s390x: CPU Topology
Subsystems (2)
v15 cover s390x: CPU Topology
Hi,
Two important changes in this series:
- The ordering of the Topology List Entries, TLE, is done just before building the System information Buffer, SYSIB, during the interception of STSI.
The adventage of this is to do the ordering only if it is needed whatever the reason for a change of the topology is, hotplug, qapi, instruction PTF(2) from the guest.
- query-topology is no longer needed, it is replaced by an extension of query-cpus-fast.
A small change that lead me to remove @Thomas RB:
- Add a trampolin function in cpu-topology code for the subsystem reset for extending to PTF reset in the following patch on PTF implementation.
@Nina and @Nico RB have been removed from reset patch too in the last series because of small changes too.
Implementation discussions
==========================
CPU models
----------
Since the facility 11, S390_FEAT_CONFIGURATION_TOPOLOGY is already in the CPU model for old QEMU we could not activate it as usual from KVM but needed a KVM capability: KVM_CAP_S390_CPU_TOPOLOGY. Checking and enabling this capability enables facility 11, S390_FEAT_CONFIGURATION_TOPOLOGY.
It is the responsibility of the admin to ensure the same CPU model for source and target host in a migration.
Migration
---------
When the target guest is started, the Multi-processor Topology Change Report (MTCR) bit is set during the creation of the vCPU by KVM. We do not need to migrate its state, in the worst case, the target guest will see the MTCR and actualize its view of the topology without necessity, but this will be done only one time.
Reset
-----
Reseting the topology is done during subsystem reset, the polarization is reset to horizontal polarization.
Topology attributes
-------------------
The topology attributes are carried by the CPU object and defined on object creation.
In the case the new attributes, socket, book, drawer, dedicated, polarity are not provided QEMU provides defaults values.
- Geometry defaults
The geometry default are based on the core-id of the core to fill the geometry in a monotone way starting with drawer 0, book 0, and filling socket 0 with the number of cores per socket, then filling socket 1, socket 2 ... etc until the book is complete and all books until the first drawer is complete before starting with the next drawer.
This allows to keep existing start scripts and Libvirt existing interface until it is extended.
- Modifiers defaults
Default polarization is horizontal
Default dedication is not dedicated.
Dynamic topology modification
-----------------------------
QAPI interface is extended with:
- a command: 'x-set-cpu-topology'
- a query: extension of 'query-cpus-fast'
- an event: 'CPU_POLARITY_CHANGE'
The admin may use query-cpus-fast to verify the topology provided to the guest and x-set-cpu-topology to modify it.
The event CPU_POLARITY_CHANGE is sent when the guest successfuly uses the PTF(2) instruction to request a polarization change. In that case, the admin is supposed to modify the CPU provisioning accordingly.
Testing
=======
To use the QEMU patches, you will need Linux V6-rc1 or newer, or use the following Linux mainline patches:
f5ecfee94493 2022-07-20 KVM: s390: resetting the Topology-Change-Report 24fe0195bc19 2022-07-20 KVM: s390: guest support for topology function 0130337ec45b 2022-07-20 KVM: s390: Cleanup ipte lock access and SIIF fac..
Currently this code is for KVM only, I have no idea if it is interesting to provide a TCG patch. If ever it will be done in another series.
Documentation
=============
To have a better understanding of the S390x CPU Topology and its implementation in QEMU you can have a look at the documentation in the last patch of this series.
The admin will want to match the host and the guest topology, taking into account that the guest does not recognize multithreading. Consequently, two vCPU assigned to threads of the same real CPU should preferably be assigned to the same socket of the guest machine.
Regards,
Pierre
Pierre Morel (11):
s390x/cpu topology: adding s390 specificities to CPU topology s390x/cpu topology: add topology entries on CPU hotplug target/s390x/cpu topology: handle STSI(15) and build the SYSIB s390x/sclp: reporting the maximum nested topology entries s390x/cpu topology: resetting the Topology-Change-Report s390x/cpu topology: interception of PTF instruction target/s390x/cpu topology: activating CPU topology qapi/s390x/cpu topology: x-set-cpu-topology monitor command
machine: adding s390 topology to query-cpu-fast qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event docs/s390x/cpu topology: document s390x cpu topology
docs/system/s390x/cpu-topology.rst | 294 ++++++++++++++++++ docs/system/target-s390x.rst | 1 + qapi/machine-target.json | 59 ++++ qapi/machine.json | 27 +- include/hw/boards.h | 10 +- include/hw/s390x/cpu-topology.h | 71 +++++ include/hw/s390x/s390-virtio-ccw.h | 6 + include/hw/s390x/sclp.h | 4 +- include/monitor/hmp.h | 1 + target/s390x/cpu.h | 79 +++++ target/s390x/kvm/kvm_s390x.h | 1 + hw/core/machine-qmp-cmds.c | 2 + hw/core/machine-smp.c | 48 ++- hw/core/machine.c | 4 + hw/s390x/cpu-topology.c | 471 +++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 28 +- hw/s390x/sclp.c | 5 + softmmu/vl.c | 6 + target/s390x/cpu-sysemu.c | 27 ++ target/s390x/cpu.c | 7 + target/s390x/cpu_models.c | 1 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++ target/s390x/kvm/kvm.c | 45 ++- hmp-commands.hx | 16 + hw/s390x/meson.build | 1 + qemu-options.hx | 7 +- target/s390x/kvm/meson.build | 3 +- 27 files changed, 1539 insertions(+), 20 deletions(-) create mode 100644 docs/system/s390x/cpu-topology.rst create mode 100644 include/hw/s390x/cpu-topology.h create mode 100644 hw/s390x/cpu-topology.c create mode 100644 target/s390x/kvm/cpu_topology.c
Signature · 181 lines
2.31.1 Since v14: - move the ordering of TLE to just before filling the SYSIB, optimize TLE ordering to be done on need only. (Cedric in previous series) - remove 'query-topology' and instead extend 'query-cpus-fast' (Daniel) - rename POLARITY_CHANGE to CPU_POLARITY_CHANGE (Thomas) - Divers bugs correction and doc changes (Thomas, Nina) - Separate topology and entitlement, simplify pft handling (Nina) - add the resetting of all CPU to horizontal polarity once implementing PTF interpretation Since v13: - Suppress the topology device to simplify the code (Cedric) - moved reset of MTCR from device reset into subsystem reset and removed previous reviewed-by from Nico and Janis - No need for Migration - No need for machine dependencies (Christian, Thomas) - Adding all features, drawer/book and dynamic (Cedric) - since v12 - suppress new CPU flag "disable-topology" just use ctop - no use of special fields in CCW machine or in CPU - modifications in documentation - insert documentation in tree (Cedric) - moved cpu-topology.c from target/s390 to target/s390/kvm to compile smoothly (without topology) for TCG (Cedric) - since v11 - new CPU flag "disable-topology" I would have take "topology" if I was able to have it false on default. (Christian, Thomas) - Build the topology during the interception of the STSI instruction. (Cedric) - return CC3 in case the calculated SYSIB length is greater than 4096. (Janis) - minor corections on documentation - since v10 - change machine attribute "topology-disable" to "topology" (Cedric) - Add preliminary patch for machine properties (Cedric) - Use next machine as 7.2 (Cedric / Connie) - Remove unecessary mutex (Thomas) - use ENOTSUP return value for kvm_s390_topology_set_mtcr() (Cedric) - Add explanation on container and cpu TLEs (Thomas) - use again cpu and socket count in topology structure (Cedric) - Suppress the S390TopoTLE structure and integrate the TLE masks to the socket structure. (-) - the STSI instruction now finds the topology from the machine (Cedric) - since v9 - remove books and drawers - remove thread denying and replace with a merge of cores * threads to specify the CPUs available to the guest - add a class option to avoid topology on older machines (Cedric) - Allocate a SYSIB buffer of the maximal length to avoid overflow. (Nico, Janis) - suppress redundancy of smp parameters in topology and use directly the machine smp structure - Early check for topology support (Cedric) - since v8 - Linux patches are now mainline - simplification of the implementation (Janis) - Migration, new machine definition (Thomas) - Documentation - since v7 - Coherence with the Linux patch series changes for MTCR get (Pierre) - check return values during new CPU creation (Thomas) - Improving codding style and argument usages (Thomas) - since v6 - Changes on smp args in qemu-options (Daniel) - changed comments in machine.jason (Daniel) - Added reset (Janosch) - since v5 - rebasing on newer QEMU version - reworked most lines above 80 characters. - since v4 - Added drawer and books to topology - Added numa topology - Added documentation - since v3 - Added migration (Thomas) - Separated STSI instruction from KVM to prepare TCG (Thomas) - Take care of endianess to prepare TCG (Thomas) - Added comments on STSI CPU container and PFT instruction (Thomas) - Moved enabling the instructions as the last patch (Thomas)
PATCH v15 1/11 s390x/cpu topology: adding s390 specificities to CPU topology
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
--- qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++- include/hw/s390x/cpu-topology.h | 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++ softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++ qemu-options.hx | 7 +++-- 10 files changed, 115 insertions(+), 12 deletions(-) create mode 100644 include/hw/s390x/cpu-topology.h
qapi/machine.json +12 -2
@@ -900,13 +900,15 @@ # a CPU is being hotplugged. # # @node-id: NUMA node ID the CPU belongs to -# @socket-id: socket number within node/board the CPU belongs to +# @drawer-id: drawer number within node/board the CPU belongs to (since 8.0) +# @book-id: book number within drawer/node/board the CPU belongs to (since 8.0) +# @socket-id: socket number within book/node/board the CPU belongs to # @die-id: die number within socket the CPU belongs to (since 4.1) # @cluster-id: cluster number within die the CPU belongs to (since 7.1) # @core-id: core number within cluster the CPU belongs to # @thread-id: thread number within core the CPU belongs to # -# Note: currently there are 6 properties that could be present +# Note: currently there are 8 properties that could be present # but management should be prepared to pass through other # properties with device_add command to allow for future # interface extension. This also requires the filed names to be kept in
@@ -916,6 +918,8 @@ ## { 'struct': 'CpuInstanceProperties', 'data': { '*node-id': 'int', + '*drawer-id': 'int', + '*book-id': 'int', '*socket-id': 'int', '*die-id': 'int', '*cluster-id': 'int',
@@ -1465,6 +1469,10 @@ # # @cpus: number of virtual CPUs in the virtual machine # +# @drawers: number of drawers in the CPU topology (since 8.0) +# +# @books: number of books in the CPU topology (since 8.0) +# # @sockets: number of sockets in the CPU topology # # @dies: number of dies per socket in the CPU topology
@@ -1481,6 +1489,8 @@ ## { 'struct': 'SMPConfiguration', 'data': { '*cpus': 'int', + '*drawers': 'int', + '*books': 'int', '*sockets': 'int', '*dies': 'int', '*clusters': 'int',
include/hw/boards.h +9 -1
@@ -131,12 +131,16 @@ typedef struct {* @clusters_supported - whether clusters are supported by the machine * @has_clusters - whether clusters are explicitly specified in the user * provided SMP configuration +* @books_supported - whether books are supported by the machine +* @drawers_supported - whether drawers are supported by the machine */ typedef struct {bool prefer_sockets;bool dies_supported;bool clusters_supported;bool has_clusters; +bool books_supported; +bool drawers_supported; }SMPCompatProps ;/**
@@ -301,7 +305,9 @@ typedef struct DeviceMemoryState {/** * CpuTopology: * @cpus: the number of present logical processors on the machine -* @sockets: the number of sockets on the machine +* @drawers: the number of drawers on the machine +* @books: the number of books in one drawer +* @sockets: the number of sockets in one book * @dies: the number of dies in one socket * @clusters: the number of clusters in one die * @cores: the number of cores in one cluster
@@ -310,6 +316,8 @@ typedef struct DeviceMemoryState {*/ typedef struct CpuTopology { unsigned int cpus; + unsigned int drawers; + unsigned int books; unsigned int sockets; unsigned int dies; unsigned int clusters;
include/hw/s390x/cpu-topology.h new file +24 -0
@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H + +#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { +POLARITY_HORIZONTAL , +POLARITY_VERTICAL , +POLARITY_VERTICAL_LOW =1 , +POLARITY_VERTICAL_MEDIUM , +POLARITY_VERTICAL_HIGH , +POLARITY_MAX , +}; + +#endif
target/s390x/cpu.h +5 -0
@@ -131,6 +131,11 @@ struct CPUArchState {#if !defined(CONFIG_USER_ONLY )uint32_t core_id;/* PoP "CPU address", same as cpu_index */ +int32_t socket_id; +int32_t book_id; +int32_t drawer_id; +bool dedicated; +uint8_t entitlement;/* Used only for vertical polarization */ uint64_t cpuid;#endif
hw/core/machine-smp.c +41 -7
105 lines not on this page · raw message
hw/core/machine.c +4 -0
@@ -828,6 +828,8 @@ static void machine_get_smp(Object *obj, Visitor *v, const char *name,MachineState *ms =MACHINE (obj);SMPConfiguration *config = &(SMPConfiguration ){ .has_cpus = true, .cpus = ms->smp.cpus, + .has_drawers = true, .drawers = ms->smp.drawers, + .has_books = true, .books = ms->smp.books, .has_sockets = true, .sockets = ms->smp.sockets, .has_dies = true, .dies = ms->smp.dies, .has_clusters = true, .clusters = ms->smp.clusters,
@@ -1093,6 +1095,8 @@ static void machine_initfn(Object *obj)/* default to mc->default_cpus */ ms->smp.cpus = mc->default_cpus; ms->smp.max_cpus = mc->default_cpus; + ms->smp.drawers =1 ; + ms->smp.books =1 ; ms->smp.sockets =1 ; ms->smp.dies =1 ; ms->smp.clusters =1 ;
hw/s390x/s390-virtio-ccw.c +2 -0
@@ -733,6 +733,8 @@ static void ccw_machine_class_init(ObjectClass *oc, void *data) mc->no_sdcard =1 ; mc->max_cpus =S390_MAX_CPUS ; mc->has_hotpluggable_cpus = true; + mc->smp_props.books_supported = true; + mc->smp_props.drawers_supported = true;assert (!mc->get_hotplug_handler); mc->get_hotplug_handler = s390_get_hotplug_handler; mc->cpu_index_to_instance_props = s390_cpu_index_to_props;
softmmu/vl.c +6 -0
@@ -727,6 +727,12 @@ static QemuOptsList qemu_smp_opts = { { .name = "cpus", .type = QEMU_OPT_NUMBER, + }, { + .name = "drawers", + .type = QEMU_OPT_NUMBER, + }, { + .name = "books", + .type = QEMU_OPT_NUMBER, }, { .name = "sockets", .type = QEMU_OPT_NUMBER,
target/s390x/cpu.c +7 -0
@@ -37,6 +37,7 @@#ifndef CONFIG_USER_ONLY #include "sysemu/reset.h" #endif +#include "hw/s390x/cpu-topology.h" #define CR0_RESET 0xE0UL#define CR14_RESET 0xC2000000UL;
@@ -259,6 +260,12 @@ static gchar *s390_gdb_arch_name(CPUState *cs)static Property s390x_cpu_properties[] = { #if !defined(CONFIG_USER_ONLY)DEFINE_PROP_UINT32 ("core-id" ,S390CPU , env.core_id,0 ), +DEFINE_PROP_INT32 ("socket-id" ,S390CPU , env.socket_id,-1 ), +DEFINE_PROP_INT32 ("book-id" ,S390CPU , env.book_id,-1 ), +DEFINE_PROP_INT32 ("drawer-id" ,S390CPU , env.drawer_id,-1 ), +DEFINE_PROP_BOOL ("dedicated" ,S390CPU , env.dedicated, false), +DEFINE_PROP_UINT8 ("polarity" ,S390CPU , env.entitlement, +POLARITY_VERTICAL_MEDIUM ), #endifDEFINE_PROP_END_OF_LIST () };
qemu-options.hx +5 -2
@@ -250,11 +250,14 @@ SRST ERST DEF("smp", HAS_ARG, QEMU_OPTION_smp, - "-smp [[cpus=]n][,maxcpus=maxcpus][,sockets=sockets][,dies=dies][,clusters=clusters][,cores=cores][,threads=threads]\n" + "-smp [[cpus=]n][,maxcpus=maxcpus][,drawers=drawers][,books=books][,sockets=sockets]\n" + " [,dies=dies][,clusters=clusters][,cores=cores][,threads=threads]\n" " set the number of initial CPUs to 'n' [default=1]\n" " maxcpus= maximum number of total CPUs, including\n" " offline CPUs for hotplug, etc\n" - " sockets= number of sockets on the machine board\n" + " drawers= number of drawers on the machine board\n" + " books= number of books in one drawer\n" + " sockets= number of sockets in one book\n" " dies= number of dies in one socket\n" " clusters= number of clusters in one die\n" " cores= number of cores in one cluster\n"
Signature · 1 line
2.31.1
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
Quoted 3 lines · Pierre Morel
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication
Nit: s/toplogy/topology/
Quoted 9 lines
and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>---
Apart from the nit:
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
Quoted 11 lines · Thomas Huth
Quoted 3 lines · Pierre Morel
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedicationNit: s/toplogy/topology/
Quoted 4 lines
and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Quoted 9 lines
Quoted 5 lines
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>---
Apart from the nit:
Reviewed-by Thomas Huth <[email protected]>
Thanks,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
In reply to Pierre MorelNit patch title: s390x/cpu topology: add s390 specifics to CPU topology ?
Quoted 3 lines · Pierre Morel
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication
^o
Quoted 22 lines
and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++- include/hw/s390x/cpu-topology.h | 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++ softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++ qemu-options.hx | 7 +++-- 10 files changed, 115 insertions(+), 12 deletions(-) create mode 100644 include/hw/s390x/cpu-topology.h
[...]
Quoted 19 lines
include/hw/s390x/cpu-topology.h new file +12
-0@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H +
include/hw/s390x/cpu-topology.h line 22
Quoted 10 lines
include/hw/s390x/cpu-topology.h new file +10
-0+#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { +POLARITY_HORIZONTAL , +POLARITY_VERTICAL , +POLARITY_VERTICAL_LOW =1 , +POLARITY_VERTICAL_MEDIUM , +POLARITY_VERTICAL_HIGH , +POLARITY_MAX , +};
Probably a good idea to keep the S390 prefix. This works, but aliasing VERTICAL and VERTICAL_LOW is not entirely straight forward.
Why not have two enum?
enum s390_topology_polarity {
S390_POLARITY_HORIZONTAL,
S390_POLARITY_VERTICAL,
};enum s390_topology_entitlement {
S390_ENTITLEMENT_LOW = 1,
S390_ENTITLEMENT_MEDIUM,
S390_ENTITLEMENT_HIGH,
S390_ENTITLEMENT_MAX,
};Maybe add an ENTITLEMENT_INVALID/NONE, if you need that, as first value.
Quoted 2 lines
include/hw/s390x/cpu-topology.h new file +1
-0+#endif
[...]
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
@Thomas, these changes look good to me.
What do you think, do I make the change and keep your RB ?
Quoted 2 lines · Nina Schoetterl-Glausch
Nit patch title: s390x/cpu topology: add s390 specifics to CPU topology ?
OK
Quoted 5 lines · Pierre Morel
Quoted 3 lines
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication^o
Yes thx
Quoted 63 lines
Quoted 22 lines
and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++- include/hw/s390x/cpu-topology.h | 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++ softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++ qemu-options.hx | 7 +++-- 10 files changed, 115 insertions(+), 12 deletions(-)create mode 100644
include/hw/s390x/cpu-topology.h[...]
Quoted 19 lines
include/hw/s390x/cpu-topology.h new file +12
-0@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H +include/hw/s390x/cpu-topology.h line 22
Quoted 10 lines
include/hw/s390x/cpu-topology.h new file +10
-0+#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { +POLARITY_HORIZONTAL , +POLARITY_VERTICAL , +POLARITY_VERTICAL_LOW =1 , +POLARITY_VERTICAL_MEDIUM , +POLARITY_VERTICAL_HIGH , +POLARITY_MAX , +};Probably a good idea to keep the S390 prefix. This works, but aliasing VERTICAL and
VERTICAL_LOWis not entirely straight forward.Why not have two enum?
enum s390_topology_polarity { S390_POLARITY_HORIZONTAL, S390_POLARITY_VERTICAL, };
Quoted 8 lines
enum s390_topology_entitlement { S390_ENTITLEMENT_LOW = 1, S390_ENTITLEMENT_MEDIUM, S390_ENTITLEMENT_HIGH, S390_ENTITLEMENT_MAX, };Maybe add an ENTITLEMENT_INVALID/NONE, if you need that, as first value.
If Thomas agree, I do the changes.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
Quoted 3 lines · Pierre Morel
@Thomas, these changes look good to me.
What do you think, do I make the change and keep your RB ?
Yes, splitting that enum sounds cleaner, indeed.
Thomas
Quoted 92 lines · Nina Schoetterl-Glausch
Quoted 2 lines
Nit patch title: s390x/cpu topology: add s390 specifics to CPU topology ?
OK
Quoted 5 lines · Pierre Morel
Quoted 3 lines
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication^o
Yes thx
Quoted 63 lines
Quoted 22 lines
and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.
Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>---
qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++-include/hw/s390x/cpu-topology.h| 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++
softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++
qemu-options.hx | 7 +++--10 files changed, 115 insertions(+), 12 deletions(-)create mode 100644
include/hw/s390x/cpu-topology.h[...]
Quoted 19 lines
include/hw/s390x/cpu-topology.h new file +12
-0@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H +include/hw/s390x/cpu-topology.h line 22
Quoted 10 lines
include/hw/s390x/cpu-topology.h new file +10
-0+#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { + POLARITY_HORIZONTAL, + POLARITY_VERTICAL, + POLARITY_VERTICAL_LOW = 1, + POLARITY_VERTICAL_MEDIUM, + POLARITY_VERTICAL_HIGH, + POLARITY_MAX, +};Probably a good idea to keep the S390 prefix. This works, but aliasing VERTICAL and
VERTICAL_LOWis not entirely straight forward.Why not have two enum?
enums390_topology_polarity{
S390_POLARITY_HORIZONTAL,
S390_POLARITY_VERTICAL,
};Quoted 8 lines
enum
s390_topology_entitlement{
S390_ENTITLEMENT_LOW = 1,
S390_ENTITLEMENT_MEDIUM,
S390_ENTITLEMENT_HIGH,
S390_ENTITLEMENT_MAX,
};
Maybe add an ENTITLEMENT_INVALID/NONE, if you need that, as first value.If Thomas agree, I do the changes.
Regards,
Pierre
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
In reply to Pierre MorelQuoted 25 lines · Pierre Morel
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++- include/hw/s390x/cpu-topology.h | 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++ softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++ qemu-options.hx | 7 +++-- 10 files changed, 115 insertions(+), 12 deletions(-) create mode 100644 include/hw/s390x/cpu-topology.h
[...]
Quoted 19 lines
include/hw/s390x/cpu-topology.h new file +12
-0@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H +
include/hw/s390x/cpu-topology.h line 22
Quoted 10 lines
include/hw/s390x/cpu-topology.h new file +10
-0+#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { +POLARITY_HORIZONTAL , +POLARITY_VERTICAL , +POLARITY_VERTICAL_LOW =1 , +POLARITY_VERTICAL_MEDIUM , +POLARITY_VERTICAL_HIGH , +POLARITY_MAX , +};
IMO you should define the polarization and entitlement enums as qapi enums.
The polarization is passed as data in the qapi event when the polarization changes.
And the entitlement is passed to qemu when modifying the topology.
[...]
Re: [PATCH v15 01/11] s390x/cpu topology: adding s390 specificities to CPU topology
Quoted 54 lines · Nina Schoetterl-Glausch
Quoted 25 lines · Pierre Morel
S390 adds two new SMP levels, drawers and books to the CPU topology.
The S390 CPU have specific toplogy features like dedication and polarity to give to the guest indications on the host vCPUs scheduling and help the guest take the best decisions on the scheduling of threads on the vCPUs.Let us provide the SMP properties with books and drawers levels and S390 CPU with dedication and polarity,
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 14 ++++++++-- include/hw/boards.h | 10 ++++++- include/hw/s390x/cpu-topology.h | 24 +++++++++++++++++ target/s390x/cpu.h | 5 ++++ hw/core/machine-smp.c | 48 ++++++++++++++++++++++++++++----- hw/core/machine.c | 4 +++ hw/s390x/s390-virtio-ccw.c | 2 ++ softmmu/vl.c | 6 +++++ target/s390x/cpu.c | 7 +++++ qemu-options.hx | 7 +++-- 10 files changed, 115 insertions(+), 12 deletions(-)create mode 100644
include/hw/s390x/cpu-topology.h[...]
Quoted 27 lines
include/hw/s390x/cpu-topology.h new file +20
-0@@ -0,0 +1,24 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ +#ifndef HW_S390X_CPU_TOPOLOGY_H +#define HW_S390X_CPU_TOPOLOGY_H + +#define S390_TOPOLOGY_CPU_IFL 0x03 + +enum s390_topology_polarity { +POLARITY_HORIZONTAL , +POLARITY_VERTICAL , +POLARITY_VERTICAL_LOW =1 , +POLARITY_VERTICAL_MEDIUM , + POLARITY_VERTICAL_HIGH,
Quoted 8 lines
include/hw/s390x/cpu-topology.h line 22
Quoted 2 lines
include/hw/s390x/cpu-topology.h new file +2
-0+ POLARITY_MAX, +};IMO you should define the polarization and entitlement enums as qapi enums.
The polarization is passed as data in the qapi event when the polarization changes.
And the entitlement is passed to qemu when modifying the topology.
(Thought I already answered but... seems no.)
looks good, thanks, I change this.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 2/11 s390x/cpu topology: add topology entries on CPU hotplug
In reply to Pierre MorelThe topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socket
The possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-) create mode 100644 hw/s390x/cpu-topology.c
include/hw/s390x/cpu-topology.h +24 -0
@@ -10,6 +10,9 @@#ifndef HW_S390X_CPU_TOPOLOGY_H #define HW_S390X_CPU_TOPOLOGY_H +#include "qemu/queue.h" +#include "hw/boards.h" +#define S390_TOPOLOGY_CPU_IFL 0x03 enum s390_topology_polarity {
@@ -21,4 +24,25 @@ enum s390_topology_polarity { POLARITY_MAX, }; +typedef struct S390Topology { +uint8_t *cores_per_socket; +CpuTopology *smp; +}S390Topology ; + +#ifdef CONFIG_KVM +bool s390_has_topology (void ); +void s390_topology_set_cpu (MachineState *ms,S390CPU *cpu,Error **errp); +#else +static inline bool s390_has_topology (void ) +{ +return false; +} +static inline void s390_topology_set_cpu (MachineState *ms, +S390CPU *cpu, +Error **errp) {} +#endif + +extern S390Topology s390_topology; +int s390_socket_nb (S390CPU *cpu); +#endif
hw/s390x/cpu-topology.c new file +256 -0
256 lines not on this page · raw message
hw/s390x/s390-virtio-ccw.c +21 -2
@@ -44,6 +44,7 @@#include "hw/s390x/pv.h" #include "migration/blocker.h" #include "qapi/visitor.h" +#include "hw/s390x/cpu-topology.h" static Error *pv_mig_blocker;
@@ -310,10 +311,18 @@ static void s390_cpu_plug(HotplugHandler *hotplug_dev, {MachineState *ms =MACHINE (hotplug_dev);S390CPU *cpu =S390_CPU (dev); +ERRP_GUARD ();g_assert (!ms->possible_cpus->cpus[cpu->env.core_id].cpu); ms->possible_cpus->cpus[cpu->env.core_id].cpu =OBJECT (dev); +if (s390_has_topology ()) { +s390_topology_set_cpu (ms, cpu, errp); +if (*errp) { +return ; + } + } +if (dev->hotplugged) {raise_irq_cpu_hotplug (); }
@@ -551,11 +560,21 @@ static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) sizeof(CPUArchId) * max_cpus); ms->possible_cpus->len = max_cpus;for (i =0 ; i < ms->possible_cpus->len; i++) { +CpuInstanceProperties *props = &ms->possible_cpus->cpus[i].props; + ms->possible_cpus->cpus[i].type = ms->cpu_type; ms->possible_cpus->cpus[i].vcpus_count =1 ; ms->possible_cpus->cpus[i].arch_id = i; - ms->possible_cpus->cpus[i].props.has_core_id = true; - ms->possible_cpus->cpus[i].props.core_id = i; + + props->has_core_id = true; + props->core_id = i; + props->has_socket_id = true; + props->socket_id = i / ms->smp.cores; + props->has_book_id = true; + props->book_id = i / (ms->smp.cores * ms->smp.sockets); + props->has_drawer_id = true; + props->drawer_id = i / + (ms->smp.cores * ms->smp.sockets * ms->smp.books); }return ms->possible_cpus;
hw/s390x/meson.build +1 -0
@@ -24,6 +24,7 @@ s390x_ss.add(when: 'CONFIG_KVM', if_true: files( 's390-stattrib-kvm.c', 'pv.c', 's390-pci-kvm.c', + 'cpu-topology.c', )) s390x_ss.add(when: 'CONFIG_TCG', if_true: files( 'tod-tcg.c',
Signature · 1 line
2.31.1
Re: [PATCH v15 02/11] s390x/cpu topology: add topology entries on CPU hotplug
Quoted 30 lines · Pierre Morel
The topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socketThe possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-) create mode 100644 hw/s390x/cpu-topology.c
[...]
