[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-ppc] [RFC 10/17] pseries: Rewrite CAS PVR compatibility logic
From: |
David Gibson |
Subject: |
Re: [Qemu-ppc] [RFC 10/17] pseries: Rewrite CAS PVR compatibility logic |
Date: |
Mon, 31 Oct 2016 16:44:19 +1100 |
User-agent: |
Mutt/1.7.1 (2016-10-04) |
On Mon, Oct 31, 2016 at 04:00:14PM +1100, Alexey Kardashevskiy wrote:
> On 30/10/16 22:12, David Gibson wrote:
> > During boot, PAPR guests negotiate CPU model support with the
> > ibm,client-architecture-support mechanism. The logic to implement this in
> > qemu is very convoluted. This cleans it up to be cleaner, using the new
> > ppc_check_compat() call.
> >
> > The new logic for choosing a compatibility mode is:
> > 1. If the guest lists the CPU's real PVR as supported *AND* no
> > maximum compatibility mode has been requested on the command line
> > then we use "raw" mode - the CPU acts with full capabilities.
> > 2. Otherwise, we pick the most recent compatibility mode which is
> > both supported by the CPU, and is advertised as supported by the
> > guest.
> > I think the original code approximated the same thing, but it's hard to be
> > sure, and I think it had some weird edge cases.
> >
> > Signed-off-by: David Gibson <address@hidden>
> > ---
> > hw/ppc/spapr_hcall.c | 107
> > +++++++++++++++++----------------------------------
> > hw/ppc/trace-events | 2 +-
> > 2 files changed, 37 insertions(+), 72 deletions(-)
> >
> > diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
> > index d93f580..3bd6d06 100644
> > --- a/hw/ppc/spapr_hcall.c
> > +++ b/hw/ppc/spapr_hcall.c
> > @@ -895,98 +895,63 @@ static void do_set_compat(CPUState *cs, void *arg)
> > ppc_set_compat(cpu, s->compat_pvr, &s->err);
> > }
> >
> > -#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 void cas_handle_compat_cpu(PowerPCCPUClass *pcc, uint32_t pvr,
> > - unsigned max_lvl, unsigned *compat_lvl,
> > - unsigned *compat_pvr)
> > -{
> > - unsigned lvl = get_compat_level(pvr);
> > - bool is205, is206, is207;
> > -
> > - if (!lvl) {
> > - return;
> > - }
> > -
> > - /* If it is a logical PVR, try to determine the highest level */
> > - is205 = (pcc->pcr_supported & PCR_COMPAT_2_05) &&
> > - (lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_05));
> > - is206 = (pcc->pcr_supported & PCR_COMPAT_2_06) &&
> > - ((lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_06)) ||
> > - (lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_06_PLUS)));
> > - is207 = (pcc->pcr_supported & PCR_COMPAT_2_07) &&
> > - (lvl == get_compat_level(CPU_POWERPC_LOGICAL_2_07));
> > -
> > - if (is205 || is206 || is207) {
> > - if (!max_lvl) {
> > - /* User did not set the level, choose the highest */
> > - if (*compat_lvl <= lvl) {
> > - *compat_lvl = lvl;
> > - *compat_pvr = pvr;
> > - }
> > - } else if (max_lvl >= lvl) {
> > - /* User chose the level, don't set higher than this */
> > - *compat_lvl = lvl;
> > - *compat_pvr = pvr;
> > - }
> > - }
> > -}
> > -
> > -static target_ulong h_client_architecture_support(PowerPCCPU *cpu_,
> > +static target_ulong h_client_architecture_support(PowerPCCPU *cpu,
> > sPAPRMachineState *spapr,
> > target_ulong opcode,
> > target_ulong *args)
> > {
> > target_ulong list = ppc64_phys_to_real(args[0]);
> > target_ulong ov_table;
> > - PowerPCCPUClass *pcc = POWERPC_CPU_GET_CLASS(cpu_);
> > CPUState *cs;
> > - bool cpu_match = false;
> > - unsigned old_compat_pvr = cpu_->compat_pvr;
> > - unsigned compat_lvl = 0, compat_pvr = 0;
> > - unsigned max_lvl = get_compat_level(cpu_->max_compat);
> > - int counter;
> > + bool explicit_match = false; /* Matched the CPU's real PVR */
> > + uint32_t max_compat = cpu->max_compat;
> > + uint32_t best_compat = 0;
> > + int i;
> > sPAPROptionVector *ov5_guest, *ov5_cas_old, *ov5_updates;
> >
> > - /* Parse PVR list */
> > - for (counter = 0; counter < 512; ++counter) {
> > + /*
> > + * We scan the supplied table of PVRs looking for two things
> > + * 1. Is our real CPU PVR in the list?
> > + * 2. What's the "best" listed logical PVR
> > + */
> > + for (i = 0; i < 512; ++i) {
> > uint32_t pvr, pvr_mask;
> >
> > - pvr_mask = ldl_be_phys(&address_space_memory, list);
> > - list += 4;
> > pvr = ldl_be_phys(&address_space_memory, list);
>
>
> LoPAPR: List-entry = 4-byte-mask : 4-byte-PVR-value
>
> This patch changes the order and "cas" fails.
Oops, yes, thanks for the catch.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
- Re: [Qemu-ppc] [RFC 17/17] pseries: Default to POWER8 compatibility mode, (continued)
- [Qemu-ppc] [RFC 16/17] ppc: Remove counter-productive "sanity checks" in migration, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 04/17] pseries: Make cpu_update during CAS unconditional, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 15/17] ppc: Check that CPU model stays consistent across migration, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 03/17] pseries: Always use core objects for CPU construction, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 07/17] ppc: Rewrite ppc_set_compat(), David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 09/17] ppc: Validate compatibility modes when setting, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 10/17] pseries: Rewrite CAS PVR compatibility logic, David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 11/17] ppc: Add ppc_set_compat_all(), David Gibson, 2016/10/30
- [Qemu-ppc] [RFC 13/17] pseries: Move CPU compatibility property to machine, David Gibson, 2016/10/30