Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
Il 18/11/2013 04:02, Alexey Kardashevskiy ha scritto: On 11/15/2013 09:40 PM, Paolo Bonzini wrote: Il 15/11/2013 06:14, Alexey Kardashevskiy ha scritto: It does not feel that we really need CPUState::kvm_cpu_id and direct calling of kvm_arch_vcpu_id() would be enough. Indeed -- and it should be kvm_ppc_vcpu_id() since other architectures do not need it. And ignore kvm_arch_vcpu_id() for spapr/ppc? Sorry, brain fart - i meant ppc_get_vcpu_dt_id but it has already been objected to earlier. So the patch is ok for me, modulo removal of kvm_cpu_id which you already proposed above. Paolo
Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
On 11/15/2013 09:40 PM, Paolo Bonzini wrote: Il 15/11/2013 06:14, Alexey Kardashevskiy ha scritto: It does not feel that we really need CPUState::kvm_cpu_id and direct calling of kvm_arch_vcpu_id() would be enough. Indeed -- and it should be kvm_ppc_vcpu_id() since other architectures do not need it. And ignore kvm_arch_vcpu_id() for spapr/ppc? -- Alexey
Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
Il 15/11/2013 06:14, Alexey Kardashevskiy ha scritto: It does not feel that we really need CPUState::kvm_cpu_id and direct calling of kvm_arch_vcpu_id() would be enough. Indeed -- and it should be kvm_ppc_vcpu_id() since other architectures do not need it. Paolo
Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
Am 15.11.2013 06:14, schrieb Alexey Kardashevskiy: diff --git a/include/qom/cpu.h b/include/qom/cpu.h index 7739e00..52fc76d 100644 --- a/include/qom/cpu.h +++ b/include/qom/cpu.h @@ -197,6 +197,7 @@ struct CPUState { bool kvm_vcpu_dirty; struct KVMState *kvm_state; struct kvm_run *kvm_run; +int kvm_cpu_id; /* TODO Move common fields from CPUArchState here. */ int cpu_index; /* used by alpha TCG */ Here you are adding a field to CPUState, fine with me. (Please add a documentation line above the struct then.) diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h index bb84767..8052f6b 100644 --- a/target-ppc/cpu.h +++ b/target-ppc/cpu.h @@ -1074,6 +1074,9 @@ struct CPUPPCState { */ uint8_t fit_period[4]; uint8_t wdt_period[4]; + +/* The CPU index used in the device tree. KVM uses this index too */ +int cpu_dt_id; But I believe I have requested a number of times not to add random fields to CPUPPCState unless they are accessed by TCG. Please place the new field in PowerPCCPU instead and put the description into the struct documentation. }; #define SET_FIT_PERIOD(a_, b_, c_, d_) \ @@ -2154,4 +2157,8 @@ static inline bool cpu_has_work(CPUState *cpu) void dump_mmu(FILE *f, fprintf_function cpu_fprintf, CPUPPCState *env); +int ppc_get_vcpu_dt_id(CPUState *cs); +CPUState *ppc_get_vcpu_by_dt_id(int cpu_dt_id); +void ppc_fixup_cpu_dt_id(PowerPCCPU *cpu); + #endif /* !defined (__CPU_PPC_H__) */ Regards, Andreas -- SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
On 16.11.2013 2:11, Andreas Färber wrote: Am 15.11.2013 06:14, schrieb Alexey Kardashevskiy: diff --git a/include/qom/cpu.h b/include/qom/cpu.h index 7739e00..52fc76d 100644 --- a/include/qom/cpu.h +++ b/include/qom/cpu.h @@ -197,6 +197,7 @@ struct CPUState { bool kvm_vcpu_dirty; struct KVMState *kvm_state; struct kvm_run *kvm_run; +int kvm_cpu_id; /* TODO Move common fields from CPUArchState here. */ int cpu_index; /* used by alpha TCG */ Here you are adding a field to CPUState, fine with me. (Please add a documentation line above the struct then.) diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h index bb84767..8052f6b 100644 --- a/target-ppc/cpu.h +++ b/target-ppc/cpu.h @@ -1074,6 +1074,9 @@ struct CPUPPCState { */ uint8_t fit_period[4]; uint8_t wdt_period[4]; + +/* The CPU index used in the device tree. KVM uses this index too */ +int cpu_dt_id; But I believe I have requested a number of times not to add random fields to CPUPPCState unless they are accessed by TCG. Please place the new field in PowerPCCPU instead and put the description into the struct documentation. It is accessed by xics.c which is used in TCG. Or I misinterpret the rule, do not I? -- With best regards Alexey Kardashevskiy -- icq: 52150396