Quoted 22 lines
hw/s390x/cpu-topology.c new file +15
-0@@ -0,0 +1,256 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* Author(s): Pierre Morel <[email protected]> + +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "hw/qdev-properties.h"
hw/s390x/cpu-topology.c line 25
Quoted 10 lines
hw/s390x/cpu-topology.c new file +10
-0+#include "hw/boards.h" +#include "qemu/typedefs.h" +#include "target/s390x/cpu.h" +#include "hw/s390x/s390-virtio-ccw.h" +#include "hw/s390x/cpu-topology.h" + +/* +* s390_topology is used to keep the topology information. +* .list: queue the topology entries inside which +* we keep the information on the CPU topology.
.list doesn't exist yet.
Quoted 10 lines
hw/s390x/cpu-topology.c new file +9
-0+* .socket: tracks information on the count of cores per socket. +* .smp: keeps track of the machine topology. +* +*/ +S390Topology s390_topology = { +/* will be initialized after the cpu model is realized */ + .cores_per_socket =NULL , + .smp =NULL , +};
[...]
hw/s390x/cpu-topology.c line 83
Quoted 7 lines
hw/s390x/cpu-topology.c new file +7
-0+ +/** +* s390_topology_cpu_default: +* @cpu: pointer to a S390CPU +* @errp: Error pointer +* +* Setup the default topology for unset attributes.
My suggestion:
Setup the default topology if no attributes are already set. Passing a CPU with some, but not all, attributes set is considered an error.
Quoted 3 lines
+ *
+ * The function accept only all all default values or all set values
+ * for the geometry topology.
acceptS, all all
If you take my suggestion, you can just drop this sentence.
Quoted 33 lines
+ *
+ * The function calculates the (drawer_id,book_id,socket_id)
+ * topology by filling the cores starting from the first socket
+ * (0, 0, 0) up to the last (smp->drawers,smp->books,smp->sockets).
+ *
+ * CPU type, polarity and dedication have defaults values set in the
+ *s390x_cpu_properties.
+ */hw/s390x/cpu-topology.c new file +24
-0+static void s390_topology_cpu_default (S390CPU *cpu,Error **errp) +{ +CpuTopology *smp = s390_topology.smp; +CPUS390XState *env = &cpu->env; + +/* All geometry topology attributes must be set or all unset */ +if ((env->socket_id <0 || env->book_id <0 || env->drawer_id <0 ) && + (env->socket_id >=0 || env->book_id >=0 || env->drawer_id >=0 )) { +error_setg (errp, +"Please define all or none of the topology geometry attributes" ); +return ; + } + +/* Check if one of the geometry topology is unset */ +if (env->socket_id <0 ) { +/* Calculate default geometry topology attributes */ + env->socket_id = (env->core_id / smp->cores) % smp->sockets; + env->book_id = (env->core_id / (smp->sockets * smp->cores)) % + smp->books; + env->drawer_id = (env->core_id / + (smp->books * smp->sockets * smp->cores)) % + smp->drawers; + } +}
[...]
Quoted 5 lines
hw/s390x/cpu-topology.c new file +5
-0+ +/** +* s390_set_core_in_socket: +* @cpu: the new S390CPU to insert in the topology structure +* @drawer_id: new drawer_id
hw/s390x/cpu-topology.c line 171
Quoted 10 lines
hw/s390x/cpu-topology.c new file +10
-0+* @book_id: new book_id +* @socket_id: new socket_id +* @creation: if is true the CPU is a new CPU and there is no old socket +* to handle. +* if is false, this is a moving the CPU and old socket count +* must be decremented. +* @errp: the error pointer +* +*/ +static void s390_set_core_in_socket(S390CPU *cpu, int drawer_id, int book_id,
Maybe name it s390_(topology_)?add_core_to_socket instead.
hw/s390x/cpu-topology.c line 179
Quoted 8 lines
hw/s390x/cpu-topology.c new file +8
-0+ int socket_id, bool creation, Error **errp) +{ + int old_socket = s390_socket_nb(cpu); +int new_socket; + +if (creation) { + new_socket = old_socket; + }else {
You need parentheses here.
Quoted 1 line
+
new_socket=drawer_id*s390_topology.smp->books+
(
Quoted 1 line
+
book_id*s390_topology.smp->sockets+
)
Quoted 18 lines
hw/s390x/cpu-topology.c new file +17
-0+ socket_id; + } + +/* Check for space on new socket */ +if ((new_socket != old_socket) && + (s390_topology.cores_per_socket[new_socket] >= + s390_topology.smp->cores)) { +error_setg (errp,"No more space on this socket" ); +return ; + } + +/* Update the count of cores in sockets */ + s390_topology.cores_per_socket[new_socket] +=1 ; +if (!creation) { + s390_topology.cores_per_socket[old_socket] -=1 ; + } +}
[...]
hw/s390x/cpu-topology.c line 227
Quoted 11 lines
hw/s390x/cpu-topology.c new file +11
-0+/** +* s390_topology_set_cpu: +* @ms: MachineState used to initialize the topology structure on +* first call. +* @cpu: the new S390CPU to insert in the topology structure +* @errp: the error pointer +* +* Called from CPU Hotplug to check and setup the CPU attributes +* before to insert the CPU in the topology. +*/ +void s390_topology_set_cpu (MachineState *ms,S390CPU *cpu,Error **errp)
The name is rather non informative.s390_topology_setup_cpu ?
hw/s390x/cpu-topology.c line 233
Quoted 6 lines
hw/s390x/cpu-topology.c new file +6
-0+{ +ERRP_GUARD (); + +/* +* We do not want to initialize the topology if the cpu model +* does not support topology consequently, we have to wait for
^
Still think there should be a comma here.
Quoted 65 lines
hw/s390x/cpu-topology.c new file +23
-0+* the first CPU to be realized, which realizes the CPU model +* to initialize the topology structures. +* +* s390_topology_set_cpu() is called from the cpu hotplug. +*/ +if (!s390_topology.cores_per_socket) { +s390_topology_init (ms); + } + +s390_topology_check (cpu, errp); +if (*errp) { +return ; + } + +/* Set the CPU inside the socket */ +s390_set_core_in_socket (cpu,0 ,0 ,0 , true, errp); +if (*errp) { +return ; + } + +/* topology tree is reflected in props */ +s390_update_cpu_props (ms, cpu); +}hw/s390x/s390-virtio-ccw.c +11
-2@@ -44,6 +44,7 @@#include "hw/s390x/pv.h" #include "migration/blocker.h" #include "qapi/visitor.h" +#include "hw/s390x/cpu-topology.h" static Error *pv_mig_blocker;@@ -310,10 +311,18 @@ static void s390_cpu_plug(HotplugHandler *hotplug_dev, {MachineState *ms =MACHINE (hotplug_dev);S390CPU *cpu =S390_CPU (dev); +ERRP_GUARD ();g_assert (!ms->possible_cpus->cpus[cpu->env.core_id].cpu); ms->possible_cpus->cpus[cpu->env.core_id].cpu =OBJECT (dev); +if (s390_has_topology ()) { +s390_topology_set_cpu (ms, cpu, errp); +if (*errp) { +return ; + } + } +if (dev->hotplugged) {raise_irq_cpu_hotplug (); }@@ -551,11 +560,21 @@ static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) sizeof(CPUArchId) * max_cpus); ms->possible_cpus->len = max_cpus;for (i =0 ; i < ms->possible_cpus->len; i++) { +CpuInstanceProperties *props = &ms->possible_cpus->cpus[i].props; + ms->possible_cpus->cpus[i].type = ms->cpu_type; ms->possible_cpus->cpus[i].vcpus_count =1 ; ms->possible_cpus->cpus[i].arch_id = i; - ms->possible_cpus->cpus[i].props.has_core_id = true; - ms->possible_cpus->cpus[i].props.core_id = i;
hw/s390x/s390-virtio-ccw.c static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) · line 577
Quoted 10 lines
hw/s390x/s390-virtio-ccw.c +10
-0+ + props->has_core_id = true; + props->core_id = i; + props->has_socket_id = true; + props->socket_id = i / ms->smp.cores; + props->has_book_id = true; + props->book_id = i / (ms->smp.cores * ms->smp.sockets); + props->has_drawer_id = true; + props->drawer_id = i / + (ms->smp.cores * ms->smp.sockets * ms->smp.books);
You need to calculate the modulus like in s390_topology_cpu_default, right?
Quoted 15 lines
hw/s390x/s390-virtio-ccw.c +0
-0}return ms->possible_cpus;hw/s390x/meson.build +1
-0@@ -24,6 +24,7 @@ s390x_ss.add(when: 'CONFIG_KVM', if_true: files( 's390-stattrib-kvm.c', 'pv.c', 's390-pci-kvm.c', + 'cpu-topology.c', )) s390x_ss.add(when: 'CONFIG_TCG', if_true: files( 'tod-tcg.c',
Re: [PATCH v15 02/11] s390x/cpu topology: add topology entries on CPU hotplug
Quoted 58 lines · Nina Schoetterl-Glausch
Quoted 30 lines · Pierre Morel
The topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socketThe possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-)create mode 100644
hw/s390x/cpu-topology.c[...]
Quoted 26 lines
hw/s390x/cpu-topology.c new file +19
-0@@ -0,0 +1,256 @@ +/* +* CPU Topology +* +* Copyright IBM Corp. 2022 +* Author(s): Pierre Morel <[email protected]> + +* This work is licensed under the terms of the GNU GPL, version 2 or (at +* your option) any later version. See the COPYING file in the top-level +* directory. +*/ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "hw/qdev-properties.h" +#include "hw/boards.h" +#include "qemu/typedefs.h" +#include "target/s390x/cpu.h" +#include "hw/s390x/s390-virtio-ccw.h"
Quoted 8 lines
hw/s390x/cpu-topology.c line 25
Quoted 6 lines
hw/s390x/cpu-topology.c new file +6
-0+#include "hw/s390x/cpu-topology.h" + +/* +* s390_topology is used to keep the topology information. +* .list: queue the topology entries inside which +* we keep the information on the CPU topology. .list doesn't exist yet.
right, I move it to the corresponding patch.
Quoted 12 lines
Quoted 11 lines
hw/s390x/cpu-topology.c new file +9
-0+* .socket: tracks information on the count of cores per socket. +* .smp: keeps track of the machine topology. +* +*/ +S390Topology s390_topology = { +/* will be initialized after the cpu model is realized */ + .cores_per_socket =NULL , + .smp =NULL , +};[...]
Quoted 11 lines
hw/s390x/cpu-topology.c line 83
Quoted 7 lines
hw/s390x/cpu-topology.c new file +7
-0+ +/** +* s390_topology_cpu_default: +* @cpu: pointer to a S390CPU +* @errp: Error pointer +* +* Setup the default topology for unset attributes. My suggestion:
Setup the default topology if no attributes are already set. Passing a CPU with some, but not all, attributes set is considered an error.
Thanks, I take your suggestion.
Quoted 51 lines
Quoted 4 lines
+ *
+ * The function accept only all all default values or all set values
+ * for the geometry topology.acceptS, all all
If you take my suggestion, you can just drop this sentence.Quoted 33 lines
+ *
+ * The function calculates the (drawer_id,book_id,socket_id)
+ * topology by filling the cores starting from the first socket
+ * (0, 0, 0) up to the last (smp->drawers,smp->books,smp->sockets).
+ *
+ * CPU type, polarity and dedication have defaults values set in the
+ *s390x_cpu_properties.
+ */hw/s390x/cpu-topology.c new file +24
-0+static void s390_topology_cpu_default (S390CPU *cpu,Error **errp) +{ +CpuTopology *smp = s390_topology.smp; +CPUS390XState *env = &cpu->env; + +/* All geometry topology attributes must be set or all unset */ +if ((env->socket_id <0 || env->book_id <0 || env->drawer_id <0 ) && + (env->socket_id >=0 || env->book_id >=0 || env->drawer_id >=0 )) { +error_setg (errp, +"Please define all or none of the topology geometry attributes" ); +return ; + } + +/* Check if one of the geometry topology is unset */ +if (env->socket_id <0 ) { +/* Calculate default geometry topology attributes */ + env->socket_id = (env->core_id / smp->cores) % smp->sockets; + env->book_id = (env->core_id / (smp->sockets * smp->cores)) % + smp->books; + env->drawer_id = (env->core_id / + (smp->books * smp->sockets * smp->cores)) % + smp->drawers; + } +}[...]
Quoted 9 lines
hw/s390x/cpu-topology.c new file +9
-0+ +/** +* s390_set_core_in_socket: +* @cpu: the new S390CPU to insert in the topology structure +* @drawer_id: new drawer_id +* @book_id: new book_id +* @socket_id: new socket_id +* @creation: if is true the CPU is a new CPU and there is no old socket +* to handle.
Quoted 8 lines
Quoted 6 lines
+ * if is false, this is a moving the CPU and old socket count
+ * must be decremented.
+ * @errp: the error pointer
+ *
+ */
+static void s390_set_core_in_socket(S390CPU *cpu, intdrawer_id, intbook_id,Maybe name it s390_(topology_)?add_core_to_socket instead.
OK, it is better
Quoted 12 lines
hw/s390x/cpu-topology.c line 179
Quoted 9 lines
hw/s390x/cpu-topology.c new file +8
-0+ int socket_id, bool creation, Error **errp) +{ + int old_socket = s390_socket_nb(cpu); +int new_socket; + +if (creation) { + new_socket = old_socket; + }else {You need parentheses here.
Quoted 5 lines
Quoted 1 line
+
new_socket=drawer_id*s390_topology.smp->books+(Quoted 1 line
+
book_id*s390_topology.smp->sockets+)Quoted 1 line
+
socket_id;
If you prefer I can us parentheses.
Quoted 19 lines
Quoted 17 lines
hw/s390x/cpu-topology.c new file +16
-0+ } + +/* Check for space on new socket */ +if ((new_socket != old_socket) && + (s390_topology.cores_per_socket[new_socket] >= + s390_topology.smp->cores)) { +error_setg (errp,"No more space on this socket" ); +return ; + } + +/* Update the count of cores in sockets */ + s390_topology.cores_per_socket[new_socket] +=1 ; +if (!creation) { + s390_topology.cores_per_socket[old_socket] -=1 ; + } +}[...]
Quoted 14 lines
hw/s390x/cpu-topology.c line 227
Quoted 11 lines
hw/s390x/cpu-topology.c new file +11
-0+/** +* s390_topology_set_cpu: +* @ms: MachineState used to initialize the topology structure on +* first call. +* @cpu: the new S390CPU to insert in the topology structure +* @errp: the error pointer +* +* Called from CPU Hotplug to check and setup the CPU attributes +* before to insert the CPU in the topology. +*/ +void s390_topology_set_cpu (MachineState *ms,S390CPU *cpu,Error **errp)The name is rather non informative.
s390_topology_setup_cpu?
OK
Quoted 9 lines
hw/s390x/cpu-topology.c line 233
Quoted 7 lines
hw/s390x/cpu-topology.c new file +6
-0+{ +ERRP_GUARD (); + +/* +* We do not want to initialize the topology if the cpu model +* does not support topology consequently, we have to wait for ^Still think there should be a comma here.
OK
Quoted 70 lines
Quoted 70 lines
hw/s390x/cpu-topology.c new file +23
-0+* the first CPU to be realized, which realizes the CPU model +* to initialize the topology structures. +* +* s390_topology_set_cpu() is called from the cpu hotplug. +*/ +if (!s390_topology.cores_per_socket) { +s390_topology_init (ms); + } + +s390_topology_check (cpu, errp); +if (*errp) { +return ; + } + +/* Set the CPU inside the socket */ +s390_set_core_in_socket (cpu,0 ,0 ,0 , true, errp); +if (*errp) { +return ; + } + +/* topology tree is reflected in props */ +s390_update_cpu_props (ms, cpu); +}hw/s390x/s390-virtio-ccw.c +15
-2@@ -44,6 +44,7 @@#include "hw/s390x/pv.h" #include "migration/blocker.h" #include "qapi/visitor.h" +#include "hw/s390x/cpu-topology.h" static Error *pv_mig_blocker;@@ -310,10 +311,18 @@ static void s390_cpu_plug(HotplugHandler *hotplug_dev, {MachineState *ms =MACHINE (hotplug_dev);S390CPU *cpu =S390_CPU (dev); +ERRP_GUARD ();g_assert (!ms->possible_cpus->cpus[cpu->env.core_id].cpu); ms->possible_cpus->cpus[cpu->env.core_id].cpu =OBJECT (dev); +if (s390_has_topology ()) { +s390_topology_set_cpu (ms, cpu, errp); +if (*errp) { +return ; + } + } +if (dev->hotplugged) {raise_irq_cpu_hotplug (); }@@ -551,11 +560,21 @@ static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) sizeof(CPUArchId) * max_cpus); ms->possible_cpus->len = max_cpus;for (i =0 ; i < ms->possible_cpus->len; i++) { +CpuInstanceProperties *props = &ms->possible_cpus->cpus[i].props; + ms->possible_cpus->cpus[i].type = ms->cpu_type; ms->possible_cpus->cpus[i].vcpus_count =1 ; ms->possible_cpus->cpus[i].arch_id = i; - ms->possible_cpus->cpus[i].props.has_core_id = true; - ms->possible_cpus->cpus[i].props.core_id = i; + + props->has_core_id = true; + props->core_id = i; + props->has_socket_id = true;
Quoted 8 lines
hw/s390x/s390-virtio-ccw.c static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) · line 577
Quoted 6 lines
hw/s390x/s390-virtio-ccw.c +6
-0+ props->socket_id = i / ms->smp.cores; + props->has_book_id = true; + props->book_id = i / (ms->smp.cores * ms->smp.sockets); + props->has_drawer_id = true; + props->drawer_id = i / + (ms->smp.cores * ms->smp.sockets * ms->smp.books);You need to calculate the modulus like in
s390_topology_cpu_default, right?
!!! yes of course, good catch, I forgot that.
Since there are two uses of this calculation, what about using inlines? like:
static inline int s390_std_socket(int n, CpuTopology *smp)
{
return (n / smp->cores) % smp->sockets;
}static inline int s390_std_book(int n, CpuTopology *smp)
{
return (n / (smp->cores * smp->sockets)) % smp->books;
}static inline int s390_std_drawer(int n, CpuTopology *smp)
{
return (n / (smp->cores * smp->sockets * smp->books)) % smp->books;
}Thanks for the comments.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 02/11] s390x/cpu topology: add topology entries on CPU hotplug
Quoted 33 lines · Pierre Morel
Quoted 31 lines · Nina Schoetterl-Glausch
Quoted 30 lines · Pierre Morel
The topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socketThe possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-)create mode 100644
hw/s390x/cpu-topology.c
[...]
Quoted 32 lines
Quoted 17 lines
Quoted 5 lines
hw/s390x/cpu-topology.c new file +4
-0+/** +* s390_set_core_in_socket: +* @cpu: the new S390CPU to insert in the topology structure +* @drawer_id: new drawer_id hw/s390x/cpu-topology.c line 171
Quoted 10 lines
hw/s390x/cpu-topology.c new file +10
-0+* @book_id: new book_id +* @socket_id: new socket_id +* @creation: if is true the CPU is a new CPU and there is no old socket +* to handle. +* if is false, this is a moving the CPU and old socket count +* must be decremented. +* @errp: the error pointer +* +*/ +static void s390_set_core_in_socket(S390CPU *cpu, int drawer_id, int book_id,Maybe name it s390_(topology_)?add_core_to_socket instead.
OK, it is better
Quoted 12 lines
hw/s390x/cpu-topology.c line 179
Quoted 9 lines
hw/s390x/cpu-topology.c new file +8
-0+ int socket_id, bool creation, Error **errp) +{ + int old_socket = s390_socket_nb(cpu); +int new_socket; + +if (creation) { + new_socket = old_socket; + }else {You need parentheses here.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
+
new_socket=drawer_id*s390_topology.smp->books+(Quoted 1 line
+
book_id*s390_topology.smp->sockets+)Quoted 1 line
+
socket_id;If you prefer I can us parentheses.
It's necessary, otherwise the multiplication of book_id and smp->sockets takes precedence.
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
hw/s390x/cpu-topology.c new file +16
-0+ } + +/* Check for space on new socket */ +if ((new_socket != old_socket) && + (s390_topology.cores_per_socket[new_socket] >= + s390_topology.smp->cores)) { +error_setg (errp,"No more space on this socket" ); +return ; + } + +/* Update the count of cores in sockets */ + s390_topology.cores_per_socket[new_socket] +=1 ; +if (!creation) { + s390_topology.cores_per_socket[old_socket] -=1 ; + } +}
[...]
Quoted 56 lines
Quoted 47 lines
Quoted 47 lines
hw/s390x/s390-virtio-ccw.c +15
-2@@ -44,6 +44,7 @@#include "hw/s390x/pv.h" #include "migration/blocker.h" #include "qapi/visitor.h" +#include "hw/s390x/cpu-topology.h" static Error *pv_mig_blocker;@@ -310,10 +311,18 @@ static void s390_cpu_plug(HotplugHandler *hotplug_dev, {MachineState *ms =MACHINE (hotplug_dev);S390CPU *cpu =S390_CPU (dev); +ERRP_GUARD ();g_assert (!ms->possible_cpus->cpus[cpu->env.core_id].cpu); ms->possible_cpus->cpus[cpu->env.core_id].cpu =OBJECT (dev); +if (s390_has_topology ()) { +s390_topology_set_cpu (ms, cpu, errp); +if (*errp) { +return ; + } + } +if (dev->hotplugged) {raise_irq_cpu_hotplug (); }@@ -551,11 +560,21 @@ static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) sizeof(CPUArchId) * max_cpus); ms->possible_cpus->len = max_cpus;for (i =0 ; i < ms->possible_cpus->len; i++) { +CpuInstanceProperties *props = &ms->possible_cpus->cpus[i].props; + ms->possible_cpus->cpus[i].type = ms->cpu_type; ms->possible_cpus->cpus[i].vcpus_count =1 ; ms->possible_cpus->cpus[i].arch_id = i; - ms->possible_cpus->cpus[i].props.has_core_id = true; - ms->possible_cpus->cpus[i].props.core_id = i; + + props->has_core_id = true; + props->core_id = i; + props->has_socket_id = true;Quoted 8 lines
hw/s390x/s390-virtio-ccw.c static const CPUArchIdList *s390_possible_cpu_arch_ids(MachineState *ms) · line 577
Quoted 6 lines
hw/s390x/s390-virtio-ccw.c +6
-0+ props->socket_id = i / ms->smp.cores; + props->has_book_id = true; + props->book_id = i / (ms->smp.cores * ms->smp.sockets); + props->has_drawer_id = true; + props->drawer_id = i / + (ms->smp.cores * ms->smp.sockets * ms->smp.books);You need to calculate the modulus like in
s390_topology_cpu_default, right?
Quoted 4 lines
!!! yes of course, good catch, I forgot that.
Since there are two uses of this calculation, what about using inlines? like:
Fine by me, I just have no idea what std is short for.
Since this would be used in s390_topology_cpu_default and calculates the default assignment, I would call it s390_topology_cpu_default_socket, book, etc. or similar
Quoted 22 lines
static inline int
s390_std_socket(int n, CpuTopology *smp){ return (n / smp->cores) % smp->sockets; }static inline int
s390_std_book(int n, CpuTopology *smp){ return (n / (smp->cores * smp->sockets)) % smp->books; }static inline int
s390_std_drawer(int n, CpuTopology *smp){ return (n / (smp->cores * smp->sockets * smp->books)) % smp->books; }Thanks for the comments.
Regards,
Pierre
Re: [PATCH v15 02/11] s390x/cpu topology: add topology entries on CPU hotplug
Quoted 68 lines · Nina Schoetterl-Glausch
Quoted 33 lines · Pierre Morel
Quoted 31 lines · Nina Schoetterl-Glausch
Quoted 30 lines · Pierre Morel
The topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socketThe possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-)create mode 100644
hw/s390x/cpu-topology.c[...]
Quoted 33 lines
Quoted 17 lines
Quoted 5 lines
hw/s390x/cpu-topology.c new file +4
-0+/** +* s390_set_core_in_socket: +* @cpu: the new S390CPU to insert in the topology structure +* @drawer_id: new drawer_id hw/s390x/cpu-topology.c line 171
Quoted 10 lines
hw/s390x/cpu-topology.c new file +10
-0+* @book_id: new book_id +* @socket_id: new socket_id +* @creation: if is true the CPU is a new CPU and there is no old socket +* to handle. +* if is false, this is a moving the CPU and old socket count +* must be decremented. +* @errp: the error pointer +* +*/ +static void s390_set_core_in_socket(S390CPU *cpu, int drawer_id, int book_id,Maybe name it s390_(topology_)?add_core_to_socket instead.
OK, it is better
Quoted 13 lines
hw/s390x/cpu-topology.c line 179
Quoted 9 lines
hw/s390x/cpu-topology.c new file +8
-0+ int socket_id, bool creation, Error **errp) +{ + int old_socket = s390_socket_nb(cpu); +int new_socket; + +if (creation) { + new_socket = old_socket; + }else {You need parentheses here.
Quoted 1 line
+
new_socket=drawer_id*s390_topology.smp->books+
Quoted 10 lines
Quoted 6 lines
Quoted 4 lines
(Quoted 1 line
+
book_id*s390_topology.smp->sockets+)Quoted 1 line
+
socket_id;If you prefer I can us parentheses.
It's necessary, otherwise the multiplication of
book_idandsmp->socketstakes precedence.Quoted 2 lines
Right, I did not understand where you want the parenthesis. I think you mean:
new_socket = (drawer_id * s390_topology.smp->books + book_id) *
s390_topology.smp->sockets + socket_id;thanks,
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 02/11] s390x/cpu topology: add topology entries on CPU hotplug
Quoted 81 lines · Pierre Morel
Quoted 68 lines · Nina Schoetterl-Glausch
Quoted 33 lines · Pierre Morel
Quoted 31 lines · Nina Schoetterl-Glausch
Quoted 30 lines · Pierre Morel
The topology information are attributes of the CPU and are specified during the CPU device creation.
On hot plug we:
- calculate the default values for the topology for drawers, books and sockets in the case they are not specified.
- verify the CPU attributes
- check that we have still room on the desired socketThe possibility to insert a CPU in a mask is dependent on the number of cores allowed in a socket, a book or a drawer, the checking is done during the hot plug of the CPU to have an immediate answer.
If the complete topology is not specified, the core is added in the physical topology based on its core ID and it gets defaults values for the modifier attributes.
This way, starting QEMU without specifying the topology can still get some advantage of the CPU topology.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 24 +++ hw/s390x/cpu-topology.c | 256 ++++++++++++++++++++++++++++++++ hw/s390x/s390-virtio-ccw.c | 23 ++- hw/s390x/meson.build | 1 + 4 files changed, 302 insertions(+), 2 deletions(-)create mode 100644
hw/s390x/cpu-topology.c[...]
