Re: [Qemu-devel] [PATCH v2 8/9] spapr: Implement processor compatibility in ibm, client-architecture-support

[Alexander Graf <agraf@suse.de>]

On 21.05.14 11:36, Alexey Kardashevskiy wrote:
> On 05/21/2014 07:21 PM, Alexander Graf wrote:
>> On 21.05.14 10:27, Alexey Kardashevskiy wrote:
>>> Modern Linux kernels support last POWERPC CPUs so when a kernel boots,
>>> in most cases it can find a matching cpu_spec in the kernel's cpu_specs
>>> list. However if the kernel is quite old, it may be missing a definition
>>> of the actual CPU. To provide an ability for old kernels to work on modern
>>> hardware, a Processor Compatibility Mode has been introduced
>>> by the PowerISA specification.
>>>
>>>   From the hardware prospective, it is supported by the Processor
>>> Compatibility Register (PCR) which is defined in PowerISA. The register
>>> enables one of the compatibility modes (2.05/2.06/2.07).
>>> Since PCR is a hypervisor privileged register and cannot be
>>> directly accessed from the guest, the mode selection is done via
>>> ibm,client-architecture-support (CAS) RTAS call using which the guest
>>> specifies what "raw" and "architected" CPU versions it supports.
>>> QEMU works out the best match, changes a "cpu-version" property of
>>> every CPU and notifies the guest about the change by setting these
>>> properties in the buffer passed as a response on a custom H_CAS hypercall.
>>>
>>> This implements ibm,client-architecture-support parameters parsing
>>> (now only for PVRs) and cooks the device tree diff with new values for
>>> "cpu-version", "ibm,ppc-interrupt-server#s" and
>>> "ibm,ppc-interrupt-server#s" properties.
>>>
>>> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
>>> ---
>>>    hw/ppc/spapr.c       |  4 ++-
>>>    hw/ppc/spapr_hcall.c | 85
>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>    trace-events         |  4 +++
>>>    3 files changed, 92 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
>>> index 14c72d9..020426a 100644
>>> --- a/hw/ppc/spapr.c
>>> +++ b/hw/ppc/spapr.c
>>> @@ -34,6 +34,7 @@
>>>    #include "sysemu/kvm.h"
>>>    #include "kvm_ppc.h"
>>>    #include "mmu-hash64.h"
>>> +#include "qom/cpu.h"
>>>      #include "hw/boards.h"
>>>    #include "hw/ppc/ppc.h"
>>> @@ -601,7 +602,8 @@ int spapr_h_cas_compose_response(target_ulong addr,
>>> target_ulong size)
>>>        _FDT((fdt_open_into(fdt_skel, fdt, size)));
>>>        g_free(fdt_skel);
>>>    -    /* Place to make changes to the tree */
>>> +    /* Fix skeleton up */
>>> +    _FDT((spapr_fixup_cpu_dt(fdt, spapr)));
>>>          /* Pack resulting tree */
>>>        _FDT((fdt_pack(fdt)));
>>> diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
>>> index 2f6aa5c..2e0a49c 100644
>>> --- a/hw/ppc/spapr_hcall.c
>>> +++ b/hw/ppc/spapr_hcall.c
>>> @@ -3,6 +3,9 @@
>>>    #include "helper_regs.h"
>>>    #include "hw/ppc/spapr.h"
>>>    #include "mmu-hash64.h"
>>> +#include "cpu-models.h"
>>> +#include "trace.h"
>>> +#include "kvm_ppc.h"
>>>      struct SPRSyncState {
>>>        CPUState *cs;
>>> @@ -752,12 +755,94 @@ out:
>>>        return ret;
>>>    }
>>>    +#define get_compat_level(cpuver) ( \
>>> +    ((cpuver) == CPU_POWERPC_LOGICAL_2_05) ? 2050 : \
>>> +    ((cpuver) == CPU_POWERPC_LOGICAL_2_06) ? 2060 : \
>>> +    ((cpuver) == CPU_POWERPC_LOGICAL_2_06_PLUS) ? 2061 : \
>>> +    ((cpuver) == CPU_POWERPC_LOGICAL_2_07) ? 2070 : 0)
>>> +
>>>    static target_ulong h_client_architecture_support(PowerPCCPU *cpu_,
>>>                                                      sPAPREnvironment *spapr,
>>>                                                      target_ulong opcode,
>>>                                                      target_ulong *args)
>>>    {
>>>        target_ulong list = args[0];
>>> +    PowerPCCPUClass *pcc_ = POWERPC_CPU_GET_CLASS(cpu_);
>>> +    CPUState *cs;
>>> +    bool cpu_match = false;
>>> +    unsigned old_cpu_version = cpu_->cpu_version;
>>> +    unsigned compat_lvl = 0, cpu_version = 0;
>>> +    unsigned max_lvl = get_compat_level(cpu_->max_compat);
>>> +
>>> +    /* Parse PVR list */
>>> +    for ( ; ; ) {
>>> +        uint32_t pvr, pvr_mask;
>>> +
>>> +        pvr_mask = rtas_ld(list, 0);
>>> +        list += 4;
>>> +        pvr = rtas_ld(list, 0);
>>> +        list += 4;
>>> +
>>> +        trace_spapr_cas_pvr_try(pvr);
>>> +        if (!max_lvl &&
>>> +            ((cpu_->env.spr[SPR_PVR] & pvr_mask) == (pvr & pvr_mask))) {
>>> +            cpu_match = true;
>>> +            cpu_version = 0;
>>> +        } else if (pvr == cpu_->cpu_version) {
>>> +            cpu_match = true;
>>> +            cpu_version = cpu_->cpu_version;
>>> +        } else if (!cpu_match) {
>>> +            /* If it is a logical PVR, try to determine the highest
>>> level */
>>> +            unsigned lvl = get_compat_level(pvr);
>>> +            if (lvl) {
>>> +                bool is205 = (pcc_->pcr_mask & PCR_COMPAT_2_05) &&
>>> +                     (lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_05));
>>> +                bool is206 = (pcc_->pcr_mask & PCR_COMPAT_2_06) &&
>>> +                    ((lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_06)) ||
>>> +                    (lvl ==
>>> get_compat_level(CPU_POWERPC_LOGICAL_2_06_PLUS)));
>>> +
>>> +                if (is205 || is206) {
>>> +                    if (!max_lvl) {
>>> +                        /* User did not set the level, choose the
>>> highest */
>>> +                        if (compat_lvl <= lvl) {
>>> +                            compat_lvl = lvl;
>>> +                            cpu_version = pvr;
>>> +                        }
>>> +                    } else if (max_lvl >= lvl) {
>>> +                        /* User chose the level, don't set higher than
>>> this */
>>> +                        compat_lvl = lvl;
>>> +                        cpu_version = pvr;
>>> +                    }
>>> +                }
>>> +            }
>>> +        }
>>> +        /* Terminator record */
>>> +        if (~pvr_mask & pvr) {
>> This loop can be used by the guest to stall QEMU for a long period of time,
>> no? Better add a safety net check somewhere to allow for early abort.
> Like what? No more than 64 PVRs would be enough?

Yeah, choose a random (high) number. If you like make it 512 to be safe 
for a few generations :).


Alex