Re: [Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
Am 15.11.2013 18:02, schrieb Alexey Kardashevskiy: On 16.11.2013 2:11, Andreas Färber wrote: Am 15.11.2013 06:14, schrieb Alexey Kardashevskiy: diff --git a/include/qom/cpu.h b/include/qom/cpu.h index 7739e00..52fc76d 100644 --- a/include/qom/cpu.h +++ b/include/qom/cpu.h @@ -197,6 +197,7 @@ struct CPUState { bool kvm_vcpu_dirty; struct KVMState *kvm_state; struct kvm_run *kvm_run; +int kvm_cpu_id; /* TODO Move common fields from CPUArchState here. */ int cpu_index; /* used by alpha TCG */ Here you are adding a field to CPUState, fine with me. (Please add a documentation line above the struct then.) diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h index bb84767..8052f6b 100644 --- a/target-ppc/cpu.h +++ b/target-ppc/cpu.h @@ -1074,6 +1074,9 @@ struct CPUPPCState { */ uint8_t fit_period[4]; uint8_t wdt_period[4]; + +/* The CPU index used in the device tree. KVM uses this index too */ +int cpu_dt_id; But I believe I have requested a number of times not to add random fields to CPUPPCState unless they are accessed by TCG. Please place the new field in PowerPCCPU instead and put the description into the struct documentation. It is accessed by xics.c which is used in TCG. Or I misinterpret the rule, do not I? Yes, you do. ;) The question is, is the field accessed using offsetof(CPUPPCState, x) such as in target-ppc/translate.c (search for TCG_AREG0). That's what CPUPPCState is still needed for - mainly GPRs and the TLBs that TCG ops load from / store to using host instructions with immediate offset. Whether it is used in a HELPER() / helper_* function called from TCG (target-ppc/*helper.c) has become less relevant since the optimization of ppc_env_get_cpu() macro. Neither of that seem to be the case here, unless I missed something. I still hope we can one day get rid of CPUPPCState by having only numeric-offset accesses from cpu+sizeof(CPUState) and all fields (including TLB, as Paolo recently pointed out) directly in PowerPCCPU. Regards, Andreas -- SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
[Qemu-devel] [PATCH v4] ppc: introduce CPUPPCState::cpu_dt_id and CPUState::kvm_cpu_id
Normally CPUState::cpu_index is used to pick the right CPU for various operations. However default consecutive numbering does not always work for POWERPC. For example, on POWER7 (which supports 4 threads per core), -smp 8,threads=4 should create CPUs with indexes 0,1,2,3,4,5,6,7 and -smp 8,threads=1 should create CPUs with indexes 0,4,8,12,16,20,24,28. These indexes are reflected in /proc/device-tree/cpus/PowerPC,POWER7@XX and used to call KVM VCPU's ioctls. In order to achieve this, kvmppc_fixup_cpu() was introduced. Roughly speaking, it multiplies cpu_index by the number of threads per core. This approach has disadvantages such as: 1. NUMA configuration stays broken after the fixup; 2. CPU-related commands from QEMU Monitor do not work properly as the accept fixed CPU indexes and the user does not really know what they are after fixup as the number of threads per core changes between CPU versions and via QEMU command line. This introduces a @cpu_dt_id field in the CPUPPCState struct which is set from @cpu_index by default but can be fixed later to conform the device tree requirements. 