Quoted 33 lines
Quoted 17 lines
Quoted 5 lines
hw/s390x/cpu-topology.c new file +4
-0+/** +* s390_set_core_in_socket: +* @cpu: the new S390CPU to insert in the topology structure +* @drawer_id: new drawer_id hw/s390x/cpu-topology.c line 171
Quoted 10 lines
hw/s390x/cpu-topology.c new file +10
-0+* @book_id: new book_id +* @socket_id: new socket_id +* @creation: if is true the CPU is a new CPU and there is no old socket +* to handle. +* if is false, this is a moving the CPU and old socket count +* must be decremented. +* @errp: the error pointer +* +*/ +static void s390_set_core_in_socket(S390CPU *cpu, int drawer_id, int book_id,Maybe name it s390_(topology_)?add_core_to_socket instead.
OK, it is better
Quoted 13 lines
hw/s390x/cpu-topology.c line 179
Quoted 9 lines
hw/s390x/cpu-topology.c new file +8
-0+ int socket_id, bool creation, Error **errp) +{ + int old_socket = s390_socket_nb(cpu); +int new_socket; + +if (creation) { + new_socket = old_socket; + }else {You need parentheses here.
Quoted 1 line
+
new_socket=drawer_id*s390_topology.smp->books+Quoted 10 lines
Quoted 6 lines
Quoted 4 lines
(Quoted 1 line
+
book_id*s390_topology.smp->sockets+)Quoted 1 line
+
socket_id;If you prefer I can us parentheses.
It's necessary, otherwise the multiplication of
book_idandsmp->socketstakes precedence.Quoted 2 lines
Quoted 5 lines
Right, I did not understand where you want the parenthesis. I think you mean:
new_socket = (drawer_id * s390_topology.smp->books + book_id) * s390_topology.smp->sockets + socket_id;
Yes, exactly. I see how it was ambiguous.
Quoted 6 lines
thanks,
Regards,
Pierre
PATCH v15 3/11 target/s390x/cpu topology: handle STSI(15) and build the SYSIB
In reply to Pierre MorelOn interception of STSI(15.1.x) the System Information Block (SYSIB) is built from the list of pre-ordered topology entries.
--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-) create mode 100644 target/s390x/kvm/cpu_topology.c
include/hw/s390x/cpu-topology.h +22 -0
@@ -24,9 +24,31 @@ enum s390_topology_polarity { POLARITY_MAX, }; +typedef union s390_topology_id { +uint64_t id; +struct { +uint8_t level5; +uint8_t drawer; +uint8_t book; +uint8_t socket; +uint8_t dedicated; +uint8_t polarity; +uint8_t type; +uint8_t origin; + }; +}s390_topology_id ; + +typedef struct S390TopologyEntry { +QTAILQ_ENTRY (S390TopologyEntry) next; +s390_topology_id id; +uint64_t mask; +}S390TopologyEntry ; +typedef struct S390Topology {uint8_t *cores_per_socket; + QTAILQ_HEAD(, S390TopologyEntry) list;CpuTopology *smp; +uint8_t polarity; }S390Topology ;#ifdef CONFIG_KVM
include/hw/s390x/sclp.h +1 -0
@@ -112,6 +112,7 @@ typedef struct CPUEntry { }QEMU_PACKED CPUEntry;#define SCLP_READ_SCP_INFO_FIXED_CPU_OFFSET 128 +#define SCLP_READ_SCP_INFO_MNEST 2 typedef struct ReadInfo { SCCBHeader h; uint16_t rnmax;
target/s390x/cpu.h +72 -0
87 lines not on this page · raw message
hw/s390x/cpu-topology.c +10 -0
@@ -31,6 +31,8 @@ S390Topology s390_topology = {/* will be initialized after the cpu model is realized */ .cores_per_socket = NULL, .smp = NULL, + .list = QTAILQ_HEAD_INITIALIZER(s390_topology.list), + .polarity = POLARITY_HORIZONTAL, };/**
@@ -65,14 +67,22 @@ bool s390_has_topology(void)* Allocate an array to keep the count of cores per socket. * The index of the array starts at socket 0 from book 0 and * drawer 0 up to the maximum allowed by the machine topology. +* +* Insert a sentinel entry using unused level5 with its maximum value. +* This entry will never be free. */ static void s390_topology_init (MachineState *ms) {CpuTopology *smp = &ms->smp; +S390TopologyEntry *entry; s390_topology.smp = smp; s390_topology.cores_per_socket =g_new0 (uint8_t, smp->sockets * smp->books * smp->drawers); + + entry =g_malloc0 (sizeof (S390TopologyEntry)); + entry->id.level5 =0xff ; +QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }/**
target/s390x/kvm/cpu_topology.c new file +335 -0
335 lines not on this page · raw message
target/s390x/kvm/kvm.c +4 -1
@@ -1919,9 +1919,12 @@ static int handle_stsi(S390CPU *cpu)if (run->s390_stsi.sel1 !=2 || run->s390_stsi.sel2 !=2 ) {return 0 ; } -/* Only sysib 3.2.2 needs post-handling for now. */ insert_stsi_3_2_2 (cpu, run->s390_stsi.addr, run->s390_stsi.ar);return 0 ; +case 15 : +insert_stsi_15_1_x (cpu, run->s390_stsi.sel2, run->s390_stsi.addr, + run->s390_stsi.ar); +return 0 ;default :return 0 ; }
target/s390x/kvm/meson.build +2 -1
@@ -1,6 +1,7 @@ s390x_ss.add(when: 'CONFIG_KVM', if_true: files( - 'kvm.c' + 'kvm.c', + 'cpu_topology.c' ), if_false: files( 'stubs.c' ))
Signature · 1 line
2.31.1
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
Quoted 15 lines · Pierre Morel
Quoted 15 lines
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 <[email protected]>--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-) create mode 100644 target/s390x/kvm/cpu_topology.c
[...]
Quoted 5 lines
Quoted 5 lines
target/s390x/cpu.h +0
-0
[...]
Quoted 8 lines
target/s390x/cpu.h typedef union SysIB { · line 636
Quoted 8 lines
target/s390x/cpu.h +8
-0+ +/* CPU type Topology List Entry */ +typedef struct SysIBTl_cpu { + uint8_t nl; + uint8_t reserved0[3]; +#define SYSIB_TLE_POLARITY_MASK 0x03 +#define SYSIB_TLE_DEDICATED 0x04 + uint8_t entitlement;
I would just call this flags, since it's multiple fields.
Quoted 8 lines
Quoted 8 lines
target/s390x/cpu.h +6
-0+uint8_t type; +uint16_t origin; +uint64_t mask; +} QEMU_PACKED QEMU_ALIGNED(8) SysIBTl_cpu; +QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 ); +
[...]
Quoted 7 lines
Quoted 7 lines
/**
target/s390x/kvm/cpu_topology.c new file +0
-0
[...]
Quoted 9 lines
target/s390x/kvm/cpu_topology.c line 66
Quoted 9 lines
target/s390x/kvm/cpu_topology.c new file +9
-0+ +/* +* Macro to check that the size of data after increment +* will not get bigger than the size of the SysIB. +*/ +#define SYSIB_GUARD (data, x) do { \ + data += x; \ + if (data > sizeof(SysIB)) { \ + return -ENOSPC; \
I would go with ENOMEM here.
Quoted 9 lines
target/s390x/kvm/cpu_topology.c line 75
Quoted 9 lines
target/s390x/kvm/cpu_topology.c new file +9
-0+ } \ + }while (0 ) + +/** +* stsi_set_tle: +* @p: A pointer to the position of the first TLE +* @level: The nested level wanted by the guest +* +* Loop inside the s390_topology.list until the sentinelle entry
s/sentinelle/sentinel/
Quoted 5 lines
Quoted 5 lines
+ * is found and for each entry:
+ * - Check usingSYSIB_GUARD()that the size of the SysIB is not
+ * reached.
+ * - Add all the container TLE needed for the level
+ * - Add the CPU TLE.
I'd focus more on *what* the function does instead of *how*.
Fill the SYSIB with the topology information as described in the PoP, nesting containers as appropriate, with the maximum nesting limited by @level.
Or something similar.
Quoted 3 lines
Quoted 3 lines
+ *
+ * Return value:
+ *s390_top_set_levelreturns the size of theSysIB_15xafter being
You forgot to rename the function here, right?
How about stsi_fill_topology_sysib or stsi_topology_fill_sysib, instead?
Quoted 2 lines
Quoted 2 lines
+ * filled with TLE on success.
+ * It returns-ENOSPCin the case we would overrun the end of the SysIB.
You would have to change to ENOMEM here than also.
Quoted 8 lines
Quoted 8 lines
+ */
target/s390x/kvm/cpu_topology.c new file +7
-0+static int stsi_set_tle (char *p,int level) +{ +S390TopologyEntry *entry; +int last_drawer =-1 ; +int last_book =-1 ; +int last_socket =-1 ; +int drawer_id =0 ;
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 101
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+int book_id =0 ; +int socket_id =0 ; +int n =sizeof (SysIB_151x); + + QTAILQ_FOREACH(entry, &s390_topology.list, next) { +int current_drawer = entry->id.drawer; +int current_book = entry->id.book; +int current_socket = entry->id.socket;
This only saves two characters, so you could just use entry->id. ...
Quoted 3 lines
target/s390x/kvm/cpu_topology.c line 104
Quoted 3 lines
target/s390x/kvm/cpu_topology.c new file +3
-0+bool drawer_change = last_drawer != current_drawer; +bool book_change = drawer_change || last_book != current_book; +bool socket_change = book_change || last_socket != current_socket;
... but keep it if it would make this line too long.
You could also rename entry, to current or cur, if you want to emphasize that.
Quoted 30 lines
Quoted 30 lines
target/s390x/kvm/cpu_topology.c new file +30
-0+ +/* If we reach the guard get out */ +if (entry->id.level5) { +break ; + } + +if (level >3 && drawer_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,3 , drawer_id++); + book_id =0 ; + } +if (level >2 && book_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,2 , book_id++); + socket_id =0 ; + } +if (socket_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,1 , socket_id++); + } + +SYSIB_GUARD (n,sizeof (SysIBTl_cpu)); + p =fill_tle_cpu (p, entry); + last_drawer = entry->id.drawer; + last_book = entry->id.book; + last_socket = entry->id.socket; + } + +return n; +}
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 142
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+ +/** +* setup_stsi: +* sysib: pointer to a SysIB to be filled with SysIB_151x data +* level: Nested level specified by the guest +* +* Setup the SysIB_151x header before calling stsi_set_tle with +* a pointer to the first TLE entry.
Same thing here with regards to describing the what.
Setup the SYSIB for STSI 15.1, the header as well as the description of the topology.
Quoted 45 lines
Quoted 45 lines
+ */
target/s390x/kvm/cpu_topology.c new file +44
-0+static int setup_stsi (SysIB_151x *sysib,int level) +{ + sysib->mnest = level; +switch (level) { +case 4 : + sysib->mag[S390_TOPOLOGY_MAG4 ] = current_machine->smp.drawers; + sysib->mag[S390_TOPOLOGY_MAG3 ] = current_machine->smp.books; + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; +case 3 : + sysib->mag[S390_TOPOLOGY_MAG3 ] = current_machine->smp.drawers * + current_machine->smp.books; + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; +case 2 : + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.drawers * + current_machine->smp.books * + current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; + } + +return stsi_set_tle (sysib->tle, level); +} + +/** +* s390_topology_add_cpu_to_entry: +* @entry: Topology entry to setup +* @cpu: the S390CPU to add +* +* Set the core bit inside the topology mask and +* increments the number of cores for the socket. +*/ +static void s390_topology_add_cpu_to_entry (S390TopologyEntry *entry, +S390CPU *cpu) +{ +set_bit (63 - (cpu->env.core_id %64 ), &entry->mask); +} + +/** +* s390_topology_new_entry: +* @id: s390_topology_id to add
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 195
Quoted 8 lines
+ * @cpu: the S390CPU to add
+ *
+ * Allocate a new entry and initialize it.
+ *
+ * returns the newly allocated entry.
+ */target/s390x/kvm/cpu_topology.c new file +2
-0+static S390TopologyEntry *s390_topology_new_entry (s390_topology_id id, +S390CPU *cpu)
This is used only once, right?
I think I'd go ahead and inline it into s390_topology_insert, since I had to go back and check if new_entry calls add_cpu when reading s390_topology_insert.
Quoted 67 lines
Quoted 67 lines
target/s390x/kvm/cpu_topology.c new file +67
-0+{ +S390TopologyEntry *entry; + + entry =g_malloc0 (sizeof (S390TopologyEntry)); + entry->id.id = id.id; +s390_topology_add_cpu_to_entry (entry, cpu); + +return entry; +} + +/** +* 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.polarity ==POLARITY_VERTICAL ) { +/* +* Vertical polarity with dedicated CPU implies +* vertical high entitlement. +*/ +if (topology_id.dedicated) { + topology_id.polarity |=POLARITY_VERTICAL_HIGH ; + }else { + topology_id.polarity |= cpu->env.entitlement; + } + } + +return topology_id; +} + +/** +* s390_topology_insert: +* @cpu: s390CPU insert. +* +* Parse the topology list to find if the entry already +* exist and add the core in it. +* If it does not exist, allocate a new entry and insert +* it in the queue from lower id to greater id. +*/ +static void s390_topology_insert (S390CPU *cpu) +{ +s390_topology_id id =s390_topology_from_cpu (cpu); +S390TopologyEntry *entry =NULL ; +S390TopologyEntry *tmp =NULL ; + + QTAILQ_FOREACH(tmp, &s390_topology.list, next) { +if (id.id == tmp->id.id) { +s390_topology_add_cpu_to_entry (tmp, cpu); +return ; + }else if (id.id < tmp->id.id) { + entry =s390_topology_new_entry (id, cpu); +QTAILQ_INSERT_BEFORE (tmp, entry, next); +return ; + } + }
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 270
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+} + +/** +* s390_order_tle: +* +* Loop over all CPU and insert it at the right place +* inside the TLE entry list. +*/
Suggestion:
s390_topology_fill_list_sorted
Fill the S390Topology list with entries according to the order specified by the PoP.
Quoted 6 lines
Quoted 6 lines
target/s390x/kvm/cpu_topology.c new file +6
-0+static void s390_order_tle (void ) +{ +CPUState *cs; + +CPU_FOREACH (cs) { +s390_topology_insert (S390_CPU (cs));
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 284
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+ } +} + +/** +* s390_free_tle: +* +* Loop over all TLE entries and free them. +* Keep the sentinelle which is the only one with level5 != 0
s/sentinelle/sentinel/
Quoted 1 line
Quoted 1 line
+ */
Suggestion:s390_topology_empty_list
Clear all entries in the S390Topology list except the sentinel.
Quoted 16 lines
Quoted 16 lines
target/s390x/kvm/cpu_topology.c new file +16
-0+static void s390_free_tle (void ) +{ +S390TopologyEntry *entry =NULL ; +S390TopologyEntry *tmp =NULL ; + + QTAILQ_FOREACH_SAFE(entry, &s390_topology.list, next, tmp) { +if (!entry->id.level5) { +QTAILQ_REMOVE (&s390_topology.list, entry, next); +g_free (entry); + } + } +} + +/** +* insert_stsi_15_1_x: +* cpu: the CPU doing the call for which we set CC
Quoted 8 lines
Quoted 8 lines
+ * sel2: the selector 2, containing the nested level
+ * addr: Guest logical address of the guest SysIB
+ * ar: the access register number
+ *
+ * Reserve a zeroed SysIB, letsetup_stsito fill it and
+ * copy the SysIB to the guest memory.
+ *
+ * In case of overflow set CC(3) and no copy is done.
Suggestion:
Emulate STSI 15.1.x, that is, perform all necessary checks and fill the SYSIB. In case the topology description is too long to fit into the SYSIB, set CC=3 and abort without writing the SYSIB.
Quoted 8 lines
Quoted 8 lines
+ */
target/s390x/kvm/cpu_topology.c new file +7
-0+void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar) +{ +SysIB sysib = {0 }; +int len; + +if (!s390_has_topology () || sel2 <2 || sel2 >SCLP_READ_SCP_INFO_MNEST ) { +setcc (cpu,3 );
Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 325
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+return ; + } + +s390_order_tle (); + + len =setup_stsi (&sysib.sysib_151x, sel2); + +if (len <0 ) {
I stumbled a bit over this, maybe rename len to r.
Quoted 41 lines
Quoted 41 lines
target/s390x/kvm/cpu_topology.c new file +10
-0+setcc (cpu,3 ); +return ; + } + + sysib.sysib_151x.length =cpu_to_be16 (len); +s390_cpu_virt_mem_write (cpu, addr, ar, &sysib, len); +setcc (cpu,0 ); + +s390_free_tle (); +}target/s390x/kvm/kvm.c +4
-1@@ -1919,9 +1919,12 @@ static int handle_stsi(S390CPU *cpu)if (run->s390_stsi.sel1 !=2 || run->s390_stsi.sel2 !=2 ) {return 0 ; } -/* Only sysib 3.2.2 needs post-handling for now. */ insert_stsi_3_2_2 (cpu, run->s390_stsi.addr, run->s390_stsi.ar);return 0 ; +case 15 : +insert_stsi_15_1_x (cpu, run->s390_stsi.sel2, run->s390_stsi.addr, + run->s390_stsi.ar); +return 0 ;default :return 0 ; }target/s390x/kvm/meson.build +2
-1@@ -1,6 +1,7 @@ s390x_ss.add(when: 'CONFIG_KVM', if_true: files( - 'kvm.c' + 'kvm.c', + 'cpu_topology.c' ), if_false: files( 'stubs.c' ))
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
Quoted 23 lines · Nina Schoetterl-Glausch
Quoted 15 lines · Pierre Morel
Quoted 15 lines
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 <[email protected]>--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-)create mode 100644
target/s390x/kvm/cpu_topology.c[...]
Quoted 5 lines
Quoted 5 lines
target/s390x/cpu.h +0
-0[...]
Quoted 10 lines
Quoted 8 lines
target/s390x/cpu.h typedef union SysIB { · line 636
Quoted 8 lines
target/s390x/cpu.h +8
-0+ +/* CPU type Topology List Entry */ +typedef struct SysIBTl_cpu { + uint8_t nl; + uint8_t reserved0[3]; +#define SYSIB_TLE_POLARITY_MASK 0x03 +#define SYSIB_TLE_DEDICATED 0x04 + uint8_t entitlement;I would just call this flags, since it's multiple fields.
OK
Quoted 18 lines
Quoted 9 lines
Quoted 9 lines
target/s390x/cpu.h +6
-0+uint8_t type; +uint16_t origin; +uint64_t mask; +} QEMU_PACKED QEMU_ALIGNED(8) SysIBTl_cpu; +QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 ); +[...]
Quoted 7 lines
Quoted 7 lines
/**
target/s390x/kvm/cpu_topology.c new file +0
-0[...]
Quoted 11 lines
Quoted 9 lines
target/s390x/kvm/cpu_topology.c line 66
Quoted 9 lines
target/s390x/kvm/cpu_topology.c new file +9
-0+ +/* +* Macro to check that the size of data after increment +* will not get bigger than the size of the SysIB. +*/ +#define SYSIB_GUARD (data, x) do { \ + data += x; \ + if (data > sizeof(SysIB)) { \ + return -ENOSPC; \I would go with ENOMEM here.
Considering your comment of the length in insert_stsi_15_1_x(), I think the best would be to return 0. Because:
- We do not need to report any error reason.
- The value will be forwarded to insert_stsi_15_1_x() as the length of the SYSIB to write
- On error we do not write the SYSIB (len = 0)
- In normal case the return value is always non zero and positive.
Quoted 12 lines
Quoted 10 lines
target/s390x/kvm/cpu_topology.c line 75
Quoted 10 lines
target/s390x/kvm/cpu_topology.c new file +9
-0+ } \ + }while (0 ) + +/** +* stsi_set_tle: +* @p: A pointer to the position of the first TLE +* @level: The nested level wanted by the guest +* +* Loop inside the s390_topology.list until the sentinelle entry s/sentinelle/sentinel/
OK, thx,
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines
+ * is found and for each entry:
+ * - Check usingSYSIB_GUARD()that the size of the SysIB is not
+ * reached.
+ * - Add all the container TLE needed for the level
+ * - Add the CPU TLE.I'd focus more on *what* the function does instead of *how*.
Right.
Quoted 5 lines
Fill the SYSIB with the topology information as described in the PoP, nesting containers as appropriate, with the maximum nesting limited by @level.
Or something similar.
Thanks, looks good to me .
Quoted 7 lines
Quoted 4 lines
Quoted 4 lines
+ *
+ * Return value:
+ *s390_top_set_levelreturns the size of theSysIB_15xafter beingYou forgot to rename the function here, right?
How aboutstsi_fill_topology_sysiborstsi_topology_fill_sysib, instead?
OK with stsi_topology_fill_sysib()
Quoted 5 lines
Quoted 3 lines
Quoted 3 lines
+ * filled with TLE on success.
+ * It returns-ENOSPCin the case we would overrun the end of the SysIB.You would have to change to ENOMEM here than also.
As discussed above, it seems to me that return 0 is even better.
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
+ */
target/s390x/kvm/cpu_topology.c new file +9
-0+static int stsi_set_tle (char *p,int level) +{ +S390TopologyEntry *entry; +int last_drawer =-1 ; +int last_book =-1 ; +int last_socket =-1 ; +int drawer_id =0 ; +int book_id =0 ; +int socket_id =0 ;
Quoted 8 lines
Quoted 6 lines
target/s390x/kvm/cpu_topology.c line 101
Quoted 6 lines
target/s390x/kvm/cpu_topology.c new file +6
-0+int n =sizeof (SysIB_151x); + + QTAILQ_FOREACH(entry, &s390_topology.list, next) { +int current_drawer = entry->id.drawer; +int current_book = entry->id.book; +int current_socket = entry->id.socket;This only saves two characters, so you could just use
entry->id. ...
OK
Quoted 6 lines
Quoted 4 lines
target/s390x/kvm/cpu_topology.c line 104
Quoted 4 lines
target/s390x/kvm/cpu_topology.c new file +3
-0+bool drawer_change = last_drawer != current_drawer; +bool book_change = drawer_change || last_book != current_book; +bool socket_change = book_change || last_socket != current_socket;... but keep it if it would make this line too long.
it is OK
Quoted 43 lines
You could also rename entry, to current or cur, if you want to emphasize that.
Quoted 30 lines
Quoted 30 lines
target/s390x/kvm/cpu_topology.c new file +30
-0+ +/* If we reach the guard get out */ +if (entry->id.level5) { +break ; + } + +if (level >3 && drawer_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,3 , drawer_id++); + book_id =0 ; + } +if (level >2 && book_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,2 , book_id++); + socket_id =0 ; + } +if (socket_change) { +SYSIB_GUARD (n,sizeof (SysIBTl_container)); + p =fill_container (p,1 , socket_id++); + } + +SYSIB_GUARD (n,sizeof (SysIBTl_cpu)); + p =fill_tle_cpu (p, entry); + last_drawer = entry->id.drawer; + last_book = entry->id.book; + last_socket = entry->id.socket; + } + +return n; +}Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 142
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+ +/** +* setup_stsi: +* sysib: pointer to a SysIB to be filled with SysIB_151x data +* level: Nested level specified by the guest +* +* Setup the SysIB_151x header before calling stsi_set_tle with +* a pointer to the first TLE entry. Same thing here with regards to describing the what.
Quoted 2 lines
Setup the SYSIB for STSI 15.1, the header as well as the description of the topology.
OK, thx
Quoted 50 lines
Quoted 50 lines
Quoted 50 lines
+ */
target/s390x/kvm/cpu_topology.c new file +48
-0+static int setup_stsi (SysIB_151x *sysib,int level) +{ + sysib->mnest = level; +switch (level) { +case 4 : + sysib->mag[S390_TOPOLOGY_MAG4 ] = current_machine->smp.drawers; + sysib->mag[S390_TOPOLOGY_MAG3 ] = current_machine->smp.books; + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; +case 3 : + sysib->mag[S390_TOPOLOGY_MAG3 ] = current_machine->smp.drawers * + current_machine->smp.books; + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; +case 2 : + sysib->mag[S390_TOPOLOGY_MAG2 ] = current_machine->smp.drawers * + current_machine->smp.books * + current_machine->smp.sockets; + sysib->mag[S390_TOPOLOGY_MAG1 ] = current_machine->smp.cores; +break ; + } + +return stsi_set_tle (sysib->tle, level); +} + +/** +* s390_topology_add_cpu_to_entry: +* @entry: Topology entry to setup +* @cpu: the S390CPU to add +* +* Set the core bit inside the topology mask and +* increments the number of cores for the socket. +*/ +static void s390_topology_add_cpu_to_entry (S390TopologyEntry *entry, +S390CPU *cpu) +{ +set_bit (63 - (cpu->env.core_id %64 ), &entry->mask); +} + +/** +* s390_topology_new_entry: +* @id: s390_topology_id to add +* @cpu: the S390CPU to add +* +* Allocate a new entry and initialize it. +*
Quoted 8 lines
Quoted 4 lines
target/s390x/kvm/cpu_topology.c line 195
Quoted 4 lines
+ * returns the newly allocated entry.
+ */target/s390x/kvm/cpu_topology.c new file +2
-0+static S390TopologyEntry *s390_topology_new_entry (s390_topology_id id, +S390CPU *cpu)This is used only once, right?
I think I'd go ahead and inline it intos390_topology_insert, since I had to go back and check ifnew_entrycallsadd_cpuwhen readings390_topology_insert.
OK
Quoted 79 lines
Quoted 68 lines
Quoted 68 lines
target/s390x/kvm/cpu_topology.c new file +67
-0+{ +S390TopologyEntry *entry; + + entry =g_malloc0 (sizeof (S390TopologyEntry)); + entry->id.id = id.id; +s390_topology_add_cpu_to_entry (entry, cpu); + +return entry; +} + +/** +* 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.polarity ==POLARITY_VERTICAL ) { +/* +* Vertical polarity with dedicated CPU implies +* vertical high entitlement. +*/ +if (topology_id.dedicated) { + topology_id.polarity |=POLARITY_VERTICAL_HIGH ; + }else { + topology_id.polarity |= cpu->env.entitlement; + } + } + +return topology_id; +} + +/** +* s390_topology_insert: +* @cpu: s390CPU insert. +* +* Parse the topology list to find if the entry already +* exist and add the core in it. +* If it does not exist, allocate a new entry and insert +* it in the queue from lower id to greater id. +*/ +static void s390_topology_insert (S390CPU *cpu) +{ +s390_topology_id id =s390_topology_from_cpu (cpu); +S390TopologyEntry *entry =NULL ; +S390TopologyEntry *tmp =NULL ; + + QTAILQ_FOREACH(tmp, &s390_topology.list, next) { +if (id.id == tmp->id.id) { +s390_topology_add_cpu_to_entry (tmp, cpu); +return ; + }else if (id.id < tmp->id.id) { + entry =s390_topology_new_entry (id, cpu); +QTAILQ_INSERT_BEFORE (tmp, entry, next); +return ; + } + }Quoted 8 lines
target/s390x/kvm/cpu_topology.c line 270
Quoted 8 lines
target/s390x/kvm/cpu_topology.c new file +8
-0+} + +/** +* s390_order_tle: +* +* Loop over all CPU and insert it at the right place +* inside the TLE entry list. +*/ Suggestion:
Quoted 4 lines
s390_topology_fill_list_sortedFill the S390Topology list with entries according to the order specified by the PoP.
OK
Quoted 17 lines
Quoted 15 lines
Quoted 5 lines
target/s390x/kvm/cpu_topology.c new file +4
-0+static void s390_order_tle (void ) +{ +CPUState *cs; +target/s390x/kvm/cpu_topology.c line 284
Quoted 10 lines
target/s390x/kvm/cpu_topology.c new file +10
-0+CPU_FOREACH (cs) { +s390_topology_insert (S390_CPU (cs)); + } +} + +/** +* s390_free_tle: +* +* Loop over all TLE entries and free them. +* Keep the sentinelle which is the only one with level5 != 0 s/sentinelle/sentinel/
yes, thx
Quoted 35 lines
Quoted 2 lines
Quoted 2 lines
+ */
Suggestion:
s390_topology_empty_listClear all entries in the S390Topology list except the sentinel.
