qemu-patch-raspberry4/hw/ppc/spapr_cpu_core.c

405 lines
11 KiB
C
Raw Normal View History

/*
* sPAPR CPU core device, acts as container of CPU thread devices.
*
* Copyright (C) 2016 Bharata B Rao <bharata@linux.vnet.ibm.com>
*
* This work is licensed under the terms of the GNU GPL, version 2 or later.
* See the COPYING file in the top-level directory.
*/
#include "qemu/osdep.h"
#include "hw/cpu/core.h"
#include "hw/ppc/spapr_cpu_core.h"
#include "target/ppc/cpu.h"
#include "hw/ppc/spapr.h"
#include "hw/boards.h"
#include "qapi/error.h"
#include "sysemu/cpus.h"
#include "sysemu/kvm.h"
#include "target/ppc/kvm_ppc.h"
#include "hw/ppc/ppc.h"
#include "target/ppc/mmu-hash64.h"
#include "sysemu/numa.h"
#include "sysemu/hw_accel.h"
#include "qemu/error-report.h"
static void spapr_cpu_reset(void *opaque)
{
PowerPCCPU *cpu = opaque;
CPUState *cs = CPU(cpu);
CPUPPCState *env = &cpu->env;
PowerPCCPUClass *pcc = POWERPC_CPU_GET_CLASS(cpu);
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuState *spapr_cpu = spapr_cpu_state(cpu);
target_ulong lpcr;
cpu_reset(cs);
/* All CPUs start halted. CPU0 is unhalted from the machine level
* reset code and the rest are explicitly started up by the guest
* using an RTAS call */
cs->halted = 1;
Fix a deadlock case in the CPU hotplug flow We need to set cs->halted to 1 before calling ppc_set_compat. The reason is that ppc_set_compat kicks up the new thread created to manage the hotplugged KVM virtual CPU and the code drives directly to KVM_RUN ioctl. When cs->halted is 1, the code: int kvm_cpu_exec(CPUState *cpu) ... if (kvm_arch_process_async_events(cpu)) { atomic_set(&cpu->exit_request, 0); return EXCP_HLT; } ... returns before it reaches KVM_RUN, giving time to the main thread to finish its job. Otherwise we can fall in a deadlock because the KVM thread will issue the KVM_RUN ioctl while the main thread is setting up KVM registers. Depending on how these jobs are scheduled we'll end up freezing QEMU. The following output shows kvm_vcpu_ioctl sleeping because it cannot get the mutex and never will. PS: kvm_vcpu_ioctl was triggered kvm_set_one_reg - compat_pvr. STATE: TASK_UNINTERRUPTIBLE|TASK_WAKEKILL PID: 61564 TASK: c000003e981e0780 CPU: 48 COMMAND: "qemu-system-ppc" #0 [c000003e982679a0] __schedule at c000000000b10a44 #1 [c000003e98267a60] schedule at c000000000b113a8 #2 [c000003e98267a90] schedule_preempt_disabled at c000000000b11910 #3 [c000003e98267ab0] __mutex_lock at c000000000b132ec #4 [c000003e98267bc0] kvm_vcpu_ioctl at c00800000ea03140 [kvm] #5 [c000003e98267d20] do_vfs_ioctl at c000000000407d30 #6 [c000003e98267dc0] ksys_ioctl at c000000000408674 #7 [c000003e98267e10] sys_ioctl at c0000000004086f8 #8 [c000003e98267e30] system_call at c00000000000b488 crash> struct -x kvm.vcpus 0xc000003da0000000 vcpus = {0xc000003db4880000, 0xc000003d52b80000, 0xc0000039e9c80000, 0xc000003d0e200000, 0xc000003d58280000, 0x0, 0x0, ...} crash> struct -x kvm_vcpu.mutex.owner 0xc000003d58280000 mutex.owner = { counter = 0xc000003a23a5c881 <- flag 1: waiters }, crash> bt 0xc000003a23a5c880 PID: 61579 TASK: c000003a23a5c880 CPU: 9 COMMAND: "CPU 4/KVM" (active) crash> struct -x kvm_vcpu.mutex.wait_list 0xc000003d58280000 mutex.wait_list = { next = 0xc000003e98267b10, prev = 0xc000003e98267b10 }, crash> struct -x mutex_waiter.task 0xc000003e98267b10 task = 0xc000003e981e0780 The following command-line was used to reproduce the problem (note: gdb and trace can change the results). $ qemu-ppc/build/ppc64-softmmu/qemu-system-ppc64 -cpu host \ -enable-kvm -m 4096 \ -smp 4,maxcpus=8,sockets=1,cores=2,threads=4 \ -display none -nographic \ -drive file=disk1.qcow2,format=qcow2 ... (qemu) device_add host-spapr-cpu-core,core-id=4 [no interaction is possible after it, only SIGKILL to take the terminal back] Signed-off-by: Jose Ricardo Ziviani <joserz@linux.ibm.com> Reviewed-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-09-02 16:19:04 +02:00
/* Set compatibility mode to match the boot CPU, which was either set
* by the machine reset code or by CAS. This should never fail.
