qemu-ppc
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-ppc] [RFC NO-MERGE 09/12] target/ppc: Flush TLB on write to PI


From: Suraj Jitindar Singh
Subject: Re: [Qemu-ppc] [RFC NO-MERGE 09/12] target/ppc: Flush TLB on write to PIDR
Date: Tue, 28 Feb 2017 13:54:16 +1100

On Mon, 2017-02-27 at 13:09 +1100, David Gibson wrote:
> On Fri, Feb 24, 2017 at 06:13:03PM +1100, Suraj Jitindar Singh wrote:
> > 
> > On Mon, 2017-02-20 at 13:18 +1100, David Gibson wrote:
> > > 
> > > On Fri, Feb 17, 2017 at 04:08:09PM +1100, Suraj Jitindar Singh
> > > wrote:
> > > > 
> > > > 
> > > > The PIDR (process id register) is used to store the id of the
> > > > currently
> > > > running process, which is used to select the process table
> > > > entry
> > > > used to
> > > > perform address translation. This means that when we write to
> > > > this
> > > > register
> > > > all the translations in the TLB become outdated as they are for
> > > > a
> > > > previously running process. Thus when this register is written
> > > > to
> > > > we need
> > > > to invalidate the TLB entries to ensure stale entries aren't
> > > > used
> > > > to
> > > > to perform translation for the new process, which would result
> > > > in
> > > > at best
> > > > segfaults or alternatively just random memory being accessed.
> > > > 
> > > > Signed-off-by: Suraj Jitindar Singh <address@hidden>
> > > How does this interact with the SLB if running in hash mode
> > > without
> > > segment tables?  Does writing the PIDR automatically invalidate
> > > the
> > > SLB, or, does the SLB need to be invalidated separately?  If the
> > > second, then writing the PIDR probably doesn't need to invalidate
> > > the
> > > qemu TLB in hash+SLB mode.
> > Currently in hash we don't touch the pidr so it's a little
> > irrelevant.
> > Entries are supposed to be automatically invalidated when the PIDR
> > is
> > written to so I think we still need this since I don't believe the
> > software is required to issue an invalidation.
> Hmm.. meaning that the SLB entries should also be discarded on a
> write
> to the PIDR?

When we (I) implement segment tables, yes.

> 
> > 
> > 
> > > 
> > > 
> > > > 
> > > > 
> > > > ---
> > > >  target/ppc/helper.h         | 1 +
> > > >  target/ppc/misc_helper.c    | 8 ++++++++
> > > >  target/ppc/translate_init.c | 8 +++++++-
> > > >  3 files changed, 16 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/target/ppc/helper.h b/target/ppc/helper.h
> > > > index 85af9df..53ff749 100644
> > > > --- a/target/ppc/helper.h
> > > > +++ b/target/ppc/helper.h
> > > > @@ -698,6 +698,7 @@ DEF_HELPER_FLAGS_1(load_601_rtcu,
> > > > TCG_CALL_NO_RWG, tl, env)
> > > >  DEF_HELPER_FLAGS_1(load_purr, TCG_CALL_NO_RWG, tl, env)
> > > >  #endif
> > > >  DEF_HELPER_2(store_sdr1, void, env, tl)
> > > > +DEF_HELPER_2(store_pidr, void, env, tl)
> > > >  DEF_HELPER_FLAGS_2(store_tbl, TCG_CALL_NO_RWG, void, env, tl)
> > > >  DEF_HELPER_FLAGS_2(store_tbu, TCG_CALL_NO_RWG, void, env, tl)
> > > >  DEF_HELPER_FLAGS_2(store_atbl, TCG_CALL_NO_RWG, void, env, tl)
> > > > diff --git a/target/ppc/misc_helper.c
> > > > b/target/ppc/misc_helper.c
> > > > index 49ba767..350057a 100644
> > > > --- a/target/ppc/misc_helper.c
> > > > +++ b/target/ppc/misc_helper.c
> > > > @@ -91,6 +91,14 @@ void helper_store_sdr1(CPUPPCState *env,
> > > > target_ulong val)
> > > >      }
> > > >  }
> > > >  
> > > > +void helper_store_pidr(CPUPPCState *env, target_ulong val)
> > > > +{
> > > > +    PowerPCCPU *cpu = ppc_env_get_cpu(env);
> > > > +
> > > > +    env->spr[SPR_BOOKS_PID] = val;
> > > > +    tlb_flush(CPU(cpu));
> > > > +}
> > > > +
> > > >  void helper_store_hid0_601(CPUPPCState *env, target_ulong val)
> > > >  {
> > > >      target_ulong hid0;
> > > > diff --git a/target/ppc/translate_init.c
> > > > b/target/ppc/translate_init.c
> > > > index 66a7f4a..bdc3894 100644
> > > > --- a/target/ppc/translate_init.c
> > > > +++ b/target/ppc/translate_init.c
> > > > @@ -394,6 +394,12 @@ static void spr_write_sdr1 (DisasContext
> > > > *ctx,
> > > > int sprn, int gprn)
> > > >      gen_helper_store_sdr1(cpu_env, cpu_gpr[gprn]);
> > > >  }
> > > >  
> > > > +/* PIDR */
> > > > +static void spr_write_pidr (DisasContext *ctx, int sprn, int
> > > > gprn)
> > > > +{
> > > > +    gen_helper_store_pidr(cpu_env, cpu_gpr[gprn]);
> > > > +}
> > > > +
> > > >  /* 64 bits PowerPC specific SPRs */
> > > >  #if defined(TARGET_PPC64)
> > > >  static void spr_read_hior (DisasContext *ctx, int gprn, int
> > > > sprn)
> > > > @@ -8170,7 +8176,7 @@ static void
> > > > gen_spr_power8_book4(CPUPPCState
> > > > *env)
> > > >                       KVM_REG_PPC_ACOP, 0);
> > > >      spr_register_kvm(env, SPR_BOOKS_PID, "PID",
> > > >                       SPR_NOACCESS, SPR_NOACCESS,
> > > > -                     &spr_read_generic, &spr_write_generic,
> > > > +                     &spr_read_generic, &spr_write_pidr,
> > > >                       KVM_REG_PPC_PID, 0);
> > > >      spr_register_kvm(env, SPR_WORT, "WORT",
> > > >                       SPR_NOACCESS, SPR_NOACCESS,



reply via email to

[Prev in Thread] Current Thread [Next in Thread]