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

[Xen-devel] Re: [PATCH V2 4/4] amd iommu: interfaces for p2m sharing



Hi, 

I've applied all of these.

Cc'ing Allen: this series changes the xen command-line rune to share EPT
and VT-D pagetables from "sharept" to "iommu=sharept".  That seems to me
like an improvement but if you don't like it, shout. 

Cheers,

Tim.

At 14:48 +0100 on 18 Apr (1303138135), Wei Wang2 wrote:
> On Monday 18 April 2011 12:26:36 Tim Deegan wrote:
> > At 13:14 +0100 on 15 Apr (1302873244), Wei Wang2 wrote:
> > > --
> > > Advanced Micro Devices GmbH
> > > Sitz: Dornach, Gemeinde Aschheim,
> > > Landkreis München Registergericht München,
> > > HRB Nr. 43632
> > > WEEE-Reg-Nr: DE 12919551
> > > Geschäftsführer:
> > > Alberto Bozzo, Andrew Bowd
> > >
> > > # HG changeset patch
> > > # User Wei Wang <wei.wang2@xxxxxxx>
> > > # Node ID b5c197b1d2ac80881d0683cdadf12ac33b762614
> > > # Parent  1d9b0e45566ec3cd0e293212d9a454920116b2b1
> > > Add a generic interface for both vtd and amd iommu p2m sharing. Also
> > > introduce a new parameter (iommu=sharept) to enable this feature.
> >
> > Why does this introduce a separate variable?  Surely "iommu=sharept"
> > should just set iommu_hap_pt_share directly.  Also, it looks like
> > "iommu=sharept" has different semantics on AMD and Intel machines
> > (i.e., it does nothing on Intel). Was that the intention?
> Hi Tim
> True, that is a little weird.. It looks like some vtd systems have compatible 
> issue of re-using p2m table for iommu, so vtd code only turns 
> iommu_hap_pt_share on after some compatible checks (in vtd_ept_share). To 
> avoid breaking vtd, I intended to keep this in the new code. But your 
> suggestion also looks good to me. We could set iommu_hap_pt_share directly 
> and turn it off if vtd_ept_share() fails. New patch has been attached.
> 
> > The rest of this series looks OK to me.  Is it safe to apply the rest
> > without this patch or should I wait for the next version and apply them
> > all together?
> Yes, it should be OK to just apply the rest. Without the last one, p2m 
> sharing 
> will not been enabled.
> Thanks,
> Wei
> 
> 
> > Cheers,
> >
> > Tim.
> >
> > > Signed-off-by: Wei Wang <wei.wang2@xxxxxxx>
> > >
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac xen/arch/x86/mm/p2m.c
> > > --- a/xen/arch/x86/mm/p2m.c       Fri Apr 15 12:10:39 2011 +0200
> > > +++ b/xen/arch/x86/mm/p2m.c       Fri Apr 15 12:13:24 2011 +0200
> > > @@ -2061,8 +2061,8 @@ int p2m_alloc_table(struct p2m_domain *p
> > >
> > >      p2m->phys_table = pagetable_from_mfn(page_to_mfn(p2m_top));
> > >
> > > -    if ( hap_enabled(d) && (boot_cpu_data.x86_vendor ==
> > > X86_VENDOR_INTEL) ) -        iommu_set_pgd(d);
> > > +    if ( hap_enabled(d) )
> > > +        iommu_share_p2m_table(d);
> > >
> > >      P2M_PRINTK("populating p2m table\n");
> > >
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac
> > > xen/drivers/passthrough/amd/iommu_map.c ---
> > > a/xen/drivers/passthrough/amd/iommu_map.c Fri Apr 15 12:10:39 2011 +0200
> > > +++ b/xen/drivers/passthrough/amd/iommu_map.c     Fri Apr 15 12:13:24 2011
> > > +0200 @@ -716,6 +716,12 @@ void amd_iommu_share_p2m(struct domain *
> > >
> > >      ASSERT( is_hvm_domain(d) && d->arch.hvm_domain.hap_enabled );
> > >
> > > +    if ( !iommu_sharept )
> > > +    {
> > > +        iommu_hap_pt_share = 0;
> > > +        return;
> > > +    }
> > > +
> > >      pgd_mfn = pagetable_get_mfn(p2m_get_pagetable(p2m_get_hostp2m(d)));
> > >      p2m_table = mfn_to_page(mfn_x(pgd_mfn));
> > >
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac
> > > xen/drivers/passthrough/amd/pci_amd_iommu.c ---