*/
ppc_set_compat(cpu, POWERPC_CPU(first_cpu)->compat_pvr, &error_abort);
env->spr[SPR_HIOR] = 0;
lpcr = env->spr[SPR_LPCR];
/* Set emulated LPCR to not send interrupts to hypervisor. Note that
* under KVM, the actual HW LPCR will be set differently by KVM itself,
* the settings below ensure proper operations with TCG in absence of
* a real hypervisor.
*
* Clearing VPM0 will also cause us to use RMOR in mmu-hash64.c for
* real mode accesses, which thankfully defaults to 0 and isn't
* accessible in guest mode.
*
* Disable Power-saving mode Exit Cause exceptions for the CPU, so
* we don't get spurious wakups before an RTAS start-cpu call.
*/
lpcr &= ~(LPCR_VPM0 | LPCR_VPM1 | LPCR_ISL | LPCR_KBV | pcc->lpcr_pm);
lpcr |= LPCR_LPES0 | LPCR_LPES1;
/* Set RMLS to the max (ie, 16G) */
lpcr &= ~LPCR_RMLS;
lpcr |= 1ull << LPCR_RMLS_SHIFT;
ppc_store_lpcr(cpu, lpcr);
/* Set a full AMOR so guest can use the AMR as it sees fit */
env->spr[SPR_AMOR] = 0xffffffffffffffffull;
spapr_cpu->vpa_addr = 0;
spapr_cpu->slb_shadow_addr = 0;
spapr_cpu->slb_shadow_size = 0;
spapr_cpu->dtl_addr = 0;
spapr_cpu->dtl_size = 0;
spapr_caps_cpu_apply(SPAPR_MACHINE(qdev_get_machine()), cpu);
2018-04-16 08:19:52 +02:00
kvm_check_mmu(cpu, &error_fatal);
}
void spapr_cpu_set_entry_state(PowerPCCPU *cpu, target_ulong nip, target_ulong r3)
{
PowerPCCPUClass *pcc = POWERPC_CPU_GET_CLASS(cpu);
CPUPPCState *env = &cpu->env;
env->nip = nip;
env->gpr[3] = r3;
kvmppc_set_reg_ppc_online(cpu, 1);
CPU(cpu)->halted = 0;
/* Enable Power-saving mode Exit Cause exceptions */
ppc_store_lpcr(cpu, env->spr[SPR_LPCR] | pcc->lpcr_pm);
}
/*
* Return the sPAPR CPU core type for @model which essentially is the CPU
* model specified with -cpu cmdline option.