Quoted 16 lines
Quoted 16 lines
target/s390x/kvm/cpu_topology.c new file +16
-0+static void s390_free_tle (void ) +{ +S390TopologyEntry *entry =NULL ; +S390TopologyEntry *tmp =NULL ; + + QTAILQ_FOREACH_SAFE(entry, &s390_topology.list, next, tmp) { +if (!entry->id.level5) { +QTAILQ_REMOVE (&s390_topology.list, entry, next); +g_free (entry); + } + } +} + +/** +* insert_stsi_15_1_x: +* cpu: the CPU doing the call for which we set CC Quoted 8 lines
Quoted 8 lines
+ * sel2: the selector 2, containing the nested level
+ * addr: Guest logical address of the guest SysIB
+ * ar: the access register number
+ *
+ * Reserve a zeroed SysIB, letsetup_stsito fill it and
+ * copy the SysIB to the guest memory.
+ *
+ * In case of overflow set CC(3) and no copy is done.Suggestion:
Quoted 3 lines
Emulate STSI 15.1.x, that is, perform all necessary checks and fill the SYSIB. In case the topology description is too long to fit into the SYSIB, set CC=3 and abort without writing the SYSIB.
OK, better thanks.
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
+ */
target/s390x/kvm/cpu_topology.c new file +9
-0+void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar) +{ +SysIB sysib = {0 }; +int len; + +if (!s390_has_topology () || sel2 <2 || sel2 >SCLP_READ_SCP_INFO_MNEST ) { +setcc (cpu,3 ); +return ; + }
Quoted 8 lines
Quoted 6 lines
target/s390x/kvm/cpu_topology.c line 325
Quoted 6 lines
target/s390x/kvm/cpu_topology.c new file +6
-0+ +s390_order_tle (); + + len =setup_stsi (&sysib.sysib_151x, sel2); + +if (len <0 ) {I stumbled a bit over this, maybe rename len to r.
Why ? it is the length used to fill the length field of the SYSIB.
May be it would be clearer if we give back the length to write and 0 on error then we have here:
if (!len) {
setcc(cpu, 3);
return;
}Quoted 11 lines
Quoted 11 lines
target/s390x/kvm/cpu_topology.c line 335
Quoted 11 lines
target/s390x/kvm/cpu_topology.c new file +10
-0+setcc (cpu,3 ); +return ; + } + + sysib.sysib_151x.length =cpu_to_be16 (len); +s390_cpu_virt_mem_write (cpu, addr, ar, &sysib, len); +setcc (cpu,0 ); + +s390_free_tle (); +}
Thanks for the comments.
If there are only comments and cosmetic changes will I get your RB if I change them according to your wishes?
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
Quoted 18 lines · Pierre Morel
Quoted 16 lines · Nina Schoetterl-Glausch
Quoted 15 lines · Pierre Morel
Quoted 15 lines
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 <[email protected]>--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-)create mode 100644
target/s390x/kvm/cpu_topology.c
[...]
Quoted 20 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
+ */
target/s390x/kvm/cpu_topology.c new file +9
-0+void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar) +{ +SysIB sysib = {0 }; +int len; + +if (!s390_has_topology () || sel2 <2 || sel2 >SCLP_READ_SCP_INFO_MNEST ) { +setcc (cpu,3 ); +return ; + }Quoted 8 lines
Quoted 6 lines
target/s390x/kvm/cpu_topology.c line 325
Quoted 6 lines
target/s390x/kvm/cpu_topology.c new file +6
-0+ +s390_order_tle (); + + len =setup_stsi (&sysib.sysib_151x, sel2); + +if (len <0 ) {I stumbled a bit over this, maybe rename len to r.
Quoted 1 line
Why ? it is the length used to fill the length field of the SYSIB.
Well a negative length doesn't make sense, that's what confused me for a bit. It's the error value of course, I proposed renaming it to the more generic r return value, to signify that the return value isn't exactly the length.
You're solution of using 0 is fine with me, too.
Quoted 10 lines
May be it would be clearer if we give back the length to write and 0 on error then we have here:
if (!len) { setcc(cpu, 3); return; }Quoted 1 line
Quoted 14 lines
Quoted 10 lines
Quoted 10 lines
target/s390x/kvm/cpu_topology.c line 335
Quoted 10 lines
target/s390x/kvm/cpu_topology.c new file +10
-0+setcc (cpu,3 ); +return ; + } + + sysib.sysib_151x.length =cpu_to_be16 (len); +s390_cpu_virt_mem_write (cpu, addr, ar, &sysib, len); +setcc (cpu,0 ); + +s390_free_tle (); +}Thanks for the comments.
If there are only comments and cosmetic changes will I get your RB if I change them according to your wishes?
Usually I review the whole series and then comment, this time sent feedback early, so the review isn't as deep. Probably easiest to give you the R-b for v16. My impression is that the series is close to final.
Quoted 4 lines
Regards,
Pierre
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
In reply to Pierre MorelQuoted 5 lines · Pierre Morel
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 <[email protected]>---
...
Quoted 91 lines
include/hw/s390x/sclp.h +1
-0@@ -112,6 +112,7 @@ typedef struct CPUEntry { }QEMU_PACKED CPUEntry;#define SCLP_READ_SCP_INFO_FIXED_CPU_OFFSET 128 +#define SCLP_READ_SCP_INFO_MNEST 2 typedef struct ReadInfo { SCCBHeader h; uint16_t rnmax;target/s390x/cpu.h +61
-0@@ -560,6 +560,25 @@ typedef struct SysIB_322 { } SysIB_322;QEMU_BUILD_BUG_ON (sizeof (SysIB_322) !=4096 ); +#define S390_TOPOLOGY_MAG 6 +#define S390_TOPOLOGY_MAG6 0 +#define S390_TOPOLOGY_MAG5 1 +#define S390_TOPOLOGY_MAG4 2 +#define S390_TOPOLOGY_MAG3 3 +#define S390_TOPOLOGY_MAG2 4 +#define S390_TOPOLOGY_MAG1 5 +/* Configuration topology */ +typedef struct SysIB_151x { +uint8_t reserved0[2 ]; +uint16_t length; +uint8_t mag[S390_TOPOLOGY_MAG ]; +uint8_t reserved1; +uint8_t mnest; +uint32_t reserved2; +char tle[]; +} QEMU_PACKED QEMU_ALIGNED(8) SysIB_151x; +QEMU_BUILD_BUG_ON (sizeof (SysIB_151x) !=16 ); + typedef union SysIB { SysIB_111 sysib_111; SysIB_121 sysib_121;@@ -567,9 +586,62 @@ typedef union SysIB {SysIB_221 sysib_221;SysIB_222 sysib_222;SysIB_322 sysib_322; +SysIB_151x sysib_151x; } SysIB;QEMU_BUILD_BUG_ON (sizeof (SysIB) !=4096 ); +/* +* 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 +* 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 ); + +/* CPU type Topology List Entry */ +typedef struct SysIBTl_cpu { + uint8_t nl; + uint8_t reserved0[3];
target/s390x/cpu.h typedef union SysIB { · line 643
Quoted 10 lines
target/s390x/cpu.h +10
-0+#define SYSIB_TLE_POLARITY_MASK 0x03 +#define SYSIB_TLE_DEDICATED 0x04 +uint8_t entitlement; +uint8_t type; +uint16_t origin; +uint64_t mask; +} QEMU_PACKED QEMU_ALIGNED(8) SysIBTl_cpu; +QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 ); + +void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar);
__u64 is not a portable type (e.g. fails when doing "make vm-build-openbsd"), please replace with uint64_t.
Thanks,
Thomas
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
Quoted 7 lines · Thomas Huth
Quoted 5 lines · Pierre Morel
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 <[email protected]>---
...
...
Quoted 2 lines
Quoted 2 lines
target/s390x/cpu.h +0
-0
...
Quoted 4 lines
target/s390x/cpu.h
Quoted 2 lines
target/s390x/cpu.h +2
-0+ +void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar);
__u64is not a portable type (e.g. fails when doing "make vm-build-openbsd"), please replace withuint64_t.
done.
Thanks
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
In reply to Pierre MorelQuoted 15 lines · Pierre Morel
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 <[email protected]>--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-) create mode 100644 target/s390x/kvm/cpu_topology.c
[...]
Quoted 18 lines
target/s390x/kvm/cpu_topology.c new file +18
-0+ +/** +* 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; +
target/s390x/kvm/cpu_topology.c line 232
Quoted 10 lines
target/s390x/kvm/cpu_topology.c new file +10
-0+if (s390_topology.polarity ==POLARITY_VERTICAL ) { +/* +* Vertical polarity with dedicated CPU implies +* vertical high entitlement. +*/ +if (topology_id.dedicated) { + topology_id.polarity |=POLARITY_VERTICAL_HIGH ; + }else { + topology_id.polarity |= cpu->env.entitlement; + }
Why |= instead of an assignment?
Anyway, I think you can get rid of this in the next version. If you define the entitlement via qapi you can just put a little switch here and convert it to the hardware definition of polarization. (Or you just do +1, but I think the switch is easier to understand)
Quoted 5 lines
target/s390x/kvm/cpu_topology.c new file +5
-0+ } + +return topology_id; +} +
[...]
Re: [PATCH v15 03/11] target/s390x/cpu topology: handle STSI(15) and build the SYSIB
Quoted 43 lines · Nina Schoetterl-Glausch
Quoted 15 lines · Pierre Morel
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 <[email protected]>--- include/hw/s390x/cpu-topology.h | 22 +++ include/hw/s390x/sclp.h | 1 + target/s390x/cpu.h | 72 +++++++ hw/s390x/cpu-topology.c | 10 + target/s390x/kvm/cpu_topology.c | 335 ++++++++++++++++++++++++++++++++ target/s390x/kvm/kvm.c | 5 +- target/s390x/kvm/meson.build | 3 +- 7 files changed, 446 insertions(+), 2 deletions(-)create mode 100644
target/s390x/kvm/cpu_topology.c[...]
Quoted 26 lines
target/s390x/kvm/cpu_topology.c new file +26
-0+ +/** +* 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.polarity ==POLARITY_VERTICAL ) { +/* +* Vertical polarity with dedicated CPU implies +* vertical high entitlement. +*/ +if (topology_id.dedicated) { + topology_id.polarity |=POLARITY_VERTICAL_HIGH ; + }else {
Quoted 8 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ topology_id.polarity |= cpu->env.entitlement; + }Why |= instead of an assignment?
Anyway, I think you can get rid of this in the next version. If you define the entitlement via qapi you can just put a little switch here and convert it to the hardware definition of polarization. (Or you just do +1, but I think the switch is easier to understand)
Oh! right thanks, it is a leftover from when dedication and polarity were in the same variable.
I change it with the QAPI enum.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 4/11 s390x/sclp: reporting the maximum nested topology entries
In reply to Pierre MorelThe maximum nested topology entries is used by the guest to know how many nested topology are available on the machine.
Let change the MNEST value from 2 to 4 in the SCLP READ INFO structure now that we support books and drawers.
--- include/hw/s390x/sclp.h | 5 +++-- hw/s390x/sclp.c | 5 +++++ 2 files changed, 8 insertions(+), 2 deletions(-)
include/hw/s390x/sclp.h +3 -2
@@ -112,12 +112,13 @@ typedef struct CPUEntry { }QEMU_PACKED CPUEntry;#define SCLP_READ_SCP_INFO_FIXED_CPU_OFFSET 128 -#define SCLP_READ_SCP_INFO_MNEST 2 +#define SCLP_READ_SCP_INFO_MNEST 4 typedef struct ReadInfo { SCCBHeader h; uint16_t rnmax; uint8_t rnsize; - uint8_t _reserved1[16 - 11]; /* 11-15 */ + uint8_t _reserved1[15 - 11]; /* 11-14 */ + uint8_t stsi_parm; /* 15-15 */ uint16_t entries_cpu; /* 16-17 */ uint16_t offset_cpu; /* 18-19 */ uint8_t _reserved2[24 - 20]; /* 20-23 */
hw/s390x/sclp.c +5 -0
@@ -20,6 +20,7 @@#include "hw/s390x/event-facility.h" #include "hw/s390x/s390-pci-bus.h" #include "hw/s390x/ipl.h" +#include "hw/s390x/cpu-topology.h" static inline SCLPDevice *get_sclp_device (void ) {
@@ -123,6 +124,10 @@ static void read_SCP_info(SCLPDevice *sclp, SCCB *sccb)return ; } +if (s390_has_topology ()) { + read_info->stsi_parm =SCLP_READ_SCP_INFO_MNEST ; + } +/* CPU information */ prepare_cpu_entries (machine, entries_start, &cpu_count); read_info->entries_cpu =cpu_to_be16 (cpu_count);
Signature · 1 line
2.31.1
Re: [PATCH v15 04/11] s390x/sclp: reporting the maximum nested topology entries
Quoted 6 lines · Pierre Morel
The maximum nested topology entries is used by the guest to know how many nested topology are available on the machine.
Let change the MNEST value from 2 to 4 in the SCLP READ INFO structure now that we support books and drawers.
Quoted 6 lines
Signed-off-by Pierre Morel <[email protected]>Reviewed-by Nina Schoetterl-Glausch <[email protected]>--- include/hw/s390x/sclp.h | 5 +++-- hw/s390x/sclp.c | 5 +++++ 2 files changed, 8 insertions(+), 2 deletions(-)
Re: [PATCH v15 04/11] s390x/sclp: reporting the maximum nested topology entries
Quoted 7 lines · Thomas Huth
Quoted 6 lines · Pierre Morel
The maximum nested topology entries is used by the guest to know how many nested topology are available on the machine.
Let change the MNEST value from 2 to 4 in the SCLP READ INFO structure now that we support books and drawers.
Quoted 9 lines
Quoted 6 lines
Signed-off-by Pierre Morel <[email protected]>Reviewed-by Nina Schoetterl-Glausch <[email protected]>---
include/hw/s390x/sclp.h| 5 +++--
hw/s390x/sclp.c | 5 +++++2 files changed, 8 insertions(+), 2 deletions(-)Reviewed-by Thomas Huth <[email protected]>
Thanks,
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 5/11 s390x/cpu topology: resetting the Topology-Change-Report
In reply to Pierre MorelDuring a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.
--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)
include/hw/s390x/cpu-topology.h +1 -0
@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif
target/s390x/cpu.h +1 -0
@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */
target/s390x/kvm/kvm_s390x.h +1 -0
@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */
hw/s390x/cpu-topology.c +12 -0
@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); } +/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{ +s390_cpu_topology_reset (); +} +/** * s390_topology_cpu_default: * @cpu: pointer to a S390CPU
hw/s390x/s390-virtio-ccw.c +3 -0
@@ -122,6 +122,9 @@ static void subsystem_reset(void)device_cold_reset (dev); } } +if (s390_has_topology ()) { +s390_topology_reset (); + } }static int virtio_ccw_hcall_notify (const uint64_t *args)
target/s390x/cpu-sysemu.c +13 -0
@@ -306,3 +306,16 @@ void s390_do_cpu_set_diag318(CPUState *cs, run_on_cpu_data arg)kvm_s390_set_diag318 (cs, arg.host_ulong); } } + +void s390_cpu_topology_reset (void ) +{ +int ret; + +if (kvm_enabled ()) { + ret =kvm_s390_topology_set_mtcr (0 ); +if (ret) { +error_report ("Failed to set Modified Topology Change Report: %s" , +strerror (-ret)); + } + } +}
target/s390x/kvm/kvm.c +17 -0
@@ -2592,6 +2592,23 @@ int kvm_s390_get_zpci_op(void)return cap_zpci_op; } +int kvm_s390_topology_set_mtcr (uint64_t attr) +{ +struct kvm_device_attr attribute = { + .group =KVM_S390_VM_CPU_TOPOLOGY , + .attr = attr, + }; + +if (!s390_has_feat (S390_FEAT_CONFIGURATION_TOPOLOGY )) { +return 0 ; + } +if (!kvm_vm_check_attr (kvm_state,KVM_S390_VM_CPU_TOPOLOGY , attr)) { +return -ENOTSUP ; + } + +return kvm_vm_ioctl (kvm_state,KVM_SET_DEVICE_ATTR , &attribute); +} +void kvm_arch_accel_class_init (ObjectClass *oc) { }
Signature · 1 line
2.31.1
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 7 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>---
...
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 92
Quoted 13 lines
hw/s390x/cpu-topology.c +5
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); } +/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology
Duplicated s390_topology_reset() in the comment.
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 98
Quoted 6 lines
hw/s390x/cpu-topology.c +6
-0+* change record. +*/ +void s390_topology_reset (void ) +{ +s390_cpu_topology_reset (); +}
With the nit fixed:
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 9 lines · Thomas Huth
Quoted 7 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>---
...
Quoted 14 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 91
Quoted 12 lines
hw/s390x/cpu-topology.c +5
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms) QTAILQ_INSERT_HEAD(&s390_topology.list, entry, next); } +/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology Duplicated
s390_topology_reset()in the comment.
right, thx.
Quoted 11 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 98
Quoted 7 lines
hw/s390x/cpu-topology.c +6
-0+* change record. +*/ +void s390_topology_reset (void ) +{ + s390_cpu_topology_reset(); +}With the nit fixed:
Reviewed-by Thomas Huth <[email protected]>
Thanks!
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
In reply to Pierre MorelQuoted 32 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 4 lines
target/s390x/cpu.h +1
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );
How about you call this s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
Quoted 21 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 10 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{
I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Quoted 69 lines
hw/s390x/cpu-topology.c +3
-0+s390_cpu_topology_reset (); +} +/** * s390_topology_cpu_default: * @cpu: pointer to a S390CPU hw/s390x/s390-virtio-ccw.c +3
-0@@ -122,6 +122,9 @@ static void subsystem_reset(void)device_cold_reset (dev); } } +if (s390_has_topology ()) { +s390_topology_reset (); + } }static int virtio_ccw_hcall_notify (const uint64_t *args)target/s390x/cpu-sysemu.c +13
-0@@ -306,3 +306,16 @@ void s390_do_cpu_set_diag318(CPUState *cs, run_on_cpu_data arg)kvm_s390_set_diag318 (cs, arg.host_ulong); } } + +void s390_cpu_topology_reset (void ) +{ +int ret; + +if (kvm_enabled ()) { + ret =kvm_s390_topology_set_mtcr (0 ); +if (ret) { +error_report ("Failed to set Modified Topology Change Report: %s" , +strerror (-ret)); + } + } +}target/s390x/kvm/kvm.c +17
-0@@ -2592,6 +2592,23 @@ int kvm_s390_get_zpci_op(void)return cap_zpci_op; } +int kvm_s390_topology_set_mtcr (uint64_t attr) +{ +struct kvm_device_attr attribute = { + .group =KVM_S390_VM_CPU_TOPOLOGY , + .attr = attr, + }; + +if (!s390_has_feat (S390_FEAT_CONFIGURATION_TOPOLOGY )) { +return 0 ; + } +if (!kvm_vm_check_attr (kvm_state,KVM_S390_VM_CPU_TOPOLOGY , attr)) { +return -ENOTSUP ; + } + +return kvm_vm_ioctl (kvm_state,KVM_SET_DEVICE_ATTR , &attribute); +} +void kvm_arch_accel_class_init (ObjectClass *oc) { }
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );
Quoted 6 lines
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 2 lines
target/s390x/cpu.h +1
-0void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );How about you call this
s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.
I am not sure about this, Thomas already gave his R-B on this patch so I prefer to stay on the original name, unless he says it is a good idea to change. Also in cpu-sysemu.c most of the function are tagged with _cpu_
Quoted 4 lines
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
There is already a single function kvm_s390_topology_set_mtcr(attr) to set the "modified state"
Quoted 23 lines
Quoted 23 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }
Quoted 12 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 9 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Here in this patch we do not intercept PTF and we have only an horizontal polarity.
So we do not need to reset the polarity for all the vCPUs, we only need it when we have vertical polarity.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 44 lines · Pierre Morel
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );Quoted 6 lines
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 2 lines
target/s390x/cpu.h +1
-0void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );How about you call this
s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.
Quoted 4 lines
I am not sure about this, Thomas already gave his R-B on this patch so I prefer to stay on the original name, unless he says it is a good idea to change.
Also incpu-sysemu.cmost of the function are tagged with _cpu_
IMO, renaming a function would be a small enough change to keep an R-b.
Quoted 8 lines
Quoted 5 lines
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
There is already a single function
kvm_s390_topology_set_mtcr(attr)to set the "modified state"
Yes, but that is for KVM only and doesn't do error handling.
So you need at least one function on top of that. What I'm suggesting is to only have one function instead of two because it gets rid of some code.
Quoted 37 lines
Quoted 24 lines
Quoted 24 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }Quoted 12 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 9 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Quoted 4 lines
Here in this patch we do not intercept PTF and we have only an horizontal polarity.
So we do not need to reset the polarity for all the vCPUs, we only need it when we have vertical polarity.
Well, with the PTF patch you don't get vertical polarity either, because you only enable the topology with patch 7.
And since it's about resetting, it fits better in this patch IMO.
Quoted 4 lines
Regards,
Pierre
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 57 lines · Nina Schoetterl-Glausch
Quoted 44 lines · Pierre Morel
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );Quoted 6 lines
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 2 lines
target/s390x/cpu.h +1
-0void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );How about you call this
s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.Quoted 4 lines
I am not sure about this, Thomas already gave his R-B on this patch so I prefer to stay on the original name, unless he says it is a good idea to change.
Also incpu-sysemu.cmost of the function are tagged with _cpu_IMO, renaming a function would be a small enough change to keep an R-b.
Quoted 6 lines
Quoted 5 lines
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
Quoted 6 lines
Quoted 2 lines
There is already a single function
kvm_s390_topology_set_mtcr(attr)to set the "modified state"Yes, but that is for KVM only and doesn't do error handling.
So you need at least one function on top of that. What I'm suggesting is to only have one function instead of two because it gets rid of some code.
OK this is right.
I rename
void s390_cpu_topology_reset(void);
to
void s390_cpu_topology_set_mtcr(int value);
and then:
- s390_cpu_topology_reset();
+ s390_cpu_topology_set_mtcr(0);
Quoted 37 lines
Quoted 37 lines
Quoted 36 lines
Quoted 23 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 10 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Quoted 8 lines
Quoted 4 lines
Here in this patch we do not intercept PTF and we have only an horizontal polarity.
So we do not need to reset the polarity for all the vCPUs, we only need it when we have vertical polarity.Well, with the PTF patch you don't get vertical polarity either, because you only enable the topology with patch 7.
And since it's about resetting, it fits better in this patch IMO.
Not in my opinion, suppose the next patch never get included it has no sense. However if there is a majority in favor of this change I will do it.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 66 lines · Pierre Morel
Quoted 57 lines · Nina Schoetterl-Glausch
Quoted 44 lines · Pierre Morel
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );Quoted 6 lines
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 2 lines
target/s390x/cpu.h +1
-0void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );How about you call this
s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.Quoted 4 lines
I am not sure about this, Thomas already gave his R-B on this patch so I prefer to stay on the original name, unless he says it is a good idea to change.
Also incpu-sysemu.cmost of the function are tagged with _cpu_IMO, renaming a function would be a small enough change to keep an R-b.
Quoted 6 lines
Quoted 5 lines
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
Quoted 6 lines
Quoted 2 lines
There is already a single function
kvm_s390_topology_set_mtcr(attr)to set the "modified state"Yes, but that is for KVM only and doesn't do error handling.
So you need at least one function on top of that. What I'm suggesting is to only have one function instead of two because it gets rid of some code.
Quoted 5 lines
OK this is right.
I rename
voids390_cpu_topology_reset(void);
to
voids390_cpu_topology_set_mtcr(int value);
I don't find mtcr very descriptive and a bit of a SIE/KVM name, it might not fit a possible future tcg implementation.
I'd just call it s390_cpu_topology_set_changed/modified, and have it take a bool, because I cannot imagine other int values to make sense.
Quoted 49 lines
and then:
(quoted diff) +1
-1- s390_cpu_topology_reset(); + s390_cpu_topology_set_mtcr(0);Quoted 44 lines
Quoted 43 lines
Quoted 26 lines
Quoted 26 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }Quoted 12 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 9 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Here in this patch we do not intercept PTF and we have only an horizontal polarity.
So we do not need to reset the polarity for all the vCPUs, we only need it when we have vertical polarity.
Quoted 6 lines
Quoted 3 lines
Well, with the PTF patch you don't get vertical polarity either, because you only enable the topology with patch 7.
And since it's about resetting, it fits better in this patch IMO.Not in my opinion, suppose the next patch never get included it has no sense.
Well, yes, but then the series would be broken, since the facility requires PTF to work.
Quoted 5 lines
However if there is a majority in favor of this change I will do it.
Regards,
Pierre
Re: [PATCH v15 05/11] s390x/cpu topology: resetting the Topology-Change-Report
Quoted 73 lines · Nina Schoetterl-Glausch
Quoted 66 lines · Pierre Morel
Quoted 57 lines · Nina Schoetterl-Glausch
Quoted 44 lines · Pierre Morel
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
During a subsystem reset the Topology-Change-Report is cleared by the machine.
Let's ask KVM to clear the Modified Topology Change Report (MTCR) bit of the SCA in the case of a subsystem reset.Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/cpu-topology.h | 1 + target/s390x/cpu.h | 1 + target/s390x/kvm/kvm_s390x.h | 1 + hw/s390x/cpu-topology.c | 12 ++++++++++++ hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/cpu-sysemu.c | 13 +++++++++++++ target/s390x/kvm/kvm.c | 17 +++++++++++++++++ 7 files changed, 48 insertions(+)include/hw/s390x/cpu-topology.h +1
-0@@ -66,5 +66,6 @@ static inline void s390_topology_set_cpu(MachineState *ms,extern S390Topology s390_topology;int s390_socket_nb (S390CPU *cpu); +void s390_topology_reset (void );#endif target/s390x/cpu.h +0
-0@@ -641,6 +641,7 @@ typedef struct SysIBTl_cpu {QEMU_BUILD_BUG_ON (sizeof (SysIBTl_cpu) !=16 );Quoted 6 lines
target/s390x/cpu.h typedef struct SysIBTl_cpu { · line 644
Quoted 2 lines
target/s390x/cpu.h +1
-0void insert_stsi_15_1_x (S390CPU *cpu,int sel2,__u64 addr,uint8_t ar); +void s390_cpu_topology_reset (void );How about you call this
s390_cpu_topology_reset_modified, so it's symmetric with the function you define in the next patch. You could also drop the "cpu" from the name.Quoted 4 lines
I am not sure about this, Thomas already gave his R-B on this patch so I prefer to stay on the original name, unless he says it is a good idea to change.
Also incpu-sysemu.cmost of the function are tagged with _cpu_IMO, renaming a function would be a small enough change to keep an R-b.
Quoted 6 lines
Quoted 5 lines
Or maybe even better, you only define a function for setting the modified state, but make it take a bool argument. This way you also get rid of some code duplication and it wouldn't harm readability IMO.
Quoted 6 lines
Quoted 2 lines
There is already a single function
kvm_s390_topology_set_mtcr(attr)to set the "modified state"Yes, but that is for KVM only and doesn't do error handling.
So you need at least one function on top of that. What I'm suggesting is to only have one function instead of two because it gets rid of some code.Quoted 5 lines
OK this is right.
I rename
voids390_cpu_topology_reset(void);
to
voids390_cpu_topology_set_mtcr(int value);
Quoted 4 lines
I don't find mtcr very descriptive and a bit of a SIE/KVM name, it might not fit a possible future tcg implementation.
I'd just call it s390_cpu_topology_set_changed/modified, and have it take a bool, because I cannot imagine other int values to make sense.