> > > a/xen/drivers/passthrough/amd/pci_amd_iommu.c     Fri Apr 15 12:10:39 2011
> > > +0200 +++ b/xen/drivers/passthrough/amd/pci_amd_iommu.c   Fri Apr 15
> > > 12:13:24 2011 +0200 @@ -458,4 +458,5 @@ const struct iommu_ops
> > > amd_iommu_ops = {
> > >      .read_msi_from_ire = amd_iommu_read_msi_from_ire,
> > >      .suspend = amd_iommu_suspend,
> > >      .resume = amd_iommu_resume,
> > > +    .share_p2m = amd_iommu_share_p2m,
> > >  };
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac xen/drivers/passthrough/iommu.c
> > > --- a/xen/drivers/passthrough/iommu.c     Fri Apr 15 12:10:39 2011 +0200
> > > +++ b/xen/drivers/passthrough/iommu.c     Fri Apr 15 12:13:24 2011 +0200
> > > @@ -50,6 +50,7 @@ bool_t __read_mostly iommu_hap_pt_share;
> > >  bool_t __read_mostly iommu_hap_pt_share;
> > >  bool_t __read_mostly amd_iommu_debug;
> > >  bool_t __read_mostly amd_iommu_perdev_intremap;
> > > +bool_t __read_mostly iommu_sharept;
> > >
> > >  static void __init parse_iommu_param(char *s)
> > >  {
> > > @@ -82,6 +83,8 @@ static void __init parse_iommu_param(cha
> > >              iommu_passthrough = 1;
> > >          else if ( !strcmp(s, "dom0-strict") )
> > >              iommu_dom0_strict = 1;
> > > +        else if ( !strcmp(s, "sharept") )
> > > +            iommu_sharept = 1;
> > >
> > >          s = ss + 1;
> > >      } while ( ss );
> > > @@ -419,6 +422,14 @@ void iommu_suspend()
> > >          ops->suspend();
> > >  }
> > >
> > > +void iommu_share_p2m_table(struct domain* d)
> > > +{
> > > +    const struct iommu_ops *ops = iommu_get_ops();
> > > +
> > > +    if ( iommu_enabled && is_hvm_domain(d) )
> > > +        ops->share_p2m(d);
> > > +}
> > > +
> > >  /*
> > >   * Local variables:
> > >   * mode: C
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac xen/drivers/passthrough/vtd/iommu.c
> > > --- a/xen/drivers/passthrough/vtd/iommu.c Fri Apr 15 12:10:39 2011 +0200
> > > +++ b/xen/drivers/passthrough/vtd/iommu.c Fri Apr 15 12:13:24 2011 +0200
> > > @@ -1747,7 +1747,7 @@ static bool_t vtd_ept_share(void)
> > >      bool_t share = TRUE;
> > >
> > >      /* sharept defaults to 0 for now, default to 1 when feature matures
> > > */ -    if ( !sharept )
> > > +    if ( !iommu_sharept )
> > >          share = FALSE;
> > >
> > >      /*
> > > @@ -2299,6 +2299,7 @@ const struct iommu_ops intel_iommu_ops =
> > >      .read_msi_from_ire = msi_msg_read_remap_rte,
> > >      .suspend = vtd_suspend,
> > >      .resume = vtd_resume,
> > > +    .share_p2m = iommu_set_pgd,
> > >  };
> > >
> > >  /*
> > > diff -r 1d9b0e45566e -r b5c197b1d2ac xen/include/xen/iommu.h
> > > --- a/xen/include/xen/iommu.h     Fri Apr 15 12:10:39 2011 +0200
> > > +++ b/xen/include/xen/iommu.h     Fri Apr 15 12:13:24 2011 +0200
> > > @@ -33,6 +33,7 @@ extern bool_t iommu_hap_pt_share;
> > >  extern bool_t iommu_hap_pt_share;
> > >  extern bool_t amd_iommu_debug;
> > >  extern bool_t amd_iommu_perdev_intremap;
> > > +extern bool_t iommu_sharept;
> > >
> > >  extern struct rangeset *mmio_ro_ranges;
> > >
> > > @@ -132,6 +133,7 @@ struct iommu_ops {
> > >      unsigned int (*read_apic_from_ire)(unsigned int apic, unsigned int
> > > reg); void (*suspend)(void);
> > >      void (*resume)(void);
> > > +    void (*share_p2m)(struct domain *d);
> > >  };
> > >
> > >  void iommu_update_ire_from_apic(unsigned int apic, unsigned int reg,
> > > unsigned int value); @@ -143,5 +145,6 @@ void iommu_resume(void);
> > >  void iommu_resume(void);
> > >
> > >  void iommu_set_dom0_mapping(struct domain *d);
> > > +void iommu_share_p2m_table(struct domain *d);
> > >
> > >  #endif /* _IOMMU_H_ */
> 
> 

