This series fixes multiple TCG issues related to storage key instructions,
fixes some TCG issues related to LOAD REAL ADDRESS and skeys, implements
lazy skey enablement under TCG, and prepares the whole skey infrastructure
for dealing with addresses beyond initial RAM (e.g., memory devices like
virtio-mem). Along, some cleanups.
To prepare for memory devices / memory hotplug, my goal was to get rid of
as many ms->ram_size users as possible. Unfortunately, I stumbled over
many other things on the way. The remaining s390x users of ms->ram_size
are:
a) hw/s390x/ipl.c: loading the FW. Won't need adjustment.
b) hw/s390x/s390-skeys.c: allocating the array for storage keys. Will need
adjustment for memory devices.
c) hw/s390x/s390-stattrib-kvm.c: will need adjustments in the future when
using memory devices with CMM.
d) hw/s390x/s390-virtio-ccw.c: fixing up / handling initital ram. Won't
need adjustment.
e) hw/s390x/sclp.c: same as d)
Especially patch 9-12 also affect KVM. The remaining ones mostly only
affect TCG.
v2 -> v3:
- "s390x/tcg: check for addressing exceptions for RRBE, SSKE and ISKE"
-- Use trigger_pgm_exception() to trigger PGM_ADDRESSING
v1 -> v2:
- "s390x/tcg: fix ignoring bit 63 when setting the storage key in SSKE"
-- Extended description
- "s390x/tcg: convert real to absolute address for RRBE, SSKE and ISKE"
-- Extended description
- "s390x/tcg: check for addressing exceptions for RRBE, SSKE and ISKE"
-- Fix compilation issue
- "s390x/mmu_helper: no need to pass access type to mmu_translate_asce()"
-- Mention commit id in description
- "s390x/mmu_helper: move address validation into mmu_translate*()"
-- Introduce and use MMU_S390_LRA
- "hw/s390x/s390-skeys: check if an address is valid before dumping the
key"
-- Add missing return
- "hw/s390x/s390-skeys: rename skeys_enabled to skeys_are_enabled"
-- Added
- "hw/s390x/s390-skeys: lazy storage key enablement under TCG"
-- Adjust error cases/message in TCG get_skeys() and set_skeys()
-- Use "enable_skeys" and return a bool
-- Adjust function documentation
Cc: Richard Henderson <richard.henderson@linaro.org>
Cc: Cornelia Huck <cohuck@redhat.com>
Cc: Halil Pasic <pasic@linux.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Thomas Huth <thuth@redhat.com>
Cc: Janosch Frank <frankja@linux.ibm.com>
Cc: Claudio Imbrenda <imbrenda@linux.ibm.com>
Cc: Jason J. Herne <jjherne@linux.ibm.com>
Cc: qemu-s390x@nongnu.org
David Hildenbrand (13):
s390x/tcg: wrap address for RRBE
s390x/tcg: fix ignoring bit 63 when setting the storage key in SSKE
s390x/tcg: convert real to absolute address for RRBE, SSKE and ISKE
s390x/tcg: check for addressing exceptions for RRBE, SSKE and ISKE
s390x/mmu_helper: no need to pass access type to mmu_translate_asce()
s390x/mmu_helper: fixup mmu_translate() documentation
s390x/mmu_helper: move address validation into mmu_translate*()
s390x/mmu_helper: avoid setting the storage key if nothing changed
hw/s390x/s390-skeys: use memory mapping to detect which storage keys
to migrate
hw/s390x/s390-skeys: use memory mapping to detect which storage keys
to dump
hw/s390x/s390-skeys: check if an address is valid before dumping the
key
hw/s390x/s390-skeys: rename skeys_enabled to skeys_are_enabled
hw/s390x/s390-skeys: lazy storage key enablement under TCG
hw/s390x/s390-skeys-kvm.c | 4 +-
hw/s390x/s390-skeys.c | 206 +++++++++++++++++++++-----------
include/hw/s390x/storage-keys.h | 65 +++++++++-
target/s390x/helper.h | 6 +-
target/s390x/mmu_helper.c | 70 +++++++----
target/s390x/s390x-internal.h | 3 +
target/s390x/tcg/excp_helper.c | 13 --
target/s390x/tcg/mem_helper.c | 53 +++++---
8 files changed, 294 insertions(+), 126 deletions(-)
--
2.31.1