OK
Quoted 50 lines
Quoted 50 lines
and then:
(quoted diff) +1
-1- s390_cpu_topology_reset(); + s390_cpu_topology_set_mtcr(0);Quoted 44 lines
Quoted 43 lines
Quoted 26 lines
Quoted 26 lines
target/s390x/cpu.h +0
-0/* MMU defines */ #define ASCE_ORIGIN (~0xfffULL)/* segment table origin */ target/s390x/kvm/kvm_s390x.h +1
-0@@ -47,5 +47,6 @@ void kvm_s390_crypto_reset(void);void kvm_s390_restart_interrupt (S390CPU *cpu);void kvm_s390_stop_interrupt (S390CPU *cpu);void kvm_s390_set_diag318 (CPUState *cs,uint64_t diag318_info); +int kvm_s390_topology_set_mtcr (uint64_t attr);#endif /* KVM_S390X_H */ hw/s390x/cpu-topology.c +0
-0@@ -85,6 +85,18 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }Quoted 12 lines
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 96
Quoted 9 lines
hw/s390x/cpu-topology.c +9
-0+/** +* s390_topology_reset: +* +* Generic reset for CPU topology, calls s390_topology_reset() +* s390_topology_reset() to reset the kernel Modified Topology +* change record. +*/ +void s390_topology_reset (void ) +{I'm wondering if you shouldn't move the reset changes you do in the next patch into this one. I don't see what they have to do with PTF emulation.
Here in this patch we do not intercept PTF and we have only an horizontal polarity.
So we do not need to reset the polarity for all the vCPUs, we only need it when we have vertical polarity.
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Well, with the PTF patch you don't get vertical polarity either, because you only enable the topology with patch 7.
And since it's about resetting, it fits better in this patch IMO.Not in my opinion, suppose the next patch never get included it has no sense.
Well, yes, but then the series would be broken, since the facility requires PTF to work.
Yes, same if the activation of the facility is not included.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 6/11 s390x/cpu topology: interception of PTF instruction
In reply to Pierre MorelWhen the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation by using the KVM_CAP_S390_CPU_TOPOLOGY KVM extension.
The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
During RESET all CPU of the configuration are placed in horizontal polarity.
--- include/hw/s390x/s390-virtio-ccw.h | 6 ++ target/s390x/cpu.h | 1 + hw/s390x/cpu-topology.c | 103 +++++++++++++++++++++++++++++ target/s390x/cpu-sysemu.c | 14 ++++ target/s390x/kvm/kvm.c | 11 +++ 5 files changed, 135 insertions(+)
include/hw/s390x/s390-virtio-ccw.h +6 -0
@@ -30,6 +30,12 @@ struct S390CcwMachineState {uint8_t loadparm[8 ]; }; +#define S390_PTF_REASON_NONE (0x00 << 8) +#define S390_PTF_REASON_DONE (0x01 << 8) +#define S390_PTF_REASON_BUSY (0x02 << 8) +#define S390_TOPO_FC_MASK 0xffUL +void s390_handle_ptf (S390CPU *cpu,uint8_t r1,uintptr_t ra); +struct S390CcwMachineClass {/*< private >*/ MachineClass parent_class;
target/s390x/cpu.h +1 -0
@@ -857,6 +857,7 @@ void s390_enable_css_support(S390CPU *cpu);void s390_do_cpu_set_diag318 (CPUState *cs,run_on_cpu_data arg);int s390_assign_subch_ioeventfd (EventNotifier *notifier,uint32_t sch_id,int vq,bool assign); +void s390_cpu_topology_set_modified (void );#ifndef CONFIG_USER_ONLY unsigned int s390_cpu_set_state (uint8_t cpu_state,S390CPU *cpu);#else
hw/s390x/cpu-topology.c +103 -0
125 lines not on this page · raw message
target/s390x/cpu-sysemu.c +14 -0
@@ -37,6 +37,7 @@#include "sysemu/sysemu.h" #include "sysemu/tcg.h" #include "hw/core/sysemu-cpu-ops.h" +#include "hw/s390x/cpu-topology.h" /* S390CPUClass::load_normal() */ static void s390_cpu_load_normal (CPUState *s)
@@ -319,3 +320,16 @@ void s390_cpu_topology_reset(void) } } } + +void s390_cpu_topology_set_modified (void ) +{ +int ret; + +if (kvm_enabled ()) { + ret =kvm_s390_topology_set_mtcr (1 ); +if (ret) { +error_report ("Failed to set Modified Topology Change Report: %s" , +strerror (-ret)); + } + } +}
target/s390x/kvm/kvm.c +11 -0
@@ -96,6 +96,7 @@#define PRIV_B9_EQBS 0x9c#define PRIV_B9_CLP 0xa0 +#define PRIV_B9_PTF 0xa2#define PRIV_B9_PCISTG 0xd0#define PRIV_B9_PCILG 0xd2#define PRIV_B9_RPCIT 0xd3
@@ -1464,6 +1465,13 @@ static int kvm_mpcifc_service_call(S390CPU *cpu, struct kvm_run *run) } } +static void kvm_handle_ptf (S390CPU *cpu,struct kvm_run *run) +{ +uint8_t r1 = (run->s390_sieic.ipb >>20 ) &0x0f ; + +s390_handle_ptf (cpu, r1,RA_IGNORED ); +} +static int handle_b9 (S390CPU *cpu,struct kvm_run *run,uint8_t ipa1) {int r =0 ;
@@ -1481,6 +1489,9 @@ static int handle_b9(S390CPU *cpu, struct kvm_run *run, uint8_t ipa1)case PRIV_B9_RPCIT : r =kvm_rpcit_service_call (cpu, run);break ; +case PRIV_B9_PTF : +kvm_handle_ptf (cpu, run); +break ;case PRIV_B9_EQBS :/* just inject exception */ r =-1 ;
Signature · 1 line
2.31.1
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
Quoted 3 lines · Pierre Morel
When the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation
s/hypervizor/hypervisor/
Quoted 4 lines
by using the
KVM_CAP_S390_CPU_TOPOLOGYKVM extension.The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
dito
Quoted 5 lines
During RESET all CPU of the configuration are placed in horizontal polarity.
Signed-off-by Pierre Morel <[email protected]>---
...
Quoted 7 lines
/**
*s390_topology_reset:
*
* Generic reset for CPU topology, callss390_topology_reset()
*s390_topology_reset()to reset the kernel Modified Topology
* change record.
+ * Then set global and all CPUs polarity toPOLARITY_HORIZONTAL.
You describe here already what's going to happen...
Quoted 5 lines
*/ void s390_topology_reset(void) { s390_cpu_topology_reset();+ /* Set global polarity to
POLARITY_HORIZONTAL*/
... then here again ...
Quoted 1 line
+ s390_topology.polarity =
POLARITY_HORIZONTAL;
... and the code is (fortunately) also very self-exaplaining...
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 185
Quoted 2 lines
hw/s390x/cpu-topology.c +2
-0+/* Set all CPU polarity to POLARITY_HORIZONTAL */ +s390_topology_set_cpus_polarity (POLARITY_HORIZONTAL );
... so I'd rather drop the two comments within the function body.
Quoted 1 line
}
(rest of the patch looks fine to me)
Thomas
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
Quoted 6 lines · Thomas Huth
Quoted 3 lines · Pierre Morel
When the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation
s/hypervizor/hypervisor/
grr, I was pretty sure you already said that to me and I did change it... seems I did not.
done now.
Quoted 7 lines
Quoted 5 lines
by using the
KVM_CAP_S390_CPU_TOPOLOGYKVM extension.The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
dito
yes, thx.
Quoted 25 lines
Quoted 6 lines
During RESET all CPU of the configuration are placed in horizontal polarity.
Signed-off-by Pierre Morel <[email protected]>---
...
Quoted 7 lines
/**
*s390_topology_reset:
*
* Generic reset for CPU topology, callss390_topology_reset()
*s390_topology_reset()to reset the kernel Modified Topology
* change record.
+ * Then set global and all CPUs polarity toPOLARITY_HORIZONTAL.You describe here already what's going to happen...
Quoted 5 lines
*/
voids390_topology_reset(void)
{
s390_cpu_topology_reset();
+ /* Set global polarity toPOLARITY_HORIZONTAL*/... then here again ...
Quoted 8 lines
Quoted 1 line
+ s390_topology.polarity =
POLARITY_HORIZONTAL;... and the code is (fortunately) also very self-exaplaining...
Quoted 2 lines
+ /* Set all CPU polarity to
POLARITY_HORIZONTAL*/ +s390_topology_set_cpus_polarity(POLARITY_HORIZONTAL);... so I'd rather drop the two comments within the function body.
OK
Quoted 7 lines
Quoted 2 lines
}
(rest of the patch looks fine to me)
Thomas
Thanks.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
In reply to Pierre MorelQuoted 20 lines · Pierre Morel
When the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation by using the
KVM_CAP_S390_CPU_TOPOLOGYKVM extension.The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
During RESET all CPU of the configuration are placed in horizontal polarity.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/s390-virtio-ccw.h | 6 ++ target/s390x/cpu.h | 1 + hw/s390x/cpu-topology.c | 103 +++++++++++++++++++++++++++++ target/s390x/cpu-sysemu.c | 14 ++++ target/s390x/kvm/kvm.c | 11 +++ 5 files changed, 135 insertions(+)
[...]
Quoted 12 lines
hw/s390x/cpu-topology.c +4
-0@@ -85,16 +85,104 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); } +/** +* s390_topology_set_cpus_polarity: +* @polarity: polarity requested by the caller +*
hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 101
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+* Set all CPU entitlement according to polarity and +* dedication. +* Default vertical entitlement is POLARITY_VERTICAL_MEDIUM as +* it does not require host modification of the CPU provisioning +* until the host decide to modify individual CPU provisioning +* using QAPI interface. +* However a dedicated vCPU will have a POLARITY_VERTICAL_HIGH +* entitlement. +*/ +static void s390_topology_set_cpus_polarity (int polarity)
Since you set the entitlement field I'd prefer _set_cpus_entitlement or similar.
Quoted 14 lines
hw/s390x/cpu-topology.c +14
-0+{ +CPUState *cs; + +CPU_FOREACH (cs) { +if (polarity ==POLARITY_HORIZONTAL ) { +S390_CPU (cs)->env.entitlement =0 ; + }else if (S390_CPU (cs)->env.dedicated) { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_HIGH ; + }else { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_MEDIUM ; + } + } +} +
[...]
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 227
Quoted 6 lines
/**
hw/s390x/cpu-topology.c +0
-0@@ -137,6 +225,21 @@ static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) (smp->books * smp->sockets * smp->cores)) % smp->drawers; }
Why are the changes below in this patch?
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 236
Quoted 9 lines
hw/s390x/cpu-topology.c +9
-0+ +/* +* Machine polarity is set inside the global s390_topology structure. +* In the case the polarity is set as horizontal set the entitlement +* to POLARITY_VERTICAL_MEDIUM which is the better equivalent when +* machine polarity is set to vertical or POLARITY_VERTICAL_HIGH if +* the vCPU is dedicated. +*/ +if (s390_topology.polarity && !env->entitlement) {
It'd be more readable if you compared against enum values by name.
I don't see why you check s390_topology.polarity. If it is horizontal then the value of the entitlement doesn't matter at all, so you can set it to whatever.
All you want to do is enforce dedicated -> VERTICAL_HIGH, right? So why don't you just add
+ if (cpu->env.dedicated && cpu->env.entitlement != POLARITY_VERTICAL_HIGH) {
+ error_setg(errp, "A dedicated cpu implies high entitlement");
+ return;
+ }
to s390_topology_check?
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 241
Quoted 5 lines
hw/s390x/cpu-topology.c +5
-0+if (env->dedicated) { + env->entitlement =POLARITY_VERTICAL_HIGH ; + }else { + env->entitlement =POLARITY_VERTICAL_MEDIUM ; + }
If it is horizontal, then setting the entitlement is pointless as it will be reset to medium on PTF.
So the current polarization is vertical and a cpu is being hotplugged, but setting the entitlement of the cpu being added is also pointless, because it's determined by the dedication. That seems weird.
Quoted 3 lines
hw/s390x/cpu-topology.c +1
-0+ } }
[...]
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
Quoted 31 lines · Nina Schoetterl-Glausch
Quoted 20 lines · Pierre Morel
When the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation by using the
KVM_CAP_S390_CPU_TOPOLOGYKVM extension.The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
During RESET all CPU of the configuration are placed in horizontal polarity.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/s390-virtio-ccw.h | 6 ++ target/s390x/cpu.h | 1 + hw/s390x/cpu-topology.c | 103 +++++++++++++++++++++++++++++ target/s390x/cpu-sysemu.c | 14 ++++ target/s390x/kvm/kvm.c | 11 +++ 5 files changed, 135 insertions(+)[...]
Quoted 8 lines
hw/s390x/cpu-topology.c +0
-0@@ -85,16 +85,104 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }
Quoted 16 lines
Quoted 4 lines
hw/s390x/cpu-topology.c +4
-0+/** +* s390_topology_set_cpus_polarity: +* @polarity: polarity requested by the caller +* hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 101
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+* Set all CPU entitlement according to polarity and +* dedication. +* Default vertical entitlement is POLARITY_VERTICAL_MEDIUM as +* it does not require host modification of the CPU provisioning +* until the host decide to modify individual CPU provisioning +* using QAPI interface. +* However a dedicated vCPU will have a POLARITY_VERTICAL_HIGH +* entitlement. +*/ +static void s390_topology_set_cpus_polarity (int polarity)Since you set the entitlement field I'd prefer
_set_cpus_entitlementor similar.
OK if you prefer.
Quoted 17 lines
Quoted 15 lines
hw/s390x/cpu-topology.c +14
-0+{ +CPUState *cs; + +CPU_FOREACH (cs) { +if (polarity ==POLARITY_HORIZONTAL ) { +S390_CPU (cs)->env.entitlement =0 ; + }else if (S390_CPU (cs)->env.dedicated) { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_HIGH ; + }else { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_MEDIUM ; + } + } +} +[...]
Quoted 1 line
Quoted 7 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 227
Quoted 5 lines
/**
hw/s390x/cpu-topology.c +0
-0@@ -137,6 +225,21 @@ static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) (smp->books * smp->sockets * smp->cores)) % smp->drawers; }Why are the changes below in this patch?
Because before thos patch we have only horizontal polarization.
Quoted 5 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 231
Quoted 5 lines
hw/s390x/cpu-topology.c +4
-0+ +/* +* Machine polarity is set inside the global s390_topology structure. +* In the case the polarity is set as horizontal set the entitlement
Sorry here an error in the comment should be :
"In the case the polarity is NOT set as horizontal..."
Quoted 7 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 236
Quoted 5 lines
hw/s390x/cpu-topology.c +5
-0+* to POLARITY_VERTICAL_MEDIUM which is the better equivalent when +* machine polarity is set to vertical or POLARITY_VERTICAL_HIGH if +* the vCPU is dedicated. +*/ +if (s390_topology.polarity && !env->entitlement) {It'd be more readable if you compared against enum values by name.
Right, I will change this to
if (s390_topology.polarity != S390_POLARITY_HORIZONTAL &&
env->entitlement == S390_ENTITLEMENT_UNSET) {Quoted 4 lines
I don't see why you check s390_topology.polarity. If it is horizontal then the value of the entitlement doesn't matter at all, so you can set it to whatever.
Right, that is why it is done only for vertical polarization (sorry for the wrong comment)
Quoted 8 lines
All you want to do is enforce dedicated ->
VERTICAL_HIGH, right? So why don't you just add(quoted diff) +4
-0+ if (cpu->env.dedicated && cpu->env.entitlement != POLARITY_VERTICAL_HIGH) { + error_setg(errp, "A dedicated cpu implies high entitlement"); + return; + } >to
s390_topology_check?
Here it is to set the default in the case the values are not provided.
But where you are right is that I should add a verification to the check function.
Quoted 9 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 241
Quoted 6 lines
hw/s390x/cpu-topology.c +5
-0+if (env->dedicated) { + env->entitlement =POLARITY_VERTICAL_HIGH ; + }else { + env->entitlement =POLARITY_VERTICAL_MEDIUM ; + }If it is horizontal, then setting the entitlement is pointless as it will be reset to medium on PTF.
That is why the polarity is tested (sorry for the bad comment)
Quoted 3 lines
So the current polarization is vertical and a cpu is being hotplugged, but setting the entitlement of the cpu being added is also pointless, because it's determined by the dedication. That seems weird.
No it is not determined by the dedication, if there is no dedication the 3 vertical values are possible.
Regards,
Pierre
Quoted 6 lines
Quoted 4 lines
hw/s390x/cpu-topology.c +1
-0+ } }[...]
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
Quoted 77 lines · Pierre Morel
Quoted 31 lines · Nina Schoetterl-Glausch
Quoted 20 lines · Pierre Morel
When the host supports the CPU topology facility, the PTF instruction with function code 2 is interpreted by the SIE, provided that the userland hypervizor activates the interpretation by using the
KVM_CAP_S390_CPU_TOPOLOGYKVM extension.The PTF instructions with function code 0 and 1 are intercepted and must be emulated by the userland hypervizor.
During RESET all CPU of the configuration are placed in horizontal polarity.
Signed-off-by Pierre Morel <[email protected]>--- include/hw/s390x/s390-virtio-ccw.h | 6 ++ target/s390x/cpu.h | 1 + hw/s390x/cpu-topology.c | 103 +++++++++++++++++++++++++++++ target/s390x/cpu-sysemu.c | 14 ++++ target/s390x/kvm/kvm.c | 11 +++ 5 files changed, 135 insertions(+)[...]
Quoted 8 lines
hw/s390x/cpu-topology.c +0
-0@@ -85,16 +85,104 @@ static void s390_topology_init(MachineState *ms)QTAILQ_INSERT_HEAD (&s390_topology.list, entry, next); }Quoted 16 lines
Quoted 4 lines
hw/s390x/cpu-topology.c +4
-0+/** +* s390_topology_set_cpus_polarity: +* @polarity: polarity requested by the caller +* hw/s390x/cpu-topology.c static void s390_topology_init(MachineState *ms) · line 101
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+* Set all CPU entitlement according to polarity and +* dedication. +* Default vertical entitlement is POLARITY_VERTICAL_MEDIUM as +* it does not require host modification of the CPU provisioning +* until the host decide to modify individual CPU provisioning +* using QAPI interface. +* However a dedicated vCPU will have a POLARITY_VERTICAL_HIGH +* entitlement. +*/ +static void s390_topology_set_cpus_polarity (int polarity)Since you set the entitlement field I'd prefer
_set_cpus_entitlementor similar.OK if you prefer.
Quoted 17 lines
Quoted 15 lines
hw/s390x/cpu-topology.c +14
-0+{ +CPUState *cs; + +CPU_FOREACH (cs) { +if (polarity ==POLARITY_HORIZONTAL ) { +S390_CPU (cs)->env.entitlement =0 ; + }else if (S390_CPU (cs)->env.dedicated) { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_HIGH ; + }else { +S390_CPU (cs)->env.entitlement =POLARITY_VERTICAL_MEDIUM ; + } + } +} +[...]
Quoted 1 line
Quoted 7 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 227
Quoted 5 lines
/**
hw/s390x/cpu-topology.c +0
-0@@ -137,6 +225,21 @@ static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) (smp->books * smp->sockets * smp->cores)) % smp->drawers; }Why are the changes below in this patch?
Quoted 1 line
Because before thos patch we have only horizontal polarization.
Not really since you only enable topology in the next patch.
Quoted 24 lines
Quoted 6 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 231
Quoted 6 lines
hw/s390x/cpu-topology.c +4
-0+ +/* +* Machine polarity is set inside the global s390_topology structure. +* In the case the polarity is set as horizontal set the entitlement Sorry here an error in the comment should be :
"In the case the polarity is NOT set as horizontal..."Quoted 7 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 236
Quoted 5 lines
hw/s390x/cpu-topology.c +5
-0+* to POLARITY_VERTICAL_MEDIUM which is the better equivalent when +* machine polarity is set to vertical or POLARITY_VERTICAL_HIGH if +* the vCPU is dedicated. +*/ +if (s390_topology.polarity && !env->entitlement) {It'd be more readable if you compared against enum values by name.
Right, I will change this to
if (s390_topology.polarity != S390_POLARITY_HORIZONTAL && env->entitlement == S390_ENTITLEMENT_UNSET) {Quoted 1 line
Quoted 6 lines
Quoted 3 lines
I don't see why you check s390_topology.polarity. If it is horizontal then the value of the entitlement doesn't matter at all, so you can set it to whatever.
Right, that is why it is done only for vertical polarization (sorry for the wrong comment)
I'm saying you don't need to check it at all. You adjust the values for vertical polarization, but you could just always do that since the values don't matter at all if the polarization is horizontal.
Quoted 11 lines
Quoted 9 lines
All you want to do is enforce dedicated ->
VERTICAL_HIGH, right? So why don't you just add(quoted diff) +4
-0+ if (cpu->env.dedicated && cpu->env.entitlement != POLARITY_VERTICAL_HIGH) { + error_setg(errp, "A dedicated cpu implies high entitlement"); + return; + } >to
s390_topology_check?Here it is to set the default in the case the values are not provided.
If no values are provided, they default to dedication=false and entitlement=medium, as defined by patch 1, which are fine and don't need to be adjusted.
Quoted 11 lines
But where you are right is that I should add a verification to the check function.
Quoted 7 lines
Quoted 6 lines
hw/s390x/cpu-topology.c +5
-0+if (env->dedicated) { + env->entitlement =POLARITY_VERTICAL_HIGH ; + }else { + env->entitlement =POLARITY_VERTICAL_MEDIUM ; + }
Quoted 4 lines
Quoted 2 lines
If it is horizontal, then setting the entitlement is pointless as it will be reset to medium on PTF.
That is why the polarity is tested (sorry for the bad comment)
I said this because I'm fine with setting it pointlessly.
Quoted 6 lines
Quoted 3 lines
So the current polarization is vertical and a cpu is being hotplugged, but setting the entitlement of the cpu being added is also pointless, because it's determined by the dedication. That seems weird.
No it is not determined by the dedication, if there is no dedication the 3 vertical values are possible.
You set it to either high or medium based on the dedication. And for horizontal polarization it obviously doesn't matter.
As far as I understand you don't need this because the default values are fine. You should add the check that if a dedicated cpu is hotplugged, then the entitlement must be high, to patch 2 and that's it, no additional changes necessary.
Quoted 12 lines
Regards,
PierreQuoted 6 lines
Quoted 4 lines
hw/s390x/cpu-topology.c +1
-0+ } }[...]
Re: [PATCH v15 06/11] s390x/cpu topology: interception of PTF instruction
Quoted 4 lines · Nina Schoetterl-Glausch
Quoted 3 lines · Pierre Morel
Quoted 1 line · Nina Schoetterl-Glausch
On Wed, 2023-02-01 at 14:20 +0100, Pierre Morel wrote:
...
Quoted 41 lines
Quoted 11 lines
Quoted 9 lines
[...]
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 227
Quoted 6 lines
/**hw/s390x/cpu-topology.c +0
-0@@ -137,6 +225,21 @@ static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) (smp->books * smp->sockets * smp->cores)) % smp->drawers; }Why are the changes below in this patch?
Because before thos patch we have only horizontal polarization.
Not really since you only enable topology in the next patch.
Quoted 28 lines
Quoted 6 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 231
Quoted 6 lines
hw/s390x/cpu-topology.c +4
-0+ +/* +* Machine polarity is set inside the global s390_topology structure. +* In the case the polarity is set as horizontal set the entitlement Sorry here an error in the comment should be :
"In the case the polarity is NOT set as horizontal..."Quoted 7 lines
hw/s390x/cpu-topology.c static void s390_topology_cpu_default(S390CPU *cpu, Error **errp) · line 236
Quoted 5 lines
hw/s390x/cpu-topology.c +5
-0+* to POLARITY_VERTICAL_MEDIUM which is the better equivalent when +* machine polarity is set to vertical or POLARITY_VERTICAL_HIGH if +* the vCPU is dedicated. +*/ +if (s390_topology.polarity && !env->entitlement) {It'd be more readable if you compared against enum values by name.
Right, I will change this to
if (s390_topology.polarity != S390_POLARITY_HORIZONTAL && env->entitlement == S390_ENTITLEMENT_UNSET) {Quoted 4 lines
I don't see why you check s390_topology.polarity. If it is horizontal then the value of the entitlement doesn't matter at all, so you can set it to whatever.
Quoted 6 lines
Quoted 2 lines
Right, that is why it is done only for vertical polarization (sorry for the wrong comment)
I'm saying you don't need to check it at all. You adjust the values for vertical polarization, but you could just always do that since the values don't matter at all if the polarization is horizontal.
OK right
Quoted 10 lines
Quoted 10 lines
Quoted 9 lines
All you want to do is enforce dedicated ->
VERTICAL_HIGH, right? So why don't you just add(quoted diff) +4
-0+ if (cpu->env.dedicated && cpu->env.entitlement != POLARITY_VERTICAL_HIGH) { + error_setg(errp, "A dedicated cpu implies high entitlement"); + return; + } >to
s390_topology_check?
Quoted 4 lines
Quoted 1 line
Here it is to set the default in the case the values are not provided.
If no values are provided, they default to dedication=false and entitlement=medium, as defined by patch 1, which are fine and don't need to be adjusted.
Right, I think I added this when working on modifying the attributes in QAPI and rebased it in the wrong patch. Here it has no sense.
Quoted 23 lines
Quoted 12 lines
But where you are right is that I should add a verification to the check function.
Quoted 7 lines
Quoted 6 lines
hw/s390x/cpu-topology.c +5
-0+if (env->dedicated) { + env->entitlement =POLARITY_VERTICAL_HIGH ; + }else { + env->entitlement =POLARITY_VERTICAL_MEDIUM ; + }Quoted 4 lines
Quoted 2 lines
If it is horizontal, then setting the entitlement is pointless as it will be reset to medium on PTF.
That is why the polarity is tested (sorry for the bad comment)
I said this because I'm fine with setting it pointlessly.
Quoted 4 lines
Quoted 3 lines
So the current polarization is vertical and a cpu is being hotplugged, but setting the entitlement of the cpu being added is also pointless, because it's determined by the dedication. That seems weird.
Quoted 5 lines
Quoted 2 lines
No it is not determined by the dedication, if there is no dedication the 3 vertical values are possible.
You set it to either high or medium based on the dedication. And for horizontal polarization it obviously doesn't matter.
on PTF yes but not on hotplug, on hotplug all 3 values are possible.
Quoted 4 lines
As far as I understand you don't need this because the default values are fine. You should add the check that if a dedicated cpu is hotplugged, then the entitlement must be high, to patch 2 and that's it, no additional changes necessary.
Yes, right.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 7/11 target/s390x/cpu topology: activating CPU topology
In reply to Pierre MorelThe KVM capability KVM_CAP_S390_CPU_TOPOLOGY is used to activate the S390_FEAT_CONFIGURATION_TOPOLOGY feature and the topology facility in the host CPU model for the guest in the case the topology is available in QEMU and in KVM.
The feature is disabled by default and fenced for SE (secure execution).
--- hw/s390x/cpu-topology.c | 2 +- target/s390x/cpu_models.c | 1 + target/s390x/kvm/kvm.c | 12 ++++++++++++ 3 files changed, 14 insertions(+), 1 deletion(-)
hw/s390x/cpu-topology.c +1 -1
@@ -55,7 +55,7 @@ int s390_socket_nb(S390CPU *cpu)*/ bool s390_has_topology (void ) { -return false; +return s390_has_feat (S390_FEAT_CONFIGURATION_TOPOLOGY ); }/**
target/s390x/cpu_models.c +1 -0
@@ -254,6 +254,7 @@ bool s390_has_feat(S390Feat feat)case S390_FEAT_SIE_CMMA :case S390_FEAT_SIE_PFMFI :case S390_FEAT_SIE_IBS : +case S390_FEAT_CONFIGURATION_TOPOLOGY :return false;break ;default :
target/s390x/kvm/kvm.c +12 -0
@@ -2470,6 +2470,18 @@ void kvm_s390_get_host_cpu_model(S390CPUModel *model, Error **errp)set_bit (S390_FEAT_UNPACK , model->features); } +/* +* If we have kernel support for CPU Topology indicate the +* configuration-topology facility. +*/ +if (kvm_check_extension (kvm_state,KVM_CAP_S390_CPU_TOPOLOGY )) { +if (kvm_vm_enable_cap (kvm_state,KVM_CAP_S390_CPU_TOPOLOGY ,0 ) <0 ) { +error_setg (errp,"KVM: Error enabling KVM_CAP_S390_CPU_TOPOLOGY" ); +return ; + } +set_bit (S390_FEAT_CONFIGURATION_TOPOLOGY , model->features); + } +/* We emulate a zPCI bus and AEN, therefore we don't need HW support */ set_bit (S390_FEAT_ZPCI , model->features);set_bit (S390_FEAT_ADAPTER_EVENT_NOTIFICATION , model->features);
Signature · 1 line
2.31.1
Re: [PATCH v15 07/11] target/s390x/cpu topology: activating CPU topology
Quoted 49 lines · Pierre Morel
The KVM capability
KVM_CAP_S390_CPU_TOPOLOGYis used to activate theS390_FEAT_CONFIGURATION_TOPOLOGYfeature and the topology facility in the host CPU model for the guest in the case the topology is available in QEMU and in KVM.The feature is disabled by default and fenced for SE (secure execution).