*/
const char *spapr_get_cpu_core_type(const char *cpu_type)
{
int len = strlen(cpu_type) - strlen(POWERPC_CPU_TYPE_SUFFIX);
char *core_type = g_strdup_printf(SPAPR_CPU_CORE_TYPE_NAME("%.*s"),
len, cpu_type);
ObjectClass *oc = object_class_by_name(core_type);
g_free(core_type);
if (!oc) {
return NULL;
}
return object_class_get_name(oc);
}
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
static bool slb_shadow_needed(void *opaque)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuState *spapr_cpu = opaque;
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
return spapr_cpu->slb_shadow_addr != 0;
}
static const VMStateDescription vmstate_spapr_cpu_slb_shadow = {
.name = "spapr_cpu/vpa/slb_shadow",
.version_id = 1,
.minimum_version_id = 1,
.needed = slb_shadow_needed,
.fields = (VMStateField[]) {
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
VMSTATE_UINT64(slb_shadow_addr, SpaprCpuState),
VMSTATE_UINT64(slb_shadow_size, SpaprCpuState),
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
VMSTATE_END_OF_LIST()
}
};
static bool dtl_needed(void *opaque)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuState *spapr_cpu = opaque;
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
return spapr_cpu->dtl_addr != 0;
}
static const VMStateDescription vmstate_spapr_cpu_dtl = {
.name = "spapr_cpu/vpa/dtl",
.version_id = 1,
.minimum_version_id = 1,
.needed = dtl_needed,
.fields = (VMStateField[]) {
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
VMSTATE_UINT64(dtl_addr, SpaprCpuState),
VMSTATE_UINT64(dtl_size, SpaprCpuState),
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
VMSTATE_END_OF_LIST()
}
};
static bool vpa_needed(void *opaque)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuState *spapr_cpu = opaque;
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
return spapr_cpu->vpa_addr != 0;
}
static const VMStateDescription vmstate_spapr_cpu_vpa = {
.name = "spapr_cpu/vpa",
.version_id = 1,
.minimum_version_id = 1,
.needed = vpa_needed,
.fields = (VMStateField[]) {
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
VMSTATE_UINT64(vpa_addr, SpaprCpuState),
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
VMSTATE_END_OF_LIST()
},
.subsections = (const VMStateDescription * []) {
&vmstate_spapr_cpu_slb_shadow,
&vmstate_spapr_cpu_dtl,
NULL
}
};
static const VMStateDescription vmstate_spapr_cpu_state = {
.name = "spapr_cpu",
.version_id = 1,
.minimum_version_id = 1,
.fields = (VMStateField[]) {
VMSTATE_END_OF_LIST()
},
spapr_cpu_core: migrate VPA related state QEMU implements the "Shared Processor LPAR" (SPLPAR) option, which allows the hypervisor to time-slice a physical processor into multiple virtual processor. The intent is to allow more guests to run, and to optimize processor utilization. The guest OS can cede idle VCPUs, so that their processing capacity may be used by other VCPUs, with the H_CEDE hcall. The guest OS can also optimize spinlocks, by confering the time-slice of a spinning VCPU to the spinlock holder if it's currently notrunning, with the H_CONFER hcall. Both hcalls depend on a "Virtual Processor Area" (VPA) to be registered by the guest OS, generally during early boot. Other per-VCPU areas can be registered: the "SLB Shadow Buffer" which allows a more efficient dispatching of VCPUs, and the "Dispatch Trace Log Buffer" (DTL) which is used to compute time stolen by the hypervisor. Both DTL and SLB Shadow areas depend on the VPA to be registered. The VPA/SLB Shadow/DTL are state that QEMU should migrate, but this doesn't happen, for no apparent reason other than it was just never coded. This causes the features listed above to stop working after migration, and it breaks the logic of the H_REGISTER_VPA hcall in the destination. The VPA is set at the guest request, ie, we don't have to migrate it before the guest has actually set it. This patch hence adds an "spapr_cpu/vpa" subsection to the recently introduced per-CPU machine data migration stream. Since DTL and SLB Shadow are optional and both depend on VPA, they get their own subsections "spapr_cpu/vpa/slb_shadow" and "spapr_cpu/vpa/dtl" hanging from the "spapr_cpu/vpa" subsection. Note that this won't break migration to older QEMUs. Is is already handled by only registering the vmstate handler for per-CPU data with newer machine types. Signed-off-by: Greg Kurz <groug@kaod.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-18 14:26:49 +02:00