> # HG changeset patch
> # User Wei Wang <wei.wang2@xxxxxxx>
> # Node ID 15a9daf8a8aedf21e941a9ed3cb250536ac6972f
> # Parent  5846cc03cf0c1a4b02eba76a7b4aa4884de5e957
> Add a generic interface for both vtd and amd iommu p2m sharing. Also 
> introduce a new parameter (iommu=sharept) to enable this feature.
> 
> Signed-off-by: Wei Wang <wei.wang2@xxxxxxx>
> 
> diff -r 5846cc03cf0c -r 15a9daf8a8ae xen/arch/x86/mm/p2m.c
> --- a/xen/arch/x86/mm/p2m.c   Mon Apr 18 15:01:52 2011 +0200
> +++ b/xen/arch/x86/mm/p2m.c   Mon Apr 18 15:37:28 2011 +0200
> @@ -2061,8 +2061,8 @@ int p2m_alloc_table(struct p2m_domain *p
>  
>      p2m->phys_table = pagetable_from_mfn(page_to_mfn(p2m_top));
>  
> -    if ( hap_enabled(d) && (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) )
> -        iommu_set_pgd(d);
> +    if ( hap_enabled(d) )
> +        iommu_share_p2m_table(d);
>  
>      P2M_PRINTK("populating p2m table\n");
>  
> diff -r 5846cc03cf0c -r 15a9daf8a8ae xen/drivers/passthrough/amd/iommu_map.c
> --- a/xen/drivers/passthrough/amd/iommu_map.c Mon Apr 18 15:01:52 2011 +0200
> +++ b/xen/drivers/passthrough/amd/iommu_map.c Mon Apr 18 15:37:28 2011 +0200
> @@ -716,6 +716,9 @@ void amd_iommu_share_p2m(struct domain *
>  
>      ASSERT( is_hvm_domain(d) && d->arch.hvm_domain.hap_enabled );
>  
> +    if ( !iommu_hap_pt_share )
> +        return;
> +
>      pgd_mfn = pagetable_get_mfn(p2m_get_pagetable(p2m_get_hostp2m(d)));
>      p2m_table = mfn_to_page(mfn_x(pgd_mfn));
>  
> @@ -726,8 +729,6 @@ void amd_iommu_share_p2m(struct domain *
>  
>          /* When sharing p2m with iommu, paging mode = 4 */
>          hd->paging_mode = IOMMU_PAGING_MODE_LEVEL_4;
> -        iommu_hap_pt_share = 1;
> -
>          AMD_IOMMU_DEBUG("Share p2m table with iommu: p2m table = 0x%lx\n",
>                          mfn_x(pgd_mfn));
>      }
> diff -r 5846cc03cf0c -r 15a9daf8a8ae 
> xen/drivers/passthrough/amd/pci_amd_iommu.c
> --- a/xen/drivers/passthrough/amd/pci_amd_iommu.c     Mon Apr 18 15:01:52 
> 2011 +0200
> +++ b/xen/drivers/passthrough/amd/pci_amd_iommu.c     Mon Apr 18 15:37:28 
> 2011 +0200
> @@ -458,4 +458,5 @@ const struct iommu_ops amd_iommu_ops = {
>      .read_msi_from_ire = amd_iommu_read_msi_from_ire,
>      .suspend = amd_iommu_suspend,
>      .resume = amd_iommu_resume,
> +    .share_p2m = amd_iommu_share_p2m,
>  };
> diff -r 5846cc03cf0c -r 15a9daf8a8ae xen/drivers/passthrough/iommu.c
> --- a/xen/drivers/passthrough/iommu.c Mon Apr 18 15:01:52 2011 +0200