Signed-off-by Pierre Morel <[email protected]>--- hw/s390x/cpu-topology.c | 2 +- target/s390x/cpu_models.c | 1 + target/s390x/kvm/kvm.c | 12 ++++++++++++ 3 files changed, 14 insertions(+), 1 deletion(-)hw/s390x/cpu-topology.c +1
-1@@ -55,7 +55,7 @@ int s390_socket_nb(S390CPU *cpu)*/ bool s390_has_topology (void ) { -return false; +return s390_has_feat (S390_FEAT_CONFIGURATION_TOPOLOGY ); }/** target/s390x/cpu_models.c +1
-0@@ -254,6 +254,7 @@ bool s390_has_feat(S390Feat feat)case S390_FEAT_SIE_CMMA :case S390_FEAT_SIE_PFMFI :case S390_FEAT_SIE_IBS : +case S390_FEAT_CONFIGURATION_TOPOLOGY :return false;break ;default :target/s390x/kvm/kvm.c +1
-0@@ -2470,6 +2470,18 @@ void kvm_s390_get_host_cpu_model(S390CPUModel *model, Error **errp)set_bit (S390_FEAT_UNPACK , model->features); } +/*
target/s390x/kvm/kvm.c void kvm_s390_get_host_cpu_model(S390CPUModel *model, Error **errp) · line 2483
Quoted 10 lines
target/s390x/kvm/kvm.c +10
-0+* If we have kernel support for CPU Topology indicate the +* configuration-topology facility. +*/ +if (kvm_check_extension (kvm_state,KVM_CAP_S390_CPU_TOPOLOGY )) { +if (kvm_vm_enable_cap (kvm_state,KVM_CAP_S390_CPU_TOPOLOGY ,0 ) <0 ) { +error_setg (errp,"KVM: Error enabling KVM_CAP_S390_CPU_TOPOLOGY" ); +return ; + } +set_bit (S390_FEAT_CONFIGURATION_TOPOLOGY , model->features); + }
Not sure, but for the other capabilities, the kvm_vm_enable_cap() is rather done in kvm_arch_init() instead ... likely that it is properly available in case you don't run with the "host" cpu model? So should the kvm_vm_enable_cap(KVM_CAP_S390_CPU_TOPOLOGY) also be moved there (but of course keep the set_bit() here in kvm_s390_get_host_cpu_model())?
Thomas
Re: [PATCH v15 07/11] target/s390x/cpu topology: activating CPU topology
Quoted 55 lines · Thomas Huth
Quoted 54 lines · Pierre Morel
The KVM capability
KVM_CAP_S390_CPU_TOPOLOGYis used to activate theS390_FEAT_CONFIGURATION_TOPOLOGYfeature and the topology facility in the host CPU model for the guest in the case the topology is available in QEMU and in KVM.The feature is disabled by default and fenced for SE (secure execution).
Signed-off-by Pierre Morel <[email protected]>---
hw/s390x/cpu-topology.c | 2 +-target/s390x/cpu_models.c| 1 + target/s390x/kvm/kvm.c | 12 ++++++++++++3 files changed, 14 insertions(+), 1 deletion(-)hw/s390x/cpu-topology.c +1
-1@@ -55,7 +55,7 @@ int s390_socket_nb(S390CPU *cpu)*/ bool s390_has_topology(void) { - return false; + return s390_has_feat(S390_FEAT_CONFIGURATION_TOPOLOGY); } /**target/s390x/cpu_models.c +0
-0@@ -254,6 +254,7 @@ bool s390_has_feat(S390Feat feat)case
S390_FEAT_SIE_CMMA:
caseS390_FEAT_SIE_PFMFI:
caseS390_FEAT_SIE_IBS:
+ caseS390_FEAT_CONFIGURATION_TOPOLOGY:
return false;
break;
default:target/s390x/kvm/kvm.c +6
-0@@ -2470,6 +2470,18 @@ void kvm_s390_get_host_cpu_model(S390CPUModel *model, Error **errp) set_bit(S390_FEAT_UNPACK, model->features); } + /* + * If we have kernel support for CPU Topology indicate the + * configuration-topology facility. + */ + if (kvm_check_extension(kvm_state, KVM_CAP_S390_CPU_TOPOLOGY)) { + if (kvm_vm_enable_cap(kvm_state, KVM_CAP_S390_CPU_TOPOLOGY,0 ) <0 ) {
Quoted 8 lines
target/s390x/kvm/kvm.c void kvm_s390_get_host_cpu_model(S390CPUModel *model, Error **errp) · line 2483
Quoted 6 lines
target/s390x/kvm/kvm.c +5
-0+ error_setg(errp, "KVM: Error enabling KVM_CAP_S390_CPU_TOPOLOGY"); + return; + } + set_bit(S390_FEAT_CONFIGURATION_TOPOLOGY, model->features); + }Not sure, but for the other capabilities, the
kvm_vm_enable_cap()is rather done inkvm_arch_init()instead ... likely that it is properly available in case you don't run with the "host" cpu model? So should thekvm_vm_enable_cap(KVM_CAP_S390_CPU_TOPOLOGY)also be moved there (but of course keep theset_bit()here inkvm_s390_get_host_cpu_model())?
I think you are right, I do this.
Thanks.
regards,
PIerre
Quoted 3 lines
Thomas
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 8/11 qapi/s390x/cpu topology: x-set-cpu-topology monitor command
In reply to Pierre MorelThe modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
--- qapi/machine-target.json | 29 +++++++++++++ include/monitor/hmp.h | 1 + hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++ 4 files changed, 134 insertions(+)
qapi/machine-target.json +29 -0
@@ -342,3 +342,32 @@ 'TARGET_S390X', 'TARGET_MIPS', 'TARGET_LOONGARCH64' ] } } + +## +# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPU +# @polarity: optional polarity, default is last polarity set by the guest +# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +# +# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0> +## +{ 'command': 'x-set-cpu-topology', + 'data': { + 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int', + '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}
include/monitor/hmp.h +1 -0
@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **));void hmp_info_stats (Monitor *mon,const QDict *qdict);void hmp_pcie_aer_inject_error (Monitor *mon,const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict);#endif
hw/s390x/cpu-topology.c +88 -0
97 lines not on this page · raw message
hmp-commands.hx +16 -0
@@ -1815,3 +1815,19 @@ SRST Dump the FDT in dtb format to *filename*. ERST #endif + +#if defined(TARGET_S390X) && defined(CONFIG_KVM) + { + .name = "x-set-cpu-topology", + .args_type = "core:l,socket:l,book:l,drawer:l,polarity:l?,dedicated:b?", + .params = "core socket book drawer [polarity] [dedicated]", + .help = "Move CPU 'core' to 'socket/book/drawer' " + "optionaly modifies polarity and dedication", + .cmd = hmp_x_set_cpu_topology, + }, + +SRST +``x-set-cpu-topology`` *core* *socket* *book* *drawer* *polarity* *dedicated* + Moves the CPU *core* to *socket* *book* *drawer* with *polarity* *dedicated*. +ERST +#endif
Signature · 1 line
2.31.1
Re: [PATCH v15 08/11] qapi/s390x/cpu topology: x-set-cpu-topology monitor command
Quoted 4 lines · Pierre Morel
The modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise
I'm not a native speaker, but "optimize" is more common, I think?
Quoted 27 lines
the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 29 +++++++++++++ include/monitor/hmp.h | 1 + hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++ 4 files changed, 134 insertions(+)qapi/machine-target.json +0
-0@@ -342,3 +342,32 @@ 'TARGET_S390X', 'TARGET_MIPS',
qapi/machine-target.json line 353
Quoted 10 lines
qapi/machine-target.json +9
-0'TARGET_LOONGARCH64' ] } } + +## +# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPU +# @polarity: optional polarity, default is last polarity set by the guest
Hmm, below you do something like that:
if (!has_polarity) {
polarity = POLARITY_VERTICAL_MEDIUM;
}... so it rather seems like medium polarity is the default? ... that looks weird. I think you should maybe rather pass the has_polarity and has_dedication flags to the s390_change_topology() function and only change the value if they have really been provided?
qapi/machine-target.json line 361
Quoted 8 lines
qapi/machine-target.json +8
-0+# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +# +# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0>
Please replace with "Since: 8.0" ... I hope we'll get it merged during this cycle!
Quoted 46 lines
qapi/machine-target.json +11
-0+{ 'command': 'x-set-cpu-topology', + 'data': { + 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int', + '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}include/monitor/hmp.h +1
-0@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **));void hmp_info_stats (Monitor *mon,const QDict *qdict);void hmp_pcie_aer_inject_error (Monitor *mon,const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict);#endif hw/s390x/cpu-topology.c +9
-0@@ -18,6 +18,10 @@#include "target/s390x/cpu.h" #include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" +#include "qapi/qapi-commands-machine-target.h" +#include "qapi/qmp/qdict.h" +#include "monitor/hmp.h" +#include "monitor/monitor.h" /* * s390_topology is used to keep the topology information. @@ -379,3 +383,87 @@ void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp)/* topology tree is reflected in props */ s390_update_cpu_props (ms, cpu); } + +/* +* qmp and hmp implementations +*/ +
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 400
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+static void s390_change_topology (int64_t core_id,int64_t socket_id, +int64_t book_id,int64_t drawer_id, +int64_t polarity,bool dedicated, +Error **errp) +{ +MachineState *ms = current_machine; +S390CPU *cpu; +ERRP_GUARD (); + + cpu = (S390CPU *)ms->possible_cpus->cpus[core_id].cpu;
I think you should add a sanity check for core_id being in a valid range ... otherwise this will cause an access beyond the end of the array if core_id is too big or negative.
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+if (!cpu) { +error_setg (errp,"Core-id %ld does not exist!" , core_id); +return ; + } + +/* Verify the new topology */ +s390_topology_check (cpu, errp); +if (*errp) { +return ; + }
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 420
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+ +/* Move the CPU into its new socket */ +s390_set_core_in_socket (cpu, drawer_id, book_id, socket_id, true, errp); + +/* All checks done, report topology in environment */ + cpu->env.drawer_id = drawer_id; + cpu->env.book_id = book_id; + cpu->env.socket_id = socket_id; + cpu->env.dedicated = dedicated; + cpu->env.entitlement = polarity;
As mentioned above, I think dedicated and polarity should only be changed if they have really been provided?
Quoted 54 lines
hw/s390x/cpu-topology.c +48
-0+/* topology tree is reflected in props */ +s390_update_cpu_props (ms, cpu); + +/* Advertise the topology change */ +s390_cpu_topology_set_modified (); +} + +void qmp_x_set_cpu_topology (int64_t core,int64_t socket, +int64_t book,int64_t drawer, +bool has_polarity,int64_t polarity, +bool has_dedicated,bool dedicated, +Error **errp) +{ +ERRP_GUARD (); + +if (!s390_has_topology ()) { +error_setg (errp,"This machine doesn't support topology" ); +return ; + } +if (!has_polarity) { + polarity =POLARITY_VERTICAL_MEDIUM ; + } +if (!has_dedicated) { + dedicated = false; + } +s390_change_topology (core, socket, book, drawer, polarity, dedicated, errp); +} + +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict) +{ +const int64_t core =qdict_get_int (qdict,"core" ); +const int64_t socket =qdict_get_int (qdict,"socket" ); +const int64_t book =qdict_get_int (qdict,"book" ); +const int64_t drawer =qdict_get_int (qdict,"drawer" ); +bool has_polarity =qdict_haskey (qdict,"polarity" ); +const int64_t polarity =qdict_get_try_int (qdict,"polarity" ,0 ); +bool has_dedicated =qdict_haskey (qdict,"dedicated" ); +const bool dedicated =qdict_get_try_bool (qdict,"dedicated" , false); +Error *local_err =NULL ; + +qmp_x_set_cpu_topology (core, socket, book, drawer, + has_polarity, polarity, + has_dedicated, dedicated, + &local_err); +if (hmp_handle_error (mon, local_err)) { +return ; + } +}hmp-commands.hx +0
-0@@ -1815,3 +1815,19 @@ SRST Dump the FDT in dtb format to *filename*.
hmp-commands.hx SRST · line 1825
Quoted 10 lines
hmp-commands.hx +8
-0ERST #endif + +#if defined(TARGET_S390X) && defined(CONFIG_KVM) + { + .name = "x-set-cpu-topology", + .args_type = "core:l,socket:l,book:l,drawer:l,polarity:l?,dedicated:b?", + .params = "core socket book drawer [polarity] [dedicated]", + .help = "Move CPU 'core' to 'socket/book/drawer' " + "optionaly modifies polarity and dedication",
optionaly ==> optionally
hmp-commands.hx SRST · line 1833
Quoted 8 lines
hmp-commands.hx +8
-0+ .cmd = hmp_x_set_cpu_topology, + }, + +SRST +``x-set-cpu-topology`` *core* *socket* *book* *drawer* *polarity* *dedicated* + Moves the CPU *core* to *socket* *book* *drawer* with *polarity* *dedicated*. +ERST +#endif
Thomas
Re: [PATCH v15 08/11] qapi/s390x/cpu topology: x-set-cpu-topology monitor command
Quoted 7 lines · Thomas Huth
Quoted 4 lines · Pierre Morel
The modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise
I'm not a native speaker, but "optimize" is more common, I think?
Yes. I will stick on it.
Quoted 42 lines
Quoted 38 lines
the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
Signed-off-by Pierre Morel <[email protected]>---
qapi/machine-target.json| 29 +++++++++++++ include/monitor/hmp.h | 1 +
hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++4 files changed, 134 insertions(+)qapi/machine-target.json +9
-0@@ -342,3 +342,32 @@ 'TARGET_S390X', 'TARGET_MIPS', 'TARGET_LOONGARCH64' ] } } + +## +# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPU +# @polarity: optional polarity, default is last polarity set by theQuoted 1 line
guest
Hmm, below you do something like that:
Quoted 4 lines
if (!has_polarity) {
polarity =POLARITY_VERTICAL_MEDIUM;
} >
... so it rather seems like medium polarity is the default? ... that looks weird. I think you should maybe rather pass thehas_polarityandhas_dedicationflags to thes390_change_topology()function and only change the value if they have really been provided?
OK, I will change this incoherence.
Quoted 11 lines
qapi/machine-target.json line 361
Quoted 9 lines
qapi/machine-target.json +8
-0+# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +# +# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0>Please replace with "Since: 8.0" ... I hope we'll get it merged during this cycle!
OK. :)
Quoted 50 lines
Quoted 50 lines
qapi/machine-target.json +11
-0+{ 'command': 'x-set-cpu-topology', + 'data': { + 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int', + '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}include/monitor/hmp.h +1
-0@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **)); void hmp_info_stats(Monitor *mon, const QDict *qdict); void hmp_pcie_aer_inject_error(Monitor *mon, const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict); #endifhw/s390x/cpu-topology.c +4
-0@@ -18,6 +18,10 @@ #include "target/s390x/cpu.h" #include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" +#include "qapi/qapi-commands-machine-target.h" +#include "qapi/qmp/qdict.h" +#include "monitor/hmp.h" +#include "monitor/monitor.h" /** s390_topology is used to keep the topology information. @@ -379,3 +383,87 @@ void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) /* topology tree is reflected in props */ s390_update_cpu_props(ms, cpu);}
hw/s390x/cpu-topology.c +9
-0+ +/* +* qmp and hmp implementations +*/ + +static void s390_change_topology (int64_t core_id,int64_t socket_id, + int64_t book_id, int64_t drawer_id, + int64_t polarity, bool dedicated, + Error **errp)
Quoted 8 lines
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 400
Quoted 6 lines
hw/s390x/cpu-topology.c +6
-0+{ + MachineState *ms = current_machine; + S390CPU *cpu; + ERRP_GUARD(); + + cpu = (S390CPU *)ms->possible_cpus->cpus[core_id].cpu;I think you should add a sanity check for
core_idbeing in a valid range ... otherwise this will cause an access beyond the end of the array ifcore_idis too big or negative.
Right, I will also change the int64_t for the declaration of the core, socket etc to uint16_t so avoiding a check for negative values.
Quoted 16 lines
Quoted 16 lines
+ if (!cpu) {
+error_setg(errp, "Core-id %ld does not exist!", core_id); + return;
+ }hw/s390x/cpu-topology.c +9
-0+ + /* Verify the new topology */ + s390_topology_check(cpu, errp); + if (*errp) { + return; + } + + /* Move the CPU into its new socket */ + s390_set_core_in_socket(cpu, drawer_id, book_id, socket_id, true,errp);
+
Quoted 8 lines
Quoted 6 lines
+ /* All checks done, report topology in environment */ +
cpu->env.drawer_id=drawer_id;
+cpu->env.book_id=book_id;
+cpu->env.socket_id=socket_id;
+cpu->env.dedicated= dedicated;
+cpu->env.entitlement= polarity;As mentioned above, I think dedicated and polarity should only be changed if they have really been provided?
right, I do this.
Quoted 62 lines
Quoted 62 lines
+ /* topology tree is reflected in props */
+s390_update_cpu_props(ms, cpu);hw/s390x/cpu-topology.c +24
-0+ + /* Advertise the topology change */ + s390_cpu_topology_set_modified(); +} + +void qmp_x_set_cpu_topology (int64_t core,int64_t socket, + int64_t book, int64_t drawer, + bool has_polarity, int64_t polarity, + bool has_dedicated, bool dedicated, + Error **errp) +{ + ERRP_GUARD(); + + if (!s390_has_topology()) { + error_setg(errp, "This machine doesn't support topology"); + return; + } + if (!has_polarity) { + polarity = POLARITY_VERTICAL_MEDIUM; + } + if (!has_dedicated) { + dedicated = false; + } + s390_change_topology(core, socket, book, drawer, polarity,dedicated, errp);
hw/s390x/cpu-topology.c +12
-0+} + +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict) +{ + const int64_t core = qdict_get_int(qdict, "core"); + const int64_t socket = qdict_get_int(qdict, "socket"); + const int64_t book = qdict_get_int(qdict, "book"); + const int64_t drawer = qdict_get_int(qdict, "drawer"); + bool has_polarity = qdict_haskey(qdict, "polarity"); + const int64_t polarity = qdict_get_try_int(qdict, "polarity", 0); + bool has_dedicated = qdict_haskey(qdict, "dedicated"); + const bool dedicated = qdict_get_try_bool(qdict, "dedicated",false);
+ Error*local_err= NULL;hw/s390x/cpu-topology.c +9
-0+ + qmp_x_set_cpu_topology(core, socket, book, drawer, + has_polarity, polarity, + has_dedicated, dedicated, + &local_err); + if (hmp_handle_error(mon, local_err)) { + return; + } +}hmp-commands.hx +0
-0@@ -1815,3 +1815,19 @@ SRSTDump the FDT in dtb format to *filename*.
ERST
#endifhmp-commands.hx +3
-0+ +#if defined(TARGET_S390X) && defined(CONFIG_KVM) + {
Quoted 8 lines
Quoted 6 lines
+ .name = "x-set-cpu-topology",
+ .args_type = "core:l,socket:l,book:l,drawer:l,polarity:l?,dedicated:b?", + .params = "core socket book drawer [polarity] [dedicated]", + .help = "Move CPU 'core' to 'socket/book/drawer' " + "optionaly modifies polarity and dedication",optionaly ==> optionally
OK
Quoted 14 lines
hmp-commands.hx SRST · line 1833
Quoted 11 lines
+ .cmd =
hmp_x_set_cpu_topology,
+ },hmp-commands.hx +3
-0+ +SRST +``x-set-cpu-topology`` *core* *socket* *book* *drawer* *polarity**dedicated*
+ Moves the CPU *core* to *socket* *book* *drawer* with *polarity* *dedicated*.hmp-commands.hx +2
-0+ERST +#endifThomas
Thanks,
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 08/11] qapi/s390x/cpu topology: x-set-cpu-topology monitor command
In reply to Pierre MorelQuoted 102 lines · Pierre Morel
The modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 29 +++++++++++++ include/monitor/hmp.h | 1 + hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++ 4 files changed, 134 insertions(+)qapi/machine-target.json +29
-0@@ -342,3 +342,32 @@ 'TARGET_S390X', 'TARGET_MIPS', 'TARGET_LOONGARCH64' ] } } + +## +# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPU +# @polarity: optional polarity, default is last polarity set by the guest +# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +# +# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0> +## +{ 'command': 'x-set-cpu-topology', + 'data': { + 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int', + '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}include/monitor/hmp.h +1
-0@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **));void hmp_info_stats (Monitor *mon,const QDict *qdict);void hmp_pcie_aer_inject_error (Monitor *mon,const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict);#endif hw/s390x/cpu-topology.c +15
-0@@ -18,6 +18,10 @@#include "target/s390x/cpu.h" #include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" +#include "qapi/qapi-commands-machine-target.h" +#include "qapi/qmp/qdict.h" +#include "monitor/hmp.h" +#include "monitor/monitor.h" /* * s390_topology is used to keep the topology information. @@ -379,3 +383,87 @@ void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp)/* topology tree is reflected in props */ s390_update_cpu_props (ms, cpu); } + +/* +* qmp and hmp implementations +*/ + +static void s390_change_topology (int64_t core_id,int64_t socket_id, +int64_t book_id,int64_t drawer_id, +int64_t polarity,bool dedicated, +Error **errp) +{ +MachineState *ms = current_machine;
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 406
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+S390CPU *cpu; +ERRP_GUARD (); + + cpu = (S390CPU *)ms->possible_cpus->cpus[core_id].cpu; +if (!cpu) { +error_setg (errp,"Core-id %ld does not exist!" , core_id); +return ; + } + +/* Verify the new topology */
!! Should verify the new topology but ...
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 407
Quoted 1 line
hw/s390x/cpu-topology.c +1
-0+s390_topology_check (cpu, errp);
... verifies the old one => I will have to change this.
Sorry, this will impact the patch 2 because I think I should modify the s390_topology_check() function to take as arguments the individual attributes.
Like:
static void s390_topology_check( uint16_t core_id, uint16_t socket_id,
uint16_t book_id, uint16_t drawer_id, uint16_t entitlement, bool dedicated, Error **errp)
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 08/11] qapi/s390x/cpu topology: x-set-cpu-topology monitor command
In reply to Pierre MorelQuoted 30 lines · Pierre Morel
The modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 29 +++++++++++++ include/monitor/hmp.h | 1 + hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++ 4 files changed, 134 insertions(+)qapi/machine-target.json +0
-0@@ -342,3 +342,32 @@ 'TARGET_S390X',
qapi/machine-target.json line 352
Quoted 10 lines
qapi/machine-target.json +8
-0'TARGET_MIPS', 'TARGET_LOONGARCH64' ] } } + +## +# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPU
I wonder if it wouldn't be more convenient for the caller if everything is optional.
Quoted 6 lines
qapi/machine-target.json +6
-0+# @polarity: optional polarity, default is last polarity set by the guest +# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +#
qapi/machine-target.json line 368
Quoted 10 lines
qapi/machine-target.json +10
-0+# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0> +## +{ 'command': 'x-set-cpu-topology', + 'data': { + 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int',
Did you consider naming those core-id, etc.? It would be consistent with query-cpus-fast/CpuInstanceProperties. Also all your variables end with _id. I don't care really just wanted to point it out.
qapi/machine-target.json line 373
Quoted 5 lines
qapi/machine-target.json +5
-0+ '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}
So apparently this is the old way of doing an experimental api.
Quoted 55 lines
Names beginning with ``x-`` used to signify "experimental". This convention has been replaced by special feature "unstable".
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
include/monitor/hmp.h +1
-0@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **));void hmp_info_stats (Monitor *mon,const QDict *qdict);void hmp_pcie_aer_inject_error (Monitor *mon,const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict);#endif hw/s390x/cpu-topology.c +22
-0@@ -18,6 +18,10 @@#include "target/s390x/cpu.h" #include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" +#include "qapi/qapi-commands-machine-target.h" +#include "qapi/qmp/qdict.h" +#include "monitor/hmp.h" +#include "monitor/monitor.h" /* * s390_topology is used to keep the topology information. @@ -379,3 +383,87 @@ void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp)/* topology tree is reflected in props */ s390_update_cpu_props (ms, cpu); } + +/* +* qmp and hmp implementations +*/ + +static void s390_change_topology (int64_t core_id,int64_t socket_id, +int64_t book_id,int64_t drawer_id, +int64_t polarity,bool dedicated, +Error **errp) +{ +MachineState *ms = current_machine; +S390CPU *cpu; +ERRP_GUARD (); + + cpu = (S390CPU *)ms->possible_cpus->cpus[core_id].cpu; +if (!cpu) { +error_setg (errp,"Core-id %ld does not exist!" , core_id); +return ;
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 413
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+ } + +/* Verify the new topology */ +s390_topology_check (cpu, errp); +if (*errp) { +return ; + } + +/* Move the CPU into its new socket */ +s390_set_core_in_socket (cpu, drawer_id, book_id, socket_id, true, errp);
The cpu isn't being created, so that should be false instead of true, right?
Quoted 45 lines
hw/s390x/cpu-topology.c +45
-0+ +/* All checks done, report topology in environment */ + cpu->env.drawer_id = drawer_id; + cpu->env.book_id = book_id; + cpu->env.socket_id = socket_id; + cpu->env.dedicated = dedicated; + cpu->env.entitlement = polarity; + +/* topology tree is reflected in props */ +s390_update_cpu_props (ms, cpu); + +/* Advertise the topology change */ +s390_cpu_topology_set_modified (); +} + +void qmp_x_set_cpu_topology (int64_t core,int64_t socket, +int64_t book,int64_t drawer, +bool has_polarity,int64_t polarity, +bool has_dedicated,bool dedicated, +Error **errp) +{ +ERRP_GUARD (); + +if (!s390_has_topology ()) { +error_setg (errp,"This machine doesn't support topology" ); +return ; + } +if (!has_polarity) { + polarity =POLARITY_VERTICAL_MEDIUM ; + } +if (!has_dedicated) { + dedicated = false; + } +s390_change_topology (core, socket, book, drawer, polarity, dedicated, errp); +} + +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict) +{ +const int64_t core =qdict_get_int (qdict,"core" ); +const int64_t socket =qdict_get_int (qdict,"socket" ); +const int64_t book =qdict_get_int (qdict,"book" ); +const int64_t drawer =qdict_get_int (qdict,"drawer" ); +bool has_polarity =qdict_haskey (qdict,"polarity" ); +const int64_t polarity =qdict_get_try_int (qdict,"polarity" ,0 ); +bool has_dedicated =qdict_haskey (qdict,"dedicated" );
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 468
Quoted 10 lines
hw/s390x/cpu-topology.c +10
-0+const bool dedicated =qdict_get_try_bool (qdict,"dedicated" , false); +Error *local_err =NULL ; + +qmp_x_set_cpu_topology (core, socket, book, drawer, + has_polarity, polarity, + has_dedicated, dedicated, + &local_err); +if (hmp_handle_error (mon, local_err)) { +return ; + }
What is the if for? The function ends anyway.
Quoted 25 lines
hw/s390x/cpu-topology.c +1
-0+}hmp-commands.hx +16
-0@@ -1815,3 +1815,19 @@ SRST Dump the FDT in dtb format to *filename*. ERST #endif + +#if defined(TARGET_S390X) && defined(CONFIG_KVM) + { + .name = "x-set-cpu-topology", + .args_type = "core:l,socket:l,book:l,drawer:l,polarity:l?,dedicated:b?", + .params = "core socket book drawer [polarity] [dedicated]", + .help = "Move CPU 'core' to 'socket/book/drawer' " + "optionaly modifies polarity and dedication", + .cmd = hmp_x_set_cpu_topology, + }, + +SRST +``x-set-cpu-topology`` *core* *socket* *book* *drawer* *polarity* *dedicated* + Moves the CPU *core* to *socket* *book* *drawer* with *polarity* *dedicated*. +ERST +#endif
Re: [PATCH v15 08/11] qapi/s390x/cpu topology: x-set-cpu-topology monitor command
Quoted 35 lines · Nina Schoetterl-Glausch
Quoted 34 lines · Pierre Morel
The modification of the CPU attributes are done through a monitor command.