.subsections = (const VMStateDescription * []) {
&vmstate_spapr_cpu_vpa,
NULL
}
};
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
static void spapr_unrealize_vcpu(PowerPCCPU *cpu, SpaprCpuCore *sc)
{
if (!sc->pre_3_0_migration) {
vmstate_unregister(NULL, &vmstate_spapr_cpu_state, cpu->machine_data);
}
qemu_unregister_reset(spapr_cpu_reset, cpu);
if (spapr_cpu_state(cpu)->icp) {
object_unparent(OBJECT(spapr_cpu_state(cpu)->icp));
}
if (spapr_cpu_state(cpu)->tctx) {
object_unparent(OBJECT(spapr_cpu_state(cpu)->tctx));
}
cpu_remove_sync(CPU(cpu));
object_unparent(OBJECT(cpu));
}
static void spapr_cpu_core_unrealize(DeviceState *dev, Error **errp)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuCore *sc = SPAPR_CPU_CORE(OBJECT(dev));
CPUCore *cc = CPU_CORE(dev);
int i;
for (i = 0; i < cc->nr_threads; i++) {
spapr_unrealize_vcpu(sc->threads[i], sc);
}
g_free(sc->threads);
}
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
static void spapr_realize_vcpu(PowerPCCPU *cpu, SpaprMachineState *spapr,
SpaprCpuCore *sc, Error **errp)
{
CPUPPCState *env = &cpu->env;
CPUState *cs = CPU(cpu);
Error *local_err = NULL;
object_property_set_bool(OBJECT(cpu), true, "realized", &local_err);
if (local_err) {
goto error;
}
/* Set time-base frequency to 512 MHz */
cpu_ppc_tb_init(env, SPAPR_TIMEBASE_FREQ);
cpu_ppc_set_vhyp(cpu, PPC_VIRTUAL_HYPERVISOR(spapr));
kvmppc_set_papr(cpu);
qemu_register_reset(spapr_cpu_reset, cpu);
spapr_cpu_reset(cpu);
spapr->irq->cpu_intc_create(spapr, cpu, &local_err);
if (local_err) {
spapr_cpu_core: add missing rollback on realization path The spapr_realize_vcpu() function doesn't rollback in case of error. This isn't a problem with coldplugged CPUs because the machine won't start and QEMU will exit. Hotplug is a different story though: the CPU thread is started under object_property_set_bool() and it assumes it can access the CPU object. If icp_create() fails, we return an error without unregistering the reset handler for this CPU, and we let the underlying QEMU thread for this CPU alive. Since spapr_cpu_core_realize() doesn't care to unrealize already realized CPUs either, but happily frees all of them anyway, the CPU thread crashes instantly: (qemu) device_add host-spapr-cpu-core,core-id=1,id=gku GKU: failing icp_create (cpu 0x11497fd0) ^^^^^^^^^^ Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffee3feaa0 (LWP 24725)] 0x00000000104c8374 in object_dynamic_cast_assert (obj=0x11497fd0, ^^^^^^^^^^^^^^ pointer to the CPU object 623 trace_object_dynamic_cast_assert(obj ? obj->class->type->name (gdb) p obj->class->type $1 = (Type) 0x0 (gdb) p * obj $2 = {class = 0x10ea9c10, free = 0x11244620, ^^^^^^^^^^ should be g_free (gdb) p g_free $3 = {<text variable, no debug info>} 0x7ffff282bef0 <g_free> obj is a dangling pointer to the CPU that was just destroyed in spapr_cpu_core_realize(). This patch adds proper rollback to both spapr_realize_vcpu() and spapr_cpu_core_realize(). Signed-off-by: Greg Kurz <groug@kaod.org> [dwg: Fixed a conflict due to a change in my tree] Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-14 23:50:42 +02:00
goto error_unregister;
}
if (!sc->pre_3_0_migration) {
vmstate_register(NULL, cs->cpu_index, &vmstate_spapr_cpu_state,
cpu->machine_data);
}
return;
