Re: [Qemu-ppc] [RFC NO-MERGE 07/12] target/ppc: Implement H_REGISTER_PROCESS_TABLE H_CALL for tcg guests

[Suraj Jitindar Singh]

On Mon, 2017-02-20 at 12:48 +1100, David Gibson wrote:
> On Fri, Feb 17, 2017 at 04:08:07PM +1100, Suraj Jitindar Singh wrote:
> > 
> > The H_REGISTER_PROCESS_TABLE H_CALL is used by a guest to indicate
> > to the
> > hypervisor where in memory its process table is and how translation
> > should
> > be performed using this process table.
> > 
> > Provide the implementation of this H_CALL for a tcg guest. If
> > called by a
> > KVM guest it will return H_FUNCTION for now with //TODO markers
> > where code
> > should be added to implement the required functionality.
> > 
> > We first check for invalid flags, then parse the flags to determine
> > the
> > operation, and then check the other parameters for valid values
> > based on
> > the operation (register new table/deregister table/maintain
> > registration).
> > The process table is then stored in the appropriate location, and
> > the
> > LPCR_[UPRT/GTSE] bits are updated as required.
> > 
> > Signed-off-by: Suraj Jitindar Singh <address@hidden>
> > ---
> >  hw/ppc/spapr_hcall.c | 64
> > +++++++++++++++++++++++++++++++++++++++++++++++++---
> >  1 file changed, 61 insertions(+), 3 deletions(-)
> > 
> > diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
> > index 7afedbd..bf80543 100644
> > --- a/hw/ppc/spapr_hcall.c
> > +++ b/hw/ppc/spapr_hcall.c
> > @@ -12,6 +12,7 @@
> >  #include "trace.h"
> >  #include "kvm_ppc.h"
> >  #include "hw/ppc/spapr_ovec.h"
> > +#include "mmu.h"
> >  
> >  struct SPRSyncState {
> >      int spr;
> > @@ -902,9 +903,66 @@ static target_ulong
> > h_register_process_table(PowerPCCPU *cpu,
> >                                               target_ulong opcode,
> >                                               target_ulong *args)
> >  {
> > -    qemu_log_mask(LOG_UNIMP, "Unimplemented SPAPR hcall 0x"
> > TARGET_FMT_lx "%s\n",
> > -                  opcode, " (H_REGISTER_PROC_TBL)");
> > -    return H_FUNCTION;
> > +    CPUPPCState *env = &cpu->env;
> > +    target_ulong flags = args[0];
> > +    target_ulong proc_tbl = args[1];
> > +    target_ulong page_size = args[2];
> > +    target_ulong table_size = args[3];
> > +    uint64_t cproc = spapr->patb_entry; /* Get the current table
> > */
> > +
> > +    if (flags & ~0x1fULL) { /* Check no reserved bits are set */
> Some #defines for the bit masks would be nice.

I'll add some.

> 
> > 
> > +        return H_PARAMETER;
> > +    } else if (!(flags & 0x10)) { /* Maintain current registration
> > */
> > +        if (!!(flags & 0x4) ^ !!(cproc & PATBE1_GR)) {
> > +            return H_PARAMETER; /* Existing Process Table Mismatch
> > */
> > +   }
> > +    } else if (!(flags & 0x8)) { /* Deregister current process
> > table */
> > +        if (kvm_enabled()) {
> > +            // TODO - how does kvm handle deregistration?
> > +       return H_FUNCTION;
> I think H_HARDWARE would be preferable.

This shouldn't be an issue as samb and my patches will come in pretty
much the same time. But sure

> 
> > 
> > +        } else {
> > +            cproc = 0x0ULL; /* Set to benign value - 0 works */
> > +        }
> > +    } else if (flags & 0x4) { /* Register new RADIX process table
> > */
> > +        if (proc_tbl & 0xfff || proc_tbl >> 60) {
> > +            return H_P2;
> > +        } else if (page_size) {
> > +            return H_P3;
> > +        } else if (table_size > 24) {
> > +            return H_P4;
> > +        }
> > +        cproc = PATBE1_GR | proc_tbl | table_size;
> > +   // TODO
> > +   // Free HTAB if non-null
> > +    } else { /* Register new HPT process table */
> > +        if (proc_tbl >> 38) {
> > +            return H_P2;
> > +        } else if (page_size & ~0x7) {
> > +             return H_P3;
> > +        } else if (table_size > 24) {
> > +            return H_P4;
> > +        }
> > +        cproc = (proc_tbl << 25) | page_size << 5 | table_size;
> > +    }
> > +
> > +    if (kvm_enabled()) {
> > +        // TODO - kvmppc_configure_v3_mmu()
> > +   return H_FUNCTION;
> > +    } else {
> > +        spapr->patb_entry = cproc; /* Save new process table */
> > +        if (flags & 0x6) { /* Partition Uses Process Tables */
> > +            env->spr[SPR_LPCR] |= LPCR_UPRT;
> > +   } else {
> > +            env->spr[SPR_LPCR] &= ~LPCR_UPRT;
> It's probably safer to update the LPCR in the kvm case as well.  In
> migration we'll need the qemu copy of it - we'd probably reload it
> from KVM before then, but safer to keep it in sync anyway.

Sounds good

> 
> > 
> > +   }
> > +   if (flags & 0x1) { /* Partition Uses Guest Translation
> > Shootdown */
> > +            env->spr[SPR_LPCR] |= LPCR_GTSE;
> > +        } else {
> > +            env->spr[SPR_LPCR] &= ~LPCR_GTSE;
> > +        }
> > +    }
> > +
> > +    return H_SUCCESS;
> >  }
> >  
> >  #define H_SIGNAL_SYS_RESET_ALL         -1