It allows to move the core inside the topology tree to optimise the cache usage in the case the host's hypervisor previously moved the CPU.
The same command allows to modify the CPU attributes modifiers like polarization entitlement and the dedicated attribute to notify the guest if the host admin modified scheduling or dedication of a vCPU.
With this knowledge the guest has the possibility to optimize the usage of the vCPUs.
The command is made experimental for the moment.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 29 +++++++++++++ include/monitor/hmp.h | 1 + hw/s390x/cpu-topology.c | 88 ++++++++++++++++++++++++++++++++++++++++ hmp-commands.hx | 16 ++++++++ 4 files changed, 134 insertions(+)qapi/machine-target.json +2
-0@@ -342,3 +342,32 @@ 'TARGET_S390X', 'TARGET_MIPS', 'TARGET_LOONGARCH64' ] } } + +##
Quoted 8 lines
qapi/machine-target.json line 352
Quoted 6 lines
qapi/machine-target.json +6
-0+# @x-set-cpu-topology: +# +# @core: the vCPU ID to be moved +# @socket: the destination socket where to move the vCPU +# @book: the destination book where to move the vCPU +# @drawer: the destination drawer where to move the vCPUI wonder if it wouldn't be more convenient for the caller if everything is optional.
Yes, it is a good point.
Quoted 13 lines
Quoted 13 lines
qapi/machine-target.json +12
-0+# @polarity: optional polarity, default is last polarity set by the guest +# @dedicated: optional, if the vCPU is dedicated to a real CPU +# +# Modifies the topology by moving the CPU inside the topology +# tree or by changing a modifier attribute of a CPU. +# +# Returns: Nothing on success, the reason on failure. +# +# Since: <next qemu stable release, eg. 1.0> +## +{ 'command': 'x-set-cpu-topology', + 'data': {
Quoted 8 lines
qapi/machine-target.json line 368
Quoted 4 lines
qapi/machine-target.json +4
-0+ 'core': 'int', + 'socket': 'int', + 'book': 'int', + 'drawer': 'int',Did you consider naming those core-id, etc.? It would be consistent with query-cpus-fast/CpuInstanceProperties. Also all your variables end with
_id. I don't care really just wanted to point it out.
OK, core-id etc. looks better
Quoted 15 lines
qapi/machine-target.json line 373
Quoted 6 lines
qapi/machine-target.json +5
-0+ '*polarity': 'int', + '*dedicated': 'bool' + }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } +}So apparently this is the old way of doing an experimental api.
Quoted 6 lines
Names beginning with ``x-`` used to signify "experimental". This convention has been replaced by special feature "unstable".
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
OK
Quoted 53 lines
Quoted 53 lines
include/monitor/hmp.h +1
-0@@ -151,5 +151,6 @@ void hmp_human_readable_text_helper(Monitor *mon, HumanReadableText *(*qmp_handler)(Error **));void hmp_info_stats (Monitor *mon,const QDict *qdict);void hmp_pcie_aer_inject_error (Monitor *mon,const QDict *qdict); +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict);#endif hw/s390x/cpu-topology.c +26
-0@@ -18,6 +18,10 @@#include "target/s390x/cpu.h" #include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" +#include "qapi/qapi-commands-machine-target.h" +#include "qapi/qmp/qdict.h" +#include "monitor/hmp.h" +#include "monitor/monitor.h" /* * s390_topology is used to keep the topology information. @@ -379,3 +383,87 @@ void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp)/* topology tree is reflected in props */ s390_update_cpu_props (ms, cpu); } + +/* +* qmp and hmp implementations +*/ + +static void s390_change_topology (int64_t core_id,int64_t socket_id, +int64_t book_id,int64_t drawer_id, +int64_t polarity,bool dedicated, +Error **errp) +{ +MachineState *ms = current_machine; +S390CPU *cpu; +ERRP_GUARD (); + + cpu = (S390CPU *)ms->possible_cpus->cpus[core_id].cpu; +if (!cpu) { +error_setg (errp,"Core-id %ld does not exist!" , core_id); +return ; + } + +/* Verify the new topology */ +s390_topology_check (cpu, errp);
Quoted 8 lines
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 413
Quoted 6 lines
hw/s390x/cpu-topology.c +6
-0+if (*errp) { +return ; + } + +/* Move the CPU into its new socket */ +s390_set_core_in_socket (cpu, drawer_id, book_id, socket_id, true, errp);The cpu isn't being created, so that should be false instead of true, right?
Yes right, should be "false"
Quoted 50 lines
Quoted 50 lines
hw/s390x/cpu-topology.c +49
-0+ +/* All checks done, report topology in environment */ + cpu->env.drawer_id = drawer_id; + cpu->env.book_id = book_id; + cpu->env.socket_id = socket_id; + cpu->env.dedicated = dedicated; + cpu->env.entitlement = polarity; + +/* topology tree is reflected in props */ +s390_update_cpu_props (ms, cpu); + +/* Advertise the topology change */ +s390_cpu_topology_set_modified (); +} + +void qmp_x_set_cpu_topology (int64_t core,int64_t socket, +int64_t book,int64_t drawer, +bool has_polarity,int64_t polarity, +bool has_dedicated,bool dedicated, +Error **errp) +{ +ERRP_GUARD (); + +if (!s390_has_topology ()) { +error_setg (errp,"This machine doesn't support topology" ); +return ; + } +if (!has_polarity) { + polarity =POLARITY_VERTICAL_MEDIUM ; + } +if (!has_dedicated) { + dedicated = false; + } +s390_change_topology (core, socket, book, drawer, polarity, dedicated, errp); +} + +void hmp_x_set_cpu_topology (Monitor *mon,const QDict *qdict) +{ +const int64_t core =qdict_get_int (qdict,"core" ); +const int64_t socket =qdict_get_int (qdict,"socket" ); +const int64_t book =qdict_get_int (qdict,"book" ); +const int64_t drawer =qdict_get_int (qdict,"drawer" ); +bool has_polarity =qdict_haskey (qdict,"polarity" ); +const int64_t polarity =qdict_get_try_int (qdict,"polarity" ,0 ); +bool has_dedicated =qdict_haskey (qdict,"dedicated" ); +const bool dedicated =qdict_get_try_bool (qdict,"dedicated" , false); +Error *local_err =NULL ; + + qmp_x_set_cpu_topology(core, socket, book, drawer,
Quoted 8 lines
hw/s390x/cpu-topology.c void s390_topology_set_cpu(MachineState *ms, S390CPU *cpu, Error **errp) · line 468
Quoted 6 lines
hw/s390x/cpu-topology.c +6
-0+ has_polarity, polarity, + has_dedicated, dedicated, + &local_err); +if (hmp_handle_error (mon, local_err)) { +return ; + }What is the if for? The function ends anyway.
Right, I take it away.
Thanks.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 9/11 machine: adding s390 topology to query-cpu-fast
In reply to Pierre MorelS390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)
qapi/machine.json +10 -3
@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int' + } +} ## # @CpuInfoFast:
@@ -70,7 +77,7 @@ # # @thread-id: ID of the underlying host thread # -# @props: properties describing to which node/socket/core/thread +# @props: properties describing to which node/drawer/book/socket/core/thread # virtual CPU belongs to, provided if supported by board # # @target: the QEMU system emulation target, which determines which
hw/core/machine-qmp-cmds.c +2 -0
@@ -30,6 +30,8 @@ static void cpustate_to_cpuinfo_s390(CpuInfoS390 *info, const CPUState *cpu)CPUS390XState *env = &s390_cpu->env; info->cpu_state = env->cpu_state; + info->dedicated = env->dedicated; + info->polarity = env->entitlement;#else abort ();#endif
Signature · 1 line
2.31.1
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 2 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
books and drawers are already handled via the entries in CpuInstanceProperties, so this sentence looks like a wrong leftover now?
I'd suggest talking about "dedication" and "polarity" instead?
Quoted 34 lines
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +9
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int' + } +} ## # @CpuInfoFast:
qapi/machine.json line 80
Quoted 5 lines
qapi/machine.json +1
-1@@ -70,7 +77,7 @@ # # @thread-id: ID of the underlying host thread # -# @props: properties describing to which node/socket/core/thread +# @props: properties describing to which node/drawer/book/socket/core/thread
I think this hunk should rather be moved to patch 1 now.
Thomas
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 5 lines · Thomas Huth
Quoted 2 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
books and drawers are already handled via the entries in CpuInstanceProperties, so this sentence looks like a wrong leftover now?
yes it is
Quoted 2 lines
I'd suggest talking about "dedication" and "polarity" instead?
OK.
Quoted 34 lines
Quoted 34 lines
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>---
qapi/machine.json | 13 ++++++++++---
hw/core/machine-qmp-cmds.c| 2 ++2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +9
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int' + } +} ## # @CpuInfoFast:@@ -70,7 +77,7 @@
Quoted 8 lines
Quoted 6 lines
qapi/machine.json +1
-1# # @thread-id: ID of the underlying host thread # -# @props: properties describing to which node/socket/core/thread +# @props: properties describing to whichnode/drawer/book/socket/core/thread
I think this hunk should rather be moved to patch 1 now.
yes it should.
Thanks.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
In reply to Pierre MorelQuoted 21 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +1
-1@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12)
qapi/machine.json line 65
Quoted 10 lines
qapi/machine.json +6
-1+# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'
I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
Thomas
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 28 lines · Thomas Huth
Quoted 27 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +3
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } }
Quoted 8 lines
qapi/machine.json line 65
Quoted 4 lines
qapi/machine.json +4
-0+{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
FWIW, I would say that the general expectation of QMP clients is that they must *always* expect new fields to appear in dicts that are returned in QMP replies. We add new fields at will on a frequent basis.
So personally I'd keep life simple and unconditionally report the new fields.
With regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 38 lines · Daniel P. Berrangé
Quoted 28 lines · Thomas Huth
Quoted 27 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +3
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } }Quoted 8 lines
qapi/machine.json line 65
Quoted 4 lines
qapi/machine.json +4
-0+{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
Quoted 3 lines
FWIW, I would say that the general expectation of QMP clients is that they must *always* expect new fields to appear in dicts that are returned in QMP replies. We add new fields at will on a frequent basis.
Did we change our policy here? I slightly remember I've been told differently in the past ... but I can't recall where this was, it's certainly been a while.
So a question to the QAPI maintainers: What's the preferred handling for new fields nowadays in such situations?
Thomas
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 43 lines · Thomas Huth
Quoted 38 lines · Daniel P. Berrangé
Quoted 28 lines · Thomas Huth
Quoted 27 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +3
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } }Quoted 8 lines
qapi/machine.json line 65
Quoted 4 lines
qapi/machine.json +4
-0+{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
Quoted 3 lines
FWIW, I would say that the general expectation of QMP clients is that they must *always* expect new fields to appear in dicts that are returned in QMP replies. We add new fields at will on a frequent basis.
Quoted 6 lines
Did we change our policy here? I slightly remember I've been told differently in the past ... but I can't recall where this was, it's certainly been a while.
So a question to the QAPI maintainers: What's the preferred handling for new fields nowadays in such situations?
I think you're mixing it up with policy for adding new fields to *input* parameters. You can't add a new mandatory field to input params of a command because no existing client will be sending that. Only optional params are viable, without a deprecation cycle. For output parameters such as this case, there's no compatibilty problem.
With regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 48 lines · Daniel P. Berrangé
Quoted 47 lines · Thomas Huth
Quoted 38 lines · Daniel P. Berrangé
Quoted 28 lines · Thomas Huth
Quoted 27 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +3
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } }Quoted 8 lines
qapi/machine.json line 65
Quoted 4 lines
qapi/machine.json +4
-0+{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
Quoted 3 lines
FWIW, I would say that the general expectation of QMP clients is that they must *always* expect new fields to appear in dicts that are returned in QMP replies. We add new fields at will on a frequent basis.
Did we change our policy here? I slightly remember I've been told differently in the past ... but I can't recall where this was, it's certainly been a while.
Quoted 8 lines
Quoted 2 lines
So a question to the QAPI maintainers: What's the preferred handling for new fields nowadays in such situations?
I think you're mixing it up with policy for adding new fields to *input* parameters. You can't add a new mandatory field to input params of a command because no existing client will be sending that. Only optional params are viable, without a deprecation cycle. For output parameters such as this case, there's no compatibilty problem.
Ah, right, I likely mixed it up with that. Thanks for the clarification!
Thomas
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
In reply to Daniel P. BerrangéQuoted 42 lines · Daniel P. Berrangé
Quoted 28 lines · Thomas Huth
Quoted 27 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +3
-2@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } }Quoted 8 lines
qapi/machine.json line 65
Quoted 4 lines
qapi/machine.json +4
-0+{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int'I think it would also be better to mark the new fields as optional and only return them if the guest has the topology enabled, to avoid confusing clients that were written before this change.
FWIW, I would say that the general expectation of QMP clients is that they must *always* expect new fields to appear in dicts that are returned in QMP replies. We add new fields at will on a frequent basis.
Quoted 5 lines
So personally I'd keep life simple and unconditionally report the new fields.
With regards,
Daniel
OK, thanks both of you.
I will then keep the simple way.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
In reply to Pierre MorelQuoted 50 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +10
-3@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int' + } +} ## # @CpuInfoFast:@@ -70,7 +77,7 @@ # # @thread-id: ID of the underlying host thread # -# @props: properties describing to which node/socket/core/thread +# @props: properties describing to which node/drawer/book/socket/core/thread # virtual CPU belongs to, provided if supported by board # # @target: the QEMU system emulation target, which determines whichhw/core/machine-qmp-cmds.c +0
-0
hw/core/machine-qmp-cmds.c static void cpustate_to_cpuinfo_s390(CpuInfoS390 *info, const CPUState *cpu) · line 34
Quoted 5 lines
hw/core/machine-qmp-cmds.c +2
-0@@ -30,6 +30,8 @@ static void cpustate_to_cpuinfo_s390(CpuInfoS390 *info, const CPUState *cpu)CPUS390XState *env = &s390_cpu->env; info->cpu_state = env->cpu_state; + info->dedicated = env->dedicated; + info->polarity = env->entitlement;
Might want to do s/polarity/entitlement on the whole patch to make this more coherent.
Reviewed-by: Nina Schoetterl-Glausch if you fix the issues in Thomas' first reply.
Quoted 3 lines
hw/core/machine-qmp-cmds.c +0
-0#else abort ();#endif
Re: [PATCH v15 09/11] machine: adding s390 topology to query-cpu-fast
Quoted 53 lines · Nina Schoetterl-Glausch
Quoted 52 lines · Pierre Morel
S390x provides two more topology containers above the sockets, books and drawers.
Let's add these CPU attributes to the QAPI command query-cpu-fast.
Signed-off-by Pierre Morel <[email protected]>--- qapi/machine.json | 13 ++++++++++--- hw/core/machine-qmp-cmds.c | 2 ++ 2 files changed, 12 insertions(+), 3 deletions(-)qapi/machine.json +10
-3@@ -53,11 +53,18 @@ # # Additional information about a virtual S390 CPU # -# @cpu-state: the virtual CPU's state +# @cpu-state: the virtual CPU's state (since 2.12) +# @dedicated: the virtual CPU's dedication (since 8.0) +# @polarity: the virtual CPU's polarity (since 8.0) # # Since: 2.12 ## -{ 'struct': 'CpuInfoS390', 'data': { 'cpu-state': 'CpuS390State' } } +{ 'struct': 'CpuInfoS390', + 'data': { 'cpu-state': 'CpuS390State', + 'dedicated': 'bool', + 'polarity': 'int' + } +} ## # @CpuInfoFast:@@ -70,7 +77,7 @@ # # @thread-id: ID of the underlying host thread # -# @props: properties describing to which node/socket/core/thread +# @props: properties describing to which node/drawer/book/socket/core/thread # virtual CPU belongs to, provided if supported by board # # @target: the QEMU system emulation target, which determines whichhw/core/machine-qmp-cmds.c +0
-0@@ -30,6 +30,8 @@ static void cpustate_to_cpuinfo_s390(CpuInfoS390 *info, const CPUState *cpu)CPUS390XState *env = &s390_cpu->env;
Quoted 7 lines
hw/core/machine-qmp-cmds.c static void cpustate_to_cpuinfo_s390(CpuInfoS390 *info, const CPUState *cpu) · line 34
Quoted 3 lines
hw/core/machine-qmp-cmds.c +2
-0info->cpu_state = env->cpu_state; + info->dedicated = env->dedicated; + info->polarity = env->entitlement;Might want to do s/polarity/entitlement on the whole patch to make this more coherent.
Reviewed-by: Nina Schoetterl-Glausch if you fix the issues in Thomas' first reply.
OK, substitution done and Thomas changes already done. Thanks.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
PATCH v15 10/11 qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
In reply to Pierre MorelWhen the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.
--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)
qapi/machine-target.json +30 -0
@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +# +# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +# +## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] } +}
hw/s390x/cpu-topology.c +2 -0
@@ -19,6 +19,7 @@#include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" #include "qapi/qapi-commands-machine-target.h" +#include "qapi/qapi-events-machine-target.h" #include "qapi/qmp/qdict.h" #include "monitor/hmp.h" #include "monitor/monitor.h"
@@ -163,6 +164,7 @@ void s390_handle_ptf(S390CPU *cpu, uint8_t r1, uintptr_t ra) s390_topology.polarity = fc;s390_cpu_topology_set_modified ();s390_topology_set_cpus_polarity (fc); +qapi_event_send_cpu_polarity_change (fc);setcc (cpu,0 ); }break ;
Signature · 1 line
2.31.1
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 38 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +19
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +#
qapi/machine-target.json line 402
Quoted 10 lines
qapi/machine-target.json +10
-0+# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +# +## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }
I wonder if you should depend on CONFIG_KVM or not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
Patch looks good.
Quoted 21 lines
qapi/machine-target.json +1
-0+}hw/s390x/cpu-topology.c +2
-0@@ -19,6 +19,7 @@#include "hw/s390x/s390-virtio-ccw.h" #include "hw/s390x/cpu-topology.h" #include "qapi/qapi-commands-machine-target.h" +#include "qapi/qapi-events-machine-target.h" #include "qapi/qmp/qdict.h" #include "monitor/hmp.h" #include "monitor/monitor.h" @@ -163,6 +164,7 @@ void s390_handle_ptf(S390CPU *cpu, uint8_t r1, uintptr_t ra) s390_topology.polarity = fc;s390_cpu_topology_set_modified ();s390_topology_set_cpus_polarity (fc); +qapi_event_send_cpu_polarity_change (fc);setcc (cpu,0 ); }break ;
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 45 lines · Nina Schoetterl-Glausch
Quoted 44 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +25
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +# +# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +#
Quoted 8 lines
qapi/machine-target.json line 402
Quoted 4 lines
qapi/machine-target.json +4
-0+## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }I wonder if you should depend on
CONFIG_KVMor not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.
TCG vs. KVM should be as transparent as we can make it.
If only KVM can get into the state where the event is emitted, say because the state is only possible with features only KVM supports, then making the event conditional on KVM makes sense. Of course, when another accelerator acquires these features, we need to emit the event there as well, which will involve adjusting the condition.
Quoted 3 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
docs/devel/qapi-code-gen.rst:
Special features
~~~~~~~~~~~~~~~~
Feature "deprecated" marks a command, event, enum value, or struct member as deprecated. It is not supported elsewhere so far. Interfaces so marked may be withdrawn in future releases in accordance with QEMU's deprecation policy.
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
See also -compat parameters unstable-input, unstable-output, both intended for "testing the future".
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
Events are part of the interface just like command returns are. Not emitting an event in a situation where it was emitted before can easily break things. Only when the situation is no longer possible, the event can be removed safely.
Questions?
[...]
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 56 lines · Markus Armbruster
Quoted 45 lines · Nina Schoetterl-Glausch
Quoted 44 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +25
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +# +# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +#Quoted 8 lines
qapi/machine-target.json line 402
Quoted 4 lines
qapi/machine-target.json +4
-0+## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }I wonder if you should depend on
CONFIG_KVMor not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.
Quoted 7 lines
TCG vs. KVM should be as transparent as we can make it.
If only KVM can get into the state where the event is emitted, say because the state is only possible with features only KVM supports, then making the event conditional on KVM makes sense. Of course, when another accelerator acquires these features, we need to emit the event there as well, which will involve adjusting the condition.
That's the case here, KVM supports the feature, TCG doesn't, although there is no reason it couldn't in the future.
Quoted 10 lines
Quoted 4 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
docs/devel/qapi-code-gen.rst:Special features
~~~~~~~~~~~~~~~~
Quoted 8 lines
Feature "deprecated" marks a command, event, enum value, or struct member as deprecated. It is not supported elsewhere so far. Interfaces so marked may be withdrawn in future releases in accordance with QEMU's deprecation policy.
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
Yeah, I saw that, but wasn't aware of -compat, thanks.
Quoted 4 lines
See also -compat parameters unstable-input, unstable-output, both intended for "testing the future".
Quoted 8 lines
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
Events are part of the interface just like command returns are. Not emitting an event in a situation where it was emitted before can easily break things. Only when the situation is no longer possible, the event can be removed safely.
@Pierre, seems it would be a good idea to mark all changes to qmp unstable, not just set-cpu-topology, can just remove it later after all.
Quoted 5 lines
Questions?
[...]
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 3 lines · Nina Schoetterl-Glausch
Quoted 2 lines · Markus Armbruster
Nina Schoetterl-Glausch <[email protected]> writes:
...
Quoted 30 lines
Quoted 11 lines
Quoted 5 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
docs/devel/qapi-code-gen.rst:Special features
~~~~~~~~~~~~~~~~Quoted 8 lines
Feature "deprecated" marks a command, event, enum value, or struct member as deprecated. It is not supported elsewhere so far. Interfaces so marked may be withdrawn in future releases in accordance with QEMU's deprecation policy.
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
Yeah, I saw that, but wasn't aware of -compat, thanks.
Quoted 8 lines
See also -compat parameters unstable-input, unstable-output, both intended for "testing the future".
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
Quoted 7 lines
Quoted 4 lines
Events are part of the interface just like command returns are. Not emitting an event in a situation where it was emitted before can easily break things. Only when the situation is no longer possible, the event can be removed safely.
@Pierre, seems it would be a good idea to mark all changes to qmp unstable, not just set-cpu-topology, can just remove it later after all.
OK.
Just curious: how do you think this simple event matching the guest request 1 on 1 may evolve?
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 43 lines · Pierre Morel
Quoted 3 lines · Nina Schoetterl-Glausch
Quoted 2 lines · Markus Armbruster
Nina Schoetterl-Glausch <[email protected]> writes:
...
Quoted 35 lines
Quoted 11 lines
Quoted 5 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
docs/devel/qapi-code-gen.rst:Special features
~~~~~~~~~~~~~~~~Quoted 8 lines
Feature "deprecated" marks a command, event, enum value, or struct member as deprecated. It is not supported elsewhere so far. Interfaces so marked may be withdrawn in future releases in accordance with QEMU's deprecation policy.
Feature "unstable" marks a command, event, enum value, or struct member as unstable. It is not supported elsewhere so far. Interfaces so marked may be withdrawn or changed incompatibly in future releases.
Yeah, I saw that, but wasn't aware of -compat, thanks.
Quoted 12 lines
See also -compat parameters unstable-input, unstable-output, both intended for "testing the future".
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
Events are part of the interface just like command returns are. Not emitting an event in a situation where it was emitted before can easily break things. Only when the situation is no longer possible, the event can be removed safely.
Quoted 7 lines
Quoted 2 lines
@Pierre, seems it would be a good idea to mark all changes to qmp unstable, not just set-cpu-topology, can just remove it later after all.
OK.
Just curious: how do you think this simple event matching the guest request 1 on 1 may evolve?
Well, I don't know really, making it unstable is just more conservative for now. But this way you can prototype/implement support in libvirt for topology and then make the interface stable one you've confirmed that everything works as intended.
Here is something to think about: The architecture allows rejection of the PTF polarization change request, if a request is currently in process. A possible design to allow the same semantics for qemu/libvirt would be to set a polarization_change_in_progess bit when a PTF request occurs and refuse subsequent requests until this bit was reset via qmp. This wouldn't result in the event data structure being changed, but its semantics, since it isn't fired for every request anymore.
Quoted 4 lines
Regards,
Pierre
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
In reply to Nina Schoetterl-GlauschQuoted 54 lines · Nina Schoetterl-Glausch
Quoted 38 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +19
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +#qapi/machine-target.json line 402
Quoted 10 lines
qapi/machine-target.json +10
-0+# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +# +## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }I wonder if you should depend on
CONFIG_KVMor not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.
Quoted 3 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
The intention of this flag is to allow mgmt apps to make a usage policy decision.
Libvirt's policy is that we'll never use features marked unstable.
IOW, the consequence of marking it unstable is that it'll likely go unused until the unstable marker gets removed.
Using 'unstable' is useful if you want to get complex code merged before you're quite happy with the design, and then iterate on the impl in-tree. This is OK if there's no urgent need for apps to consume the feature. If you want the feature to be used for real though, the unstable flag is not desirable and you need to finalize the design.
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
As Markus says, that's not a safe assumption. If a mgmt app is expecting to receive an event, ceasing to emit it would likely be considered a regression.
With regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 1 line · Daniel P. Berrangé
On Wed, Feb 08, 2023 at 06:35:39PM +0100, Nina Schoetterl-Glausch wrote:
[...]
Quoted 12 lines
Quoted 3 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.The intention of this flag is to allow mgmt apps to make a usage policy decision.
Libvirt's policy is that we'll never use features marked unstable.
IOW, the consequence of marking it unstable is that it'll likely go unused until the unstable marker gets removed.
Quoted 6 lines
Using 'unstable' is useful if you want to get complex code merged before you're quite happy with the design, and then iterate on the impl in-tree. This is OK if there's no urgent need for apps to consume the feature. If you want the feature to be used for real though, the unstable flag is not desirable and you need to finalize the design.
Another use of 'unstable' is debugging aids. Making these stable can be plenty of pain for precious little gain.
[...]
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
In reply to Daniel P. BerrangéQuoted 55 lines · Daniel P. Berrangé
Quoted 54 lines · Nina Schoetterl-Glausch
Quoted 38 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +19
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +#qapi/machine-target.json line 402
Quoted 10 lines
qapi/machine-target.json +10
-0+# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +# +## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }I wonder if you should depend on
CONFIG_KVMor not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.
Quoted 8 lines
Quoted 3 lines
I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.The intention of this flag is to allow mgmt apps to make a usage policy decision.
Libvirt's policy is that we'll never use features marked unstable.
Does it enforce that, e.g via compat policies?
If so, I assume there is some way to allow use of unstable features in libvirt for development? If for example you're prototyping a new mgmt feature that uses unstable commands.
Quoted 22 lines
IOW, the consequence of marking it unstable is that it'll likely go unused until the unstable marker gets removed.
Using 'unstable' is useful if you want to get complex code merged before you're quite happy with the design, and then iterate on the impl in-tree. This is OK if there's no urgent need for apps to consume the feature. If you want the feature to be used for real though, the unstable flag is not desirable and you need to finalize the design.
Quoted 3 lines
Also I guess one can remove qemu events without breaking backwards compatibility, since they just won't be emitted? Unless I guess you specify that a event must occur under certain situations and the client waits on it?
As Markus says, that's not a safe assumption. If a mgmt app is expecting to receive an event, ceasing to emit it would likely be considered a regression.
With regards,
Daniel
Re: [PATCH v15 10/11] qapi/s390x/cpu topology: CPU_POLARITY_CHANGE qapi event
Quoted 60 lines · Nina Schoetterl-Glausch
Quoted 59 lines · Daniel P. Berrangé
Quoted 57 lines · Nina Schoetterl-Glausch
Quoted 38 lines · Pierre Morel
When the guest asks to change the polarity this change is forwarded to the admin using QAPI.