spapr_cpu_core: add missing rollback on realization path The spapr_realize_vcpu() function doesn't rollback in case of error. This isn't a problem with coldplugged CPUs because the machine won't start and QEMU will exit. Hotplug is a different story though: the CPU thread is started under object_property_set_bool() and it assumes it can access the CPU object. If icp_create() fails, we return an error without unregistering the reset handler for this CPU, and we let the underlying QEMU thread for this CPU alive. Since spapr_cpu_core_realize() doesn't care to unrealize already realized CPUs either, but happily frees all of them anyway, the CPU thread crashes instantly: (qemu) device_add host-spapr-cpu-core,core-id=1,id=gku GKU: failing icp_create (cpu 0x11497fd0) ^^^^^^^^^^ Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffee3feaa0 (LWP 24725)] 0x00000000104c8374 in object_dynamic_cast_assert (obj=0x11497fd0, ^^^^^^^^^^^^^^ pointer to the CPU object 623 trace_object_dynamic_cast_assert(obj ? obj->class->type->name (gdb) p obj->class->type $1 = (Type) 0x0 (gdb) p * obj $2 = {class = 0x10ea9c10, free = 0x11244620, ^^^^^^^^^^ should be g_free (gdb) p g_free $3 = {<text variable, no debug info>} 0x7ffff282bef0 <g_free> obj is a dangling pointer to the CPU that was just destroyed in spapr_cpu_core_realize(). This patch adds proper rollback to both spapr_realize_vcpu() and spapr_cpu_core_realize(). Signed-off-by: Greg Kurz <groug@kaod.org> [dwg: Fixed a conflict due to a change in my tree] Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-14 23:50:42 +02:00
error_unregister:
qemu_unregister_reset(spapr_cpu_reset, cpu);
cpu_remove_sync(CPU(cpu));
error:
error_propagate(errp, local_err);
}
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
static PowerPCCPU *spapr_create_vcpu(SpaprCpuCore *sc, int i, Error **errp)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuCoreClass *scc = SPAPR_CPU_CORE_GET_CLASS(sc);
CPUCore *cc = CPU_CORE(sc);
Object *obj;
char *id;
CPUState *cs;
PowerPCCPU *cpu;
Error *local_err = NULL;
obj = object_new(scc->cpu_type);
cs = CPU(obj);
cpu = POWERPC_CPU(obj);
cs->cpu_index = cc->core_id + i;
spapr_set_vcpu_id(cpu, cs->cpu_index, &local_err);
if (local_err) {
goto err;
}
cpu->node_id = sc->node_id;
id = g_strdup_printf("thread[%d]", i);
object_property_add_child(OBJECT(sc), id, obj, &local_err);
g_free(id);
if (local_err) {
goto err;
}
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
cpu->machine_data = g_new0(SpaprCpuState, 1);
object_unref(obj);
return cpu;
err:
object_unref(obj);
error_propagate(errp, local_err);
return NULL;
}
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
static void spapr_delete_vcpu(PowerPCCPU *cpu, SpaprCpuCore *sc)
{
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuState *spapr_cpu = spapr_cpu_state(cpu);
cpu->machine_data = NULL;
g_free(spapr_cpu);
object_unparent(OBJECT(cpu));
}
static void spapr_cpu_core_realize(DeviceState *dev, Error **errp)
{
/* We don't use SPAPR_MACHINE() in order to exit gracefully if the user
* tries to add a sPAPR CPU core to a non-pseries machine.
*/
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprMachineState *spapr =
(SpaprMachineState *) object_dynamic_cast(qdev_get_machine(),
TYPE_SPAPR_MACHINE);
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuCore *sc = SPAPR_CPU_CORE(OBJECT(dev));
CPUCore *cc = CPU_CORE(OBJECT(dev));
Error *local_err = NULL;
int i, j;
if (!spapr) {
error_setg(errp, TYPE_SPAPR_CPU_CORE " needs a pseries machine");
return;
}
sc->threads = g_new(PowerPCCPU *, cc->nr_threads);
for (i = 0; i < cc->nr_threads; i++) {
sc->threads[i] = spapr_create_vcpu(sc, i, &local_err);
if (local_err) {
goto err;
}
}
for (j = 0; j < cc->nr_threads; j++) {
spapr_realize_vcpu(sc->threads[j], spapr, sc, &local_err);
if (local_err) {
