpc/kernel/machine_kexec_64.c
arch/powerpc/kernel/misc_64.S
arch/powerpc/kvm/book3s_64_mmu_host.c
arch/powerpc/mm/hash64_4k.c
arch/powerpc/mm/hash64_64k.c
arch/powerpc/mm/hash_native_64.c
arch/powerpc/mm/hash_utils_64.c
arch/powerpc/mm/hugepage-hash64.c
arch/powerpc/mm/hugetlbpage-hash64.c
arch/powerpc/platforms/ps3/htab.c
arch/powerpc/platforms/pseries/lpar.c