The admin is supposed to take according decisions concerning CPU provisioning.Signed-off-by Pierre Morel <[email protected]>--- qapi/machine-target.json | 30 ++++++++++++++++++++++++++++++ hw/s390x/cpu-topology.c | 2 ++ 2 files changed, 32 insertions(+)qapi/machine-target.json +19
-0@@ -371,3 +371,33 @@ }, 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM' ] } } + +## +# @CPU_POLARITY_CHANGE: +# +# Emitted when the guest asks to change the polarity. +# +# @polarity: polarity specified by the guest +# +# The guest can tell the host (via the PTF instruction) whether the +# CPUs should be provisioned using horizontal or vertical polarity. +# +# On horizontal polarity the host is expected to provision all vCPUs +# equally. +# On vertical polarity the host can provision each vCPU differently. +# The guest will get information on the details of the provisioning +# the next time it uses the STSI(15) instruction. +# +# Since: 8.0 +#qapi/machine-target.json line 402
Quoted 10 lines
qapi/machine-target.json +10
-0+# Example: +# +# <- { "event": "CPU_POLARITY_CHANGE", +# "data": { "polarity": 0 }, +# "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } +# +## +{ 'event': 'CPU_POLARITY_CHANGE', + 'data': { 'polarity': 'int' }, + 'if': { 'all': [ 'TARGET_S390X', 'CONFIG_KVM'] }I wonder if you should depend on
CONFIG_KVMor not. If tcg gets topology support it will use the same event and right now it would just never be emitted. On the other hand it's more conservative this way.I also wonder if you should add 'feature' : [ 'unstable' ].
On the upside, it would mark the event as unstable, but I don't know what the consequences are exactly.
Quoted 6 lines
Quoted 4 lines
The intention of this flag is to allow mgmt apps to make a usage policy decision.
Libvirt's policy is that we'll never use features marked unstable.
Does it enforce that, e.g via compat policies?
The policy is applied at time of code review, in that we'll not merge patches that use features marked unstable.
Quoted 2 lines
If so, I assume there is some way to allow use of unstable features in libvirt for development? If for example you're prototyping a new mgmt feature that uses unstable commands.
You can prototype usage in libvirt in a fork of course, but we won't take patches into the libvirt upstream repo.
Alternatively in some cases the the libvirt QMP passthrough can be used for experiemnts (eg virsh qemu-monitor-command ) in a non-production envionrment.
With regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
PATCH v15 11/11 docs/s390x/cpu topology: document s390x cpu topology
In reply to Pierre MorelAdd some basic examples for the definition of cpu topology in s390x.
--- docs/system/s390x/cpu-topology.rst | 294 +++++++++++++++++++++++++++++ docs/system/target-s390x.rst | 1 + 2 files changed, 295 insertions(+) create mode 100644 docs/system/s390x/cpu-topology.rst
docs/system/s390x/cpu-topology.rst new file +294 -0
@@ -0,0 +1,294 @@ +CPU topology on s390x +===================== + +Since QEMU 8.0, CPU topology on s390x provides up to 4 levels of +topology containers: drawers, books, sockets and cores. + +The first three containers define a tree hierarchy, the last one +provides the placement of the CPUs inside the parent container and +3 CPU attributes: + +- CPU type +- polarity entitlement +- dedication + +Note also that since 7.2 threads are no longer supported in the topology +and the ``-smp`` command line argument accepts only ``threads=1``. + +Prerequisites +------------- + +To use CPU topology a Linux QEMU/KVM machine providing the CPU topology facility +(STFLE bit 11) is required. + +However, since this facility has been enabled by default in an early version +of QEMU, we use a capability, ``KVM_CAP_S390_CPU_TOPOLOGY``, to notify KVM +that QEMU is supporting CPU topology. + +Enabling CPU topology +--------------------- + +Currently, CPU topology is only enabled in the host model by default. + +Enabling CPU topology in a CPU model is done by setting the CPU flag +``ctop`` to ``on`` like in: + +.. code-block:: bash + + -cpu gen16b,ctop=on + +Having the topology disabled by default allows migration between +old and new QEMU without adding new flags. + +Default topology usage +---------------------- + +The CPU topology, can be specified on the QEMU command line +with the ``-smp`` or the ``-device`` QEMU command arguments +without using any new attributes. +In this case, the topology will be calculated by simply adding +to the topology the cores based on the core-id starting with +core-0 at position 0 of socket-0, book-0, drawer-0 with default +modifier attributes: horizontal polarity and no dedication. + +In the following machine we define 8 sockets with 4 cores each. +Note that s390x QEMU machines do not implement multithreading. + +.. code-block:: bash + + $ qemu-system-s390x -m 2G \ + -cpu gen16b,ctop=on \ + -smp cpus=5,sockets=8,cores=4,maxcpus=32 \ + -device host-s390x-cpu,core-id=14 \ + +New CPUs can be plugged using the device_add hmp command like in: + +.. code-block:: bash + + (qemu) device_add gen16b-s390x-cpu,core-id=9 + +The core-id defines the placement of the core in the topology by +starting with core 0 in socket 0 up to maxcpus. + +In the example above: + +* There are 5 CPUs provided to the guest with the ``-smp`` command line + They will take the core-ids 0,1,2,3,4 + As we have 4 cores in a socket, we have 4 CPUs provided + to the guest in socket 0, with core-ids 0,1,2,3. + The last cpu, with core-id 4, will be on socket 1. + +* the core with ID 14 provided by the ``-device`` command line will + be placed in socket 3, with core-id 14 + +* the core with ID 9 provided by the ``device_add`` qmp command will + be placed in socket 2, with core-id 9 + +Polarity and dedication +----------------------- + +Polarity can be of two types: horizontal or vertical. + +The horizontal polarization specifies that all guest's vCPUs get +almost the same amount of provisioning of real CPU by the host. + +The vertical polarization specifies that guest's vCPU can get +different real CPU provisions: + +- a vCPU with Vertical high entitlement specifies that this + vCPU gets 100% of the real CPU provisioning. + +- a vCPU with Vertical medium entitlement specifies that this + vCPU shares the real CPU with other vCPUs. + +- a vCPU with Vertical low entitlement specifies that this + vCPU only get real CPU provisioning when no other vCPU need it. + +In the case a vCPU with vertical high entitlement does not use +the real CPU, the unused "slack" can be dispatched to other vCPU +with medium or low entitlement. + +A subsystem reset puts all vCPU of the configuration into the +horizontal polarization. + +The admin specifies the dedicated bit when the vCPU is dedicated +to a single real CPU. + +As for the Linux admin, the dedicated bit is an indication on the +affinity of a vCPU for a real CPU while the entitlement indicates the +sharing or exclusivity of use. + +Defining the topology on command line +------------------------------------- + +The topology can be defined entirely during the CPU definition, +with the exception of CPU 0 which must be defined with the -smp +argument. + +For example, here we set the position of the cores 1,2,3 on +drawer 1, book 1, socket 2 and cores 0,9 and 14 on drawer 0, +book 0, socket 0 with all horizontal polarity and not dedicated. +The core 4, will be set on its default position on socket 1 +(since we have 4 core per socket) and we define it with dedication and +vertical high entitlement. + +.. code-block:: bash + + $ qemu-system-s390x -m 2G \ + -cpu gen16b,ctop=on \ + -smp cpus=1,sockets=8,cores=4,maxcpus=32 \ + \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=1 \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=2 \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=3 \ + \ + -device gen16b-s390x-cpu,drawer-id=0,book-id=0,socket-id=0,core-id=9 \ + -device gen16b-s390x-cpu,drawer-id=0,book-id=0,socket-id=0,core-id=14 \ + \ + -device gen16b-s390x-cpu,core-id=4,dedicated=on,polarity=3 \ + +QAPI interface for topology +--------------------------- + +Let's start QEMU with the following command: + +.. code-block:: bash + + sudo /usr/local/bin/qemu-system-s390x \ + -enable-kvm \ + -cpu z14,ctop=on \ + -smp 1,drawers=3,books=3,sockets=2,cores=2,maxcpus=36 \ + \ + -device z14-s390x-cpu,core-id=19,polarity=3 \ + -device z14-s390x-cpu,core-id=11,polarity=1 \ + -device z14-s390x-cpu,core-id=112,polarity=3 \ + ... + +and see the result when using of the QAPI interface. + +addons to query-cpus-fast ++++++++++++++++++++++++++ + +The command query-cpus-fast allows the admin to query the topology +tree and modifiers for all configured vCPU. + +.. code-block:: QMP + + -> { "execute": "query-cpus-fast" } + { + "return": [ + { + "dedicated": false, + "thread-id": 3631238, + "props": { + "core-id": 0, + "socket-id": 0, + "drawer-id": 0, + "book-id": 0 + }, + "cpu-state": "operating", + "qom-path": "/machine/unattached/device[0]", + "polarity": 2, + "cpu-index": 0, + "target": "s390x" + }, + { + "dedicated": false, + "thread-id": 3631248, + "props": { + "core-id": 19, + "socket-id": 9, + "drawer-id": 0, + "book-id": 2 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[0]", + "polarity": 3, + "cpu-index": 19, + "target": "s390x" + }, + { + "dedicated": false, + "thread-id": 3631249, + "props": { + "core-id": 11, + "socket-id": 5, + "drawer-id": 0, + "book-id": 1 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[1]", + "polarity": 1, + "cpu-index": 11, + "target": "s390x" + }, + { + "dedicated": true, + "thread-id": 3631250, + "props": { + "core-id": 112, + "socket-id": 56, + "drawer-id": 3, + "book-id": 14 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[2]", + "polarity": 3, + "cpu-index": 112, + "target": "s390x" + } + ] + } + +x-set-cpu-topology +++++++++++++++++++ + +The command x-set-cpu-topology allows the admin to modify the topology +tree or the topology modifiers of a vCPU in the configuration. + +.. code-block:: QMP + + -> { "execute": "x-set-cpu-topology", + "arguments": { + "core": 11, + "socket": 0, + "book": 0, + "drawer": 0, + "polarity": 0, + "dedicated": false + } + } + <- {"return": {}} + + +event CPU_POLARITY_CHANGE ++++++++++++++++++++++++++ + +When a guest is requesting a modification of the polarity, +QEMU sends a CPU_POLARITY_CHANGE event. + +When requesting the change, the guest only specifies horizontal or +vertical polarity. +The dedication and fine grain vertical entitlement depends on admin +to set according to its response to this event. + +Note that a vertical polarized dedicated vCPU can only have a high +entitlement, this gives 6 possibilities for a vCPU polarity: + +- Horizontal +- Horizontal dedicated +- Vertical low +- Vertical medium +- Vertical high +- Vertical high dedicated + +Example of the event received when the guest issues the CPU instruction +Perform Topology Function PTF(0) to request an horizontal polarity: + +.. code-block:: QMP + + <- { "event": "CPU_POLARITY_CHANGE", + "data": { "polarity": 0 }, + "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } + +
docs/system/target-s390x.rst +1 -0
@@ -33,3 +33,4 @@ Architectural features .. toctree:: s390x/bootdevices s390x/protvirt + s390x/cpu-topology
Signature · 1 line
2.31.1
Re: [PATCH v15 11/11] docs/s390x/cpu topology: document s390x cpu topology
Quoted 10 lines · Pierre Morel
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Add some basic examples for the definition of cpu topology in s390x.
Signed-off-by Pierre Morel <[email protected]>---
docs/system/s390x/cpu-topology.rst | 294 +++++++++++++++++++++++++++++
docs/system/target-s390x.rst | 1 +2 files changed, 295 insertions(+)create mode 100644
docs/system/s390x/cpu-topology.rst
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
docs/system/s390x/cpu-topology.rst line 5
Quoted 11 lines
docs/system/s390x/cpu-topology.rst new file +5
-0@@ -0,0 +1,294 @@ +CPU topology on s390x +===================== + +Since QEMU 8.0, CPU topology on s390x provides up to 4 levels of +topology containers: drawers, books, sockets and cores.
Cores isn't a container itself, at least as defined by STSI. Maybe use "nodes" instead?
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
docs/system/s390x/cpu-topology.rst line 9
Quoted 4 lines
docs/system/s390x/cpu-topology.rst new file +4
-0+ +The first three containers define a tree hierarchy, the last one +provides the placement of the CPUs inside the parent container and +3 CPU attributes:
The latter half of the sentence is a bit confusing, it's too informed by the output of STSI.
Suggestion:
The first three are containers, defining a tree shaped hierarchy. There are three attributes associated with each core:
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
docs/system/s390x/cpu-topology.rst new file +9
-0+ +- CPU type +- polarity entitlement +- dedication + +Note also that since 7.2 threads are no longer supported in the topology +and the ``-smp`` command line argument accepts only ``threads=1``. + +Prerequisites
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 26
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+------------- + +To use CPU topology a Linux QEMU/KVM machine providing the CPU topology facility +(STFLE bit 11) is required. + +However, since this facility has been enabled by default in an early version +of QEMU, we use a capability, ``KVM_CAP_S390_CPU_TOPOLOGY``, to notify KVM +that QEMU is supporting CPU topology.
supports
Quoted 12 lines
Quoted 12 lines
Quoted 12 lines
Quoted 12 lines
docs/system/s390x/cpu-topology.rst new file +12
-0+ +Enabling CPU topology +--------------------- + +Currently, CPU topology is only enabled in the host model by default. + +Enabling CPU topology in a CPU model is done by setting the CPU flag +``ctop`` to ``on`` like in: + +.. code-block:: bash + + -cpu gen16b,ctop=on
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 46
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+ +Having the topology disabled by default allows migration between +old and new QEMU without adding new flags. + +Default topology usage +---------------------- + +The CPU topology, can be specified on the QEMU command line
Drop the comma.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
docs/system/s390x/cpu-topology.rst line 48
Quoted 2 lines
docs/system/s390x/cpu-topology.rst new file +2
-0+with the ``-smp`` or the ``-device`` QEMU command arguments +without using any new attributes.
Might want to split this up for -smp and -device.
If none of the new attributes (drawers, books) are specified for the -smp flag, <description of default>
If none of the new attributes (drawer, book, dedicated, entitlement) are specified for the -device cpu flag, the topology will be ...
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
docs/system/s390x/cpu-topology.rst line 55
Quoted 7 lines
docs/system/s390x/cpu-topology.rst new file +7
-0+In this case, the topology will be calculated by simply adding +to the topology the cores based on the core-id starting with +core-0 at position 0 of socket-0, book-0, drawer-0 with default +modifier attributes: horizontal polarity and no dedication. + +In the following machine we define 8 sockets with 4 cores each. +Note that s390x QEMU machines do not implement multithreading.
You could drop that since you already said it earlier.
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
docs/system/s390x/cpu-topology.rst line 64
Quoted 9 lines
docs/system/s390x/cpu-topology.rst new file +9
-0+ +.. code-block:: bash + + $ qemu-system-s390x -m 2G \ + -cpu gen16b,ctop=on \ + -smp cpus=5,sockets=8,cores=4,maxcpus=32 \ + -device host-s390x-cpu,core-id=14 \ + +New CPUs can be plugged using the device_add hmp command like in:
"as in" or "like so" sounds better to my non-native ears. Applies to other instances, too.
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 71
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +7
-0+ +.. code-block:: bash + + (qemu) device_add gen16b-s390x-cpu,core-id=9 + +The core-id defines the placement of the core in the topology by +starting with core 0 in socket 0 up to maxcpus.
Suggestion:
The same placement of the CPU is derived from the core-id as described above.
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
docs/system/s390x/cpu-topology.rst new file +5
-0+ +In the example above: + +* There are 5 CPUs provided to the guest with the ``-smp`` command line + They will take the core-ids 0,1,2,3,4
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 84
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+ As we have 4 cores in a socket, we have 4 CPUs provided + to the guest in socket 0, with core-ids 0,1,2,3. + The last cpu, with core-id 4, will be on socket 1. + +* the core with ID 14 provided by the ``-device`` command line will + be placed in socket 3, with core-id 14 + +* the core with ID 9 provided by the ``device_add`` qmp command will
You said it's a hmp command earlier.
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
docs/system/s390x/cpu-topology.rst line 87
Quoted 3 lines
docs/system/s390x/cpu-topology.rst new file +3
-0+ be placed in socket 2, with core-id 9 + +Polarity and dedication
s/Polarity/Polarization, same below
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
docs/system/s390x/cpu-topology.rst line 97
Quoted 10 lines
docs/system/s390x/cpu-topology.rst new file +10
-0+----------------------- + +Polarity can be of two types: horizontal or vertical. + +The horizontal polarization specifies that all guest's vCPUs get +almost the same amount of provisioning of real CPU by the host. + +The vertical polarization specifies that guest's vCPU can get +different real CPU provisions: +
No reason to capitalize vertical below, is there?
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 105
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+- a vCPU with Vertical high entitlement specifies that this + vCPU gets 100% of the real CPU provisioning. + +- a vCPU with Vertical medium entitlement specifies that this + vCPU shares the real CPU with other vCPUs. + +- a vCPU with Vertical low entitlement specifies that this + vCPU only get real CPU provisioning when no other vCPU need it.
getS, vCPUs or needS
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
docs/system/s390x/cpu-topology.rst line 107
Quoted 2 lines
docs/system/s390x/cpu-topology.rst new file +2
-0+ +In the case a vCPU with vertical high entitlement does not use
In case
Quoted 1 line
Quoted 1 line
Quoted 1 line
docs/system/s390x/cpu-topology.rst line 108
Quoted 1 line
docs/system/s390x/cpu-topology.rst new file +1
-0+the real CPU, the unused "slack" can be dispatched to other vCPU
vCPUs
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
docs/system/s390x/cpu-topology.rst line 111
Quoted 3 lines
docs/system/s390x/cpu-topology.rst new file +3
-0+with medium or low entitlement. + +A subsystem reset puts all vCPU of the configuration into the
vCPUs
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
docs/system/s390x/cpu-topology.rst line 117
Quoted 6 lines
docs/system/s390x/cpu-topology.rst new file +6
-0+horizontal polarization. + +The admin specifies the dedicated bit when the vCPU is dedicated +to a single real CPU. + +As for the Linux admin, the dedicated bit is an indication on the
Drop everything before the comma. indication of
Quoted 1 line
Quoted 1 line
Quoted 1 line
docs/system/s390x/cpu-topology.rst line 118
Quoted 1 line
docs/system/s390x/cpu-topology.rst new file +1
-0+affinity of a vCPU for a real CPU while the entitlement indicates the
affinity of a vCPU to a real CPU, while
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
docs/system/s390x/cpu-topology.rst line 121
Quoted 3 lines
docs/system/s390x/cpu-topology.rst new file +3
-0+sharing or exclusivity of use. + +Defining the topology on command line
on the command line
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
docs/system/s390x/cpu-topology.rst line 124
Quoted 3 lines
docs/system/s390x/cpu-topology.rst new file +3
-0+------------------------------------- + +The topology can be defined entirely during the CPU definition,
The topology can entirely be defined using -device cpu statements,
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
docs/system/s390x/cpu-topology.rst line 128
Quoted 4 lines
docs/system/s390x/cpu-topology.rst new file +4
-0+with the exception of CPU 0 which must be defined with the -smp +argument. + +For example, here we set the position of the cores 1,2,3 on
position ... to
Quoted 1 line
Quoted 1 line
Quoted 1 line
docs/system/s390x/cpu-topology.rst line 129
Quoted 1 line
docs/system/s390x/cpu-topology.rst new file +1
-0+drawer 1, book 1, socket 2 and cores 0,9 and 14 on drawer 0,
same here, s/on/to
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
docs/system/s390x/cpu-topology.rst line 131
Quoted 2 lines
docs/system/s390x/cpu-topology.rst new file +2
-0+book 0, socket 0 with all horizontal polarity and not dedicated. +The core 4, will be set on its default position on socket 1
set to ... in socket 1
Quoted 28 lines
Quoted 28 lines
Quoted 28 lines
Quoted 28 lines
docs/system/s390x/cpu-topology.rst new file +28
-0+(since we have 4 core per socket) and we define it with dedication and +vertical high entitlement. + +.. code-block:: bash + + $ qemu-system-s390x -m 2G \ + -cpu gen16b,ctop=on \ + -smp cpus=1,sockets=8,cores=4,maxcpus=32 \ + \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=1 \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=2 \ + -device gen16b-s390x-cpu,drawer-id=1,book-id=1,socket-id=2,core-id=3 \ + \ + -device gen16b-s390x-cpu,drawer-id=0,book-id=0,socket-id=0,core-id=9 \ + -device gen16b-s390x-cpu,drawer-id=0,book-id=0,socket-id=0,core-id=14 \ + \ + -device gen16b-s390x-cpu,core-id=4,dedicated=on,polarity=3 \ + +QAPI interface for topology +--------------------------- + +Let's start QEMU with the following command: + +.. code-block:: bash + + sudo /usr/local/bin/qemu-system-s390x \ + -enable-kvm \ + -cpu z14,ctop=on \
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 167
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+ -smp 1,drawers=3,books=3,sockets=2,cores=2,maxcpus=36 \ + \ + -device z14-s390x-cpu,core-id=19,polarity=3 \ + -device z14-s390x-cpu,core-id=11,polarity=1 \ + -device z14-s390x-cpu,core-id=112,polarity=3 \ + ... + +and see the result when using of the QAPI interface.
s/of//
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
docs/system/s390x/cpu-topology.rst line 173
Quoted 6 lines
docs/system/s390x/cpu-topology.rst new file +6
-0+ +addons to query-cpus-fast ++++++++++++++++++++++++++ + +The command query-cpus-fast allows the admin to query the topology +tree and modifiers for all configured vCPU.
vCPUs
Quoted 86 lines
Quoted 86 lines
Quoted 86 lines
Quoted 86 lines
docs/system/s390x/cpu-topology.rst new file +86
-0+ +.. code-block:: QMP + + -> { "execute": "query-cpus-fast" } + { + "return": [ + { + "dedicated": false, + "thread-id": 3631238, + "props": { + "core-id": 0, + "socket-id": 0, + "drawer-id": 0, + "book-id": 0 + }, + "cpu-state": "operating", + "qom-path": "/machine/unattached/device[0]", + "polarity": 2, + "cpu-index": 0, + "target": "s390x" + }, + { + "dedicated": false, + "thread-id": 3631248, + "props": { + "core-id": 19, + "socket-id": 9, + "drawer-id": 0, + "book-id": 2 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[0]", + "polarity": 3, + "cpu-index": 19, + "target": "s390x" + }, + { + "dedicated": false, + "thread-id": 3631249, + "props": { + "core-id": 11, + "socket-id": 5, + "drawer-id": 0, + "book-id": 1 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[1]", + "polarity": 1, + "cpu-index": 11, + "target": "s390x" + }, + { + "dedicated": true, + "thread-id": 3631250, + "props": { + "core-id": 112, + "socket-id": 56, + "drawer-id": 3, + "book-id": 14 + }, + "cpu-state": "operating", + "qom-path": "/machine/peripheral-anon/device[2]", + "polarity": 3, + "cpu-index": 112, + "target": "s390x" + } + ] + } + +x-set-cpu-topology +++++++++++++++++++ + +The command x-set-cpu-topology allows the admin to modify the topology +tree or the topology modifiers of a vCPU in the configuration. + +.. code-block:: QMP + + -> { "execute": "x-set-cpu-topology", + "arguments": { + "core": 11, + "socket": 0, + "book": 0, + "drawer": 0, + "polarity": 0, + "dedicated": false + }
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
docs/system/s390x/cpu-topology.rst line 267
Quoted 8 lines
docs/system/s390x/cpu-topology.rst new file +8
-0+ } + <- {"return": {}} + + +event CPU_POLARITY_CHANGE ++++++++++++++++++++++++++ + +When a guest is requesting a modification of the polarity,
requests
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
docs/system/s390x/cpu-topology.rst line 273
Quoted 6 lines
docs/system/s390x/cpu-topology.rst new file +6
-0+QEMU sends a CPU_POLARITY_CHANGE event. + +When requesting the change, the guest only specifies horizontal or +vertical polarity. +The dedication and fine grain vertical entitlement depends on admin +to set according to its response to this event.
It is the job of the admin to set the dedication and fine grained vertical entitlement in response to this event.
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
docs/system/s390x/cpu-topology.rst line 276
Quoted 3 lines
docs/system/s390x/cpu-topology.rst new file +3
-0+ +Note that a vertical polarized dedicated vCPU can only have a high +entitlement, this gives 6 possibilities for a vCPU polarity:
for vCPU polarization.
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
docs/system/s390x/cpu-topology.rst new file +18
-0+ +- Horizontal +- Horizontal dedicated +- Vertical low +- Vertical medium +- Vertical high +- Vertical high dedicated + +Example of the event received when the guest issues the CPU instruction +Perform Topology Function PTF(0) to request an horizontal polarity: + +.. code-block:: QMP + + <- { "event": "CPU_POLARITY_CHANGE", + "data": { "polarity": 0 }, + "timestamp": { "seconds": 1401385907, "microseconds": 422329 } } + +docs/system/target-s390x.rst +0
-0
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
index
c636f64113..ff0ffe04f3100644docs/system/target-s390x.rst +0
-0@@ -33,3 +33,4 @@ Architectural features.. toctree::
s390x/bootdevices
s390x/protvirt
+ s390x/cpu-topology
I'm not too big a fan of using "admin", I guess in reality it's mostly going to be libvirt adjusting things. Maybe reformulate sentences to get rid of it or use "outside administrative entity" or something similar. But in the end it's not too important.
IMO this series looks good overall and like it's nearing the final stages.
You use "polarity" instead of "polarization" a lot.
Since the PoP uses polarization I think that term would be preferred.
With the series as it is, one cannot set the polarization via qmp, only the entitlement of individual cpus. So only the VM can change the polarization.
Would it be desirable to also be able to set the polarization from the outside?
Like I said in one response, it would be good to consider if we need an polarization_change_in_progress state that refuses requests.
I'm guessing not, if a request is always completed before the next is handled and there is no way to lose requests.
I don't know how long it would take to change the CPU assignment and if there is a chance that could be overwhelmed by too many requests. Probably not but something worth thinking about.
Might be a good idea to add a test case that performs test via qmp. So starts an instance with some cpu topology assignments, checks that qmp returns the correct topology, hot plugs a cpu and does another check, Changes topology attributes, etc.
I guess this would be an avocado test.
Quoted 1 line · Nina Schoetterl-Glausch
IMO this series looks good overall and like it's nearing the final stages.
Thank you for your helping this.
Quoted 3 lines
You use "polarity" instead of "polarization" a lot.
Since the PoP uses polarization I think that term would be preferred.
OK
Quoted 5 lines
With the series as it is, one cannot set the polarization via qmp, only the entitlement of individual cpus. So only the VM can change the polarization.
Would it be desirable to also be able to set the polarization from the outside?
I do not think so, AFAIK this is not foreseen by the architecture.
My point of view on this is that the application running on the guest is the one knowing if it can get use of the heterogeneous CPU provisioning provided by the vertical polarization or not. Horizontal polarization being the default with an homogeneous, or considered as default, provisioning.
Quoted 8 lines
Like I said in one response, it would be good to consider if we need an
polarization_change_in_progressstate that refuses requests.
I'm guessing not, if a request is always completed before the next is handled and there is no way to lose requests.
I don't know how long it would take to change the CPU assignment and if there is a chance that could be overwhelmed by too many requests. Probably not but something worth thinking about.
Currently, guest request for a topology change is done via the sysfs interface by a userland process.
The value returned by the kernel to userland is -BUSY, for both DONE and IN_PROGRESS.
So at first sight I do not see a overwhelming problem but having a completion indication looks like a good thing to have in a future extension in both QEMU and Kernel
Quoted 6 lines
Might be a good idea to add a test case that performs test via qmp. So starts an instance with some cpu topology assignments, checks that qmp returns the correct topology, hot plugs a cpu and does another check, Changes topology attributes, etc.
I guess this would be an avocado test.
Yes you are right there is a lot to test.
There is already a test for kvm_unit_tests in review to test PTF and STSI instruction's interception. We do not use avocado as far as I know but our hades tests framework for the kind of tests you propose. I never used avocado for now but at first sight, avocado and hades look similar.
Regards,
Pierre
Signature · 2 lines
Pierre Morel IBM Lab Boeblingen