spapr_cpu_core: add missing rollback on realization path The spapr_realize_vcpu() function doesn't rollback in case of error. This isn't a problem with coldplugged CPUs because the machine won't start and QEMU will exit. Hotplug is a different story though: the CPU thread is started under object_property_set_bool() and it assumes it can access the CPU object. If icp_create() fails, we return an error without unregistering the reset handler for this CPU, and we let the underlying QEMU thread for this CPU alive. Since spapr_cpu_core_realize() doesn't care to unrealize already realized CPUs either, but happily frees all of them anyway, the CPU thread crashes instantly: (qemu) device_add host-spapr-cpu-core,core-id=1,id=gku GKU: failing icp_create (cpu 0x11497fd0) ^^^^^^^^^^ Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffee3feaa0 (LWP 24725)] 0x00000000104c8374 in object_dynamic_cast_assert (obj=0x11497fd0, ^^^^^^^^^^^^^^ pointer to the CPU object 623 trace_object_dynamic_cast_assert(obj ? obj->class->type->name (gdb) p obj->class->type $1 = (Type) 0x0 (gdb) p * obj $2 = {class = 0x10ea9c10, free = 0x11244620, ^^^^^^^^^^ should be g_free (gdb) p g_free $3 = {<text variable, no debug info>} 0x7ffff282bef0 <g_free> obj is a dangling pointer to the CPU that was just destroyed in spapr_cpu_core_realize(). This patch adds proper rollback to both spapr_realize_vcpu() and spapr_cpu_core_realize(). Signed-off-by: Greg Kurz <groug@kaod.org> [dwg: Fixed a conflict due to a change in my tree] Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-14 23:50:42 +02:00
goto err_unrealize;
}
}
return;
spapr_cpu_core: add missing rollback on realization path The spapr_realize_vcpu() function doesn't rollback in case of error. This isn't a problem with coldplugged CPUs because the machine won't start and QEMU will exit. Hotplug is a different story though: the CPU thread is started under object_property_set_bool() and it assumes it can access the CPU object. If icp_create() fails, we return an error without unregistering the reset handler for this CPU, and we let the underlying QEMU thread for this CPU alive. Since spapr_cpu_core_realize() doesn't care to unrealize already realized CPUs either, but happily frees all of them anyway, the CPU thread crashes instantly: (qemu) device_add host-spapr-cpu-core,core-id=1,id=gku GKU: failing icp_create (cpu 0x11497fd0) ^^^^^^^^^^ Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffee3feaa0 (LWP 24725)] 0x00000000104c8374 in object_dynamic_cast_assert (obj=0x11497fd0, ^^^^^^^^^^^^^^ pointer to the CPU object 623 trace_object_dynamic_cast_assert(obj ? obj->class->type->name (gdb) p obj->class->type $1 = (Type) 0x0 (gdb) p * obj $2 = {class = 0x10ea9c10, free = 0x11244620, ^^^^^^^^^^ should be g_free (gdb) p g_free $3 = {<text variable, no debug info>} 0x7ffff282bef0 <g_free> obj is a dangling pointer to the CPU that was just destroyed in spapr_cpu_core_realize(). This patch adds proper rollback to both spapr_realize_vcpu() and spapr_cpu_core_realize(). Signed-off-by: Greg Kurz <groug@kaod.org> [dwg: Fixed a conflict due to a change in my tree] Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-14 23:50:42 +02:00
err_unrealize:
while (--j >= 0) {
spapr_unrealize_vcpu(sc->threads[j], sc);
spapr_cpu_core: add missing rollback on realization path The spapr_realize_vcpu() function doesn't rollback in case of error. This isn't a problem with coldplugged CPUs because the machine won't start and QEMU will exit. Hotplug is a different story though: the CPU thread is started under object_property_set_bool() and it assumes it can access the CPU object. If icp_create() fails, we return an error without unregistering the reset handler for this CPU, and we let the underlying QEMU thread for this CPU alive. Since spapr_cpu_core_realize() doesn't care to unrealize already realized CPUs either, but happily frees all of them anyway, the CPU thread crashes instantly: (qemu) device_add host-spapr-cpu-core,core-id=1,id=gku GKU: failing icp_create (cpu 0x11497fd0) ^^^^^^^^^^ Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffee3feaa0 (LWP 24725)] 0x00000000104c8374 in object_dynamic_cast_assert (obj=0x11497fd0, ^^^^^^^^^^^^^^ pointer to the CPU object 623 