On 2/27/23 14:21, Nina Schoetterl-Glausch wrote:
On Wed, 2023-02-22 at 15:20 +0100, Pierre Morel wrote:
On interception of STSI(15.1.x) the System Information Block
(SYSIB) is built from the list of pre-ordered topology entries.
Signed-off-by: Pierre Morel <pmorel@xxxxxxxxxxxxx>
---
include/hw/s390x/cpu-topology.h | 21 +++
include/hw/s390x/sclp.h | 1 +
target/s390x/cpu.h | 72 ++++++++
hw/s390x/cpu-topology.c | 14 +-
target/s390x/kvm/cpu_topology.c | 312 ++++++++++++++++++++++++++++++++
target/s390x/kvm/kvm.c | 5 +-
target/s390x/kvm/meson.build | 3 +-
7 files changed, 425 insertions(+), 3 deletions(-)
create mode 100644 target/s390x/kvm/cpu_topology.c
diff --git a/include/hw/s390x/cpu-topology.h b/include/hw/s390x/cpu-topology.h
index fa7f885a9f..8dc42d2942 100644
--- a/include/hw/s390x/cpu-topology.h
+++ b/include/hw/s390x/cpu-topology.h
@@ -16,8 +16,29 @@
#define S390_TOPOLOGY_CPU_IFL 0x03
+typedef union s390_topology_id {
+ uint64_t id;
+ struct {
+ uint8_t level5;
You could rename this to sentinel, since that's the only use case and
if there ever is another level the sentinel implementation might need
to be changed anyway.
OK
+ uint8_t drawer;
+ uint8_t book;
+ uint8_t socket;
+ uint8_t dedicated;
+ uint8_t entitlement;
+ uint8_t type;
+ uint8_t origin;
+ };
+} s390_topology_id;
+
[...]
diff --git a/target/s390x/cpu.h b/target/s390x/cpu.h
index d654267a71..c899f4e04b 100644
--- a/target/s390x/cpu.h
+++ b/target/s390x/cpu.h
@@ -560,6 +560,25 @@ typedef struct SysIB_322 {
[...]
+/*
+ * CPU Topology List provided by STSI with fc=15 provides a list
+ * of two different Topology List Entries (TLE) types to specify
+ * the topology hierarchy.
+ *
+ * - Container Topology List Entry
+ * Defines a container to contain other Topology List Entries
+ * of any type, nested containers or CPU.
+ * - CPU Topology List Entry
+ * Specifies the CPUs position, type, entitlement and polarization
+ * of the CPUs contained in the last Container TLE.
+ *
+ * There can be theoretically up to five levels of containers, QEMU
+ * uses only three levels, the drawer's, book's and socket's level.
+ *
+ * A container of with a nesting level (NL) greater than 1 can only
s/of//
thanks.
+ * contain another container of nesting level NL-1.
+ *
+ * A container of nesting level 1 (socket), contains as many CPU TLE
+ * as needed to describe the position and qualities of all CPUs inside
+ * the container.
+ * The qualities of a CPU are polarization, entitlement and type.
+ *
+ * The CPU TLE defines the position of the CPUs of identical qualities
+ * using a 64bits mask which first bit has its offset defined by
+ * the CPU address orgin field of the CPU TLE like in:
+ * CPU address = origin * 64 + bit position within the mask
+ *
+ */
+/* Container type Topology List Entry */
+typedef struct SysIBTl_container {
+ uint8_t nl;
+ uint8_t reserved[6];
+ uint8_t id;
+} QEMU_PACKED QEMU_ALIGNED(8) SysIBTl_container;
+QEMU_BUILD_BUG_ON(sizeof(SysIBTl_container) != 8);
+
[...]
+
+/**
+ * s390_topology_from_cpu:
+ * @cpu: The S390CPU
+ *
+ * Initialize the topology id from the CPU environment.
+ */
+static s390_topology_id s390_topology_from_cpu(S390CPU *cpu)
+{
+ s390_topology_id topology_id = {0};
+
+ topology_id.drawer = cpu->env.drawer_id;
+ topology_id.book = cpu->env.book_id;
+ topology_id.socket = cpu->env.socket_id;
+ topology_id.origin = cpu->env.core_id / 64;
+ topology_id.type = S390_TOPOLOGY_CPU_IFL;
+ topology_id.dedicated = cpu->env.dedicated;
+
+ if (s390_topology.polarization == S390_CPU_POLARIZATION_VERTICAL) {
+ /*
+ * Vertical polarization with dedicated CPU implies
+ * vertical high entitlement.
+ */
+ if (topology_id.dedicated) {
+ topology_id.entitlement = S390_CPU_ENTITLEMENT_HIGH;
+ } else {
+ topology_id.entitlement = cpu->env.entitlement;
+ }
I don't see why you need this if, it should already be correct.
+ }
I'd suggest the following:
* rename entitlement in s390_topology_id back to polarization, but keep entitlement everywhere else.
* remove horizontal/none from CpuS390Entitlement, this way the user cannot set it,
and it doesn't show up in the output of query-cpus-fast.
* this is where you convert between the two, so:
if horizontal, id.polarization = 0,
otherwise id.polarization = entitlement + 1, or a switch case.
* in patch 6 in s390_topology_set_cpus_entitlement you don't set the entitlement if the polarization
is horizontal, which is ok because of the conversion above.
I do not like to remove the horizontal entitlement from the enum because
every
existing s390 tool show "horizontal" entitlement when the polarization
is horizontal.
See, lscpu -e or "/sys/devices/system/cpu/cpu*/polarization"
Also, the user may find strange that a field is missing depending of the
polarization in the output of query-cpu-fast.
Regards,
Pierre