1. int ppc_get_vcpu_dt_id(CPUState *cs) - returns a device-tree ID for a CPU; 2. CPUState *ppc_get_vcpu_by_dt_id(int cpu_dt_id) - finds CPUState by a device-tree CPU ID. 3. void ppc_fixup_cpu_dt_id(PowerPCCPU *cpu) - calculates correct cpu_dt_id; this replaces kvmppc_fixup_cpu() as the numbers are not just about KVM but about the device tree too. This uses the new functions to: 1. fix emulated XICS hypercall handlers as they receive fixed CPU indexes; 2. fix XICS-KVM to enable in-kernel XICS on right CPU; 3. compose correct device-tree. This removes @cpu_index fixup as @cpu_dt_id is used instead so QEMU monitor can accept command-line CPU indexes again. Cc: Paul Mackerras pau...@samba.org Cc: Benjamin Herrenschmidt b...@kernel.crashing.org Cc: Paolo Bonzini pbonz...@redhat.com Cc: Bharat Bhushan bharat.bhus...@freescale.com Cc: Scott Wood scottw...@freescale.com Signed-off-by: Alexey Kardashevskiy a...@ozlabs.ru --- It does not feel that we really need CPUState::kvm_cpu_id and direct calling of kvm_arch_vcpu_id() would be enough. Please comment. Thanks. --- Changes: v4: * now there are cpu_index, cpu_dt_id and kvm_cpu_id * kvmppc_fixup_cpu() was replaced with ppc_fixup_cpu_dt_id() v3: * althouth e500 does not tweak CPU indexes, e500 and openpic-kvm are fixed too; hopefully it does not break anything. v2: * added PPC-specific ppc_get_vcpu_dt_id() and ppc_get_vcpu_by_dt_id() * fixed kvm_arch_vcpu_id() to use ppc_get_vcpu_dt_id() * fixed emulated XICS * removed kvm_arch_vcpu_id() stub for non-KVM case --- hw/intc/openpic_kvm.c | 2 +- hw/intc/xics.c | 15 +-- hw/intc/xics_kvm.c | 8 hw/ppc/e500.c | 5 +++-- hw/ppc/ppc.c| 37 + hw/ppc/spapr.c | 9 + hw/ppc/spapr_hcall.c| 2 +- hw/ppc/spapr_rtas.c | 4 ++-- include/qom/cpu.h | 1 + kvm-all.c | 1 + target-ppc/cpu.h| 7 +++ target-ppc/kvm.c| 10 +- target-ppc/translate_init.c | 4 13 files changed, 80 insertions(+), 25 deletions(-) diff --git a/hw/intc/openpic_kvm.c b/hw/intc/openpic_kvm.c index c7f7b84..75d5897 100644 --- a/hw/intc/openpic_kvm.c +++ b/hw/intc/openpic_kvm.c @@ -228,7 +228,7 @@ int kvm_openpic_connect_vcpu(DeviceState *d, CPUState *cs) encap.cap = KVM_CAP_IRQ_MPIC; encap.args[0] = opp-fd; -encap.args[1] = cs-cpu_index; +encap.args[1] = cs-kvm_cpu_id; return kvm_vcpu_ioctl(cs, KVM_ENABLE_CAP, encap); } diff --git a/hw/intc/xics.c b/hw/intc/xics.c index a05..866ee08 100644 --- a/hw/intc/xics.c +++ b/hw/intc/xics.c @@ -33,6 +33,17 @@ #include qemu/error-report.h #include qapi/visitor.h +static int get_cpu_index_by_dt_id(int cpu_dt_id) +{ +CPUState *cs = ppc_get_vcpu_by_dt_id(cpu_dt_id); + +if (cs) { +return cs-cpu_index; +} + +return -1; +} + void xics_cpu_setup(XICSState *icp, PowerPCCPU *cpu) { CPUState *cs = CPU(cpu); @@ -659,7 +670,7 @@ static target_ulong h_cppr(PowerPCCPU *cpu, sPAPREnvironment *spapr, static target_ulong h_ipi(PowerPCCPU *cpu, sPAPREnvironment *spapr, target_ulong opcode, target_ulong *args) { -target_ulong server = args[0]; +target_ulong server = get_cpu_index_by_dt_id(args[0]); target_ulong mfrr = args[1]; if (server = spapr-icp-nr_servers) { @@ -728,7 +739,7 @@ static void rtas_set_xive(PowerPCCPU *cpu, sPAPREnvironment *spapr, } nr = rtas_ld(args, 0); -server = rtas_ld(args, 1); +server = get_cpu_index_by_dt_id(rtas_ld(args, 1)); priority = rtas_ld(args, 2); if (!ics_valid_irq(ics, nr) || (server = ics-icp-nr_servers) diff --git a/hw/intc/xics_kvm.c b/hw/intc/xics_kvm.c index c203646..e3556ba 100644 --- a/hw/intc/xics_kvm.c +++ b/hw/intc/xics_kvm.c @@