trace_object_dynamic_cast_assert(obj ? obj->class->type->name (gdb) p obj->class->type $1 = (Type) 0x0 (gdb) p * obj $2 = {class = 0x10ea9c10, free = 0x11244620, ^^^^^^^^^^ should be g_free (gdb) p g_free $3 = {<text variable, no debug info>} 0x7ffff282bef0 <g_free> obj is a dangling pointer to the CPU that was just destroyed in spapr_cpu_core_realize(). This patch adds proper rollback to both spapr_realize_vcpu() and spapr_cpu_core_realize(). Signed-off-by: Greg Kurz <groug@kaod.org> [dwg: Fixed a conflict due to a change in my tree] Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2018-06-14 23:50:42 +02:00
}
err:
while (--i >= 0) {
spapr_delete_vcpu(sc->threads[i], sc);
}
g_free(sc->threads);
error_propagate(errp, local_err);
}
static Property spapr_cpu_core_properties[] = {
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
DEFINE_PROP_INT32("node-id", SpaprCpuCore, node_id, CPU_UNSET_NUMA_NODE_ID),
DEFINE_PROP_BOOL("pre-3.0-migration", SpaprCpuCore, pre_3_0_migration,
false),
DEFINE_PROP_END_OF_LIST()
};
static void spapr_cpu_core_class_init(ObjectClass *oc, void *data)
{
DeviceClass *dc = DEVICE_CLASS(oc);
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
SpaprCpuCoreClass *scc = SPAPR_CPU_CORE_CLASS(oc);
dc->realize = spapr_cpu_core_realize;
dc->unrealize = spapr_cpu_core_unrealize;
dc->props = spapr_cpu_core_properties;
scc->cpu_type = data;
}
#define DEFINE_SPAPR_CPU_CORE_TYPE(cpu_model) \
{ \
.parent = TYPE_SPAPR_CPU_CORE, \
.class_data = (void *) POWERPC_CPU_TYPE_NAME(cpu_model), \
.class_init = spapr_cpu_core_class_init, \
.name = SPAPR_CPU_CORE_TYPE_NAME(cpu_model), \
}
static const TypeInfo spapr_cpu_core_type_infos[] = {
{
.name = TYPE_SPAPR_CPU_CORE,
.parent = TYPE_CPU_CORE,
.abstract = true,
spapr: Use CamelCase properly The qemu coding standard is to use CamelCase for type and structure names, and the pseries code follows that... sort of. There are quite a lot of places where we bend the rules in order to preserve the capitalization of internal acronyms like "PHB", "TCE", "DIMM" and most commonly "sPAPR". That was a bad idea - it frequently leads to names ending up with hard to read clusters of capital letters, and means they don't catch the eye as type identifiers, which is kind of the point of the CamelCase convention in the first place. In short, keeping type identifiers look like CamelCase is more important than preserving standard capitalization of internal "words". So, this patch renames a heap of spapr internal type names to a more standard CamelCase. In addition to case changes, we also make some other identifier renames: VIOsPAPR* -> SpaprVio* The reverse word ordering was only ever used to mitigate the capital cluster, so revert to the natural ordering. VIOsPAPRVTYDevice -> SpaprVioVty VIOsPAPRVLANDevice -> SpaprVioVlan Brevity, since the "Device" didn't add useful information sPAPRDRConnector -> SpaprDrc sPAPRDRConnectorClass -> SpaprDrcClass Brevity, and makes it clearer this is the same thing as a "DRC" mentioned in many other places in the code This is 100% a mechanical search-and-replace patch. It will, however, conflict with essentially any and all outstanding patches touching the spapr code. Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
2019-03-06 05:35:37 +01:00
.instance_size = sizeof(SpaprCpuCore),
.class_size = sizeof(SpaprCpuCoreClass),
},
DEFINE_SPAPR_CPU_CORE_TYPE("970_v2.2"),
DEFINE_SPAPR_CPU_CORE_TYPE("970mp_v1.0"),
DEFINE_SPAPR_CPU_CORE_TYPE("970mp_v1.1"),
DEFINE_SPAPR_CPU_CORE_TYPE("power5+_v2.1"),
DEFINE_SPAPR_CPU_CORE_TYPE("power7_v2.3"),
DEFINE_SPAPR_CPU_CORE_TYPE("power7+_v2.1"),
DEFINE_SPAPR_CPU_CORE_TYPE("power8_v2.0"),
DEFINE_SPAPR_CPU_CORE_TYPE("power8e_v2.1"),
DEFINE_SPAPR_CPU_CORE_TYPE("power8nvl_v1.0"),
DEFINE_SPAPR_CPU_CORE_TYPE("power9_v1.0"),
DEFINE_SPAPR_CPU_CORE_TYPE("power9_v2.0"),
#ifdef CONFIG_KVM
DEFINE_SPAPR_CPU_CORE_TYPE("host"),
#endif
};
DEFINE_TYPES(spapr_cpu_core_type_infos)