> +++ b/xen/drivers/passthrough/iommu.c Mon Apr 18 15:37:28 2011 +0200
> @@ -82,6 +82,8 @@ static void __init parse_iommu_param(cha
>              iommu_passthrough = 1;
>          else if ( !strcmp(s, "dom0-strict") )
>              iommu_dom0_strict = 1;
> +        else if ( !strcmp(s, "sharept") )
> +            iommu_hap_pt_share = 1;
>  
>          s = ss + 1;
>      } while ( ss );
> @@ -419,6 +421,14 @@ void iommu_suspend()
>          ops->suspend();
>  }
>  
> +void iommu_share_p2m_table(struct domain* d)
> +{
> +    const struct iommu_ops *ops = iommu_get_ops();
> +
> +    if ( iommu_enabled && is_hvm_domain(d) )
> +        ops->share_p2m(d);
> +}
> +
>  /*
>   * Local variables:
>   * mode: C
> diff -r 5846cc03cf0c -r 15a9daf8a8ae xen/drivers/passthrough/vtd/iommu.c
> --- a/xen/drivers/passthrough/vtd/iommu.c     Mon Apr 18 15:01:52 2011 +0200
> +++ b/xen/drivers/passthrough/vtd/iommu.c     Mon Apr 18 15:37:28 2011 +0200
> @@ -45,9 +45,6 @@
>  #define nr_ioapics              iosapic_get_nr_iosapics()
>  #endif
>  
> -static int sharept = 0;
> -boolean_param("sharept", sharept);
> -
>  int nr_iommus;
>  
>  static void setup_dom0_devices(struct domain *d);
> @@ -1747,7 +1744,7 @@ static bool_t vtd_ept_share(void)
>      bool_t share = TRUE;
>  
>      /* sharept defaults to 0 for now, default to 1 when feature matures */
> -    if ( !sharept )
> +    if ( !iommu_hap_pt_share )
>          share = FALSE;
>  
>      /*
> @@ -2299,6 +2296,7 @@ const struct iommu_ops intel_iommu_ops =
>      .read_msi_from_ire = msi_msg_read_remap_rte,
>      .suspend = vtd_suspend,
>      .resume = vtd_resume,
> +    .share_p2m = iommu_set_pgd,
>  };
>  
>  /*
> diff -r 5846cc03cf0c -r 15a9daf8a8ae xen/include/xen/iommu.h
> --- a/xen/include/xen/iommu.h Mon Apr 18 15:01:52 2011 +0200
> +++ b/xen/include/xen/iommu.h Mon Apr 18 15:37:28 2011 +0200
> @@ -132,6 +132,7 @@ struct iommu_ops {
>      unsigned int (*read_apic_from_ire)(unsigned int apic, unsigned int reg);
>      void (*suspend)(void);
>      void (*resume)(void);
> +    void (*share_p2m)(struct domain *d);
>  };
>  
>  void iommu_update_ire_from_apic(unsigned int apic, unsigned int reg, 
> unsigned int value);
> @@ -143,5 +144,6 @@ void iommu_resume(void);
>  void iommu_resume(void);
>  
>  void iommu_set_dom0_mapping(struct domain *d);
> +void iommu_share_p2m_table(struct domain *d);
>  
>  #endif /* _IOMMU_H_ */


-- 
Tim Deegan <Tim.Deegan@xxxxxxxxxx>
Principal Software Engineer, Xen Platform Team
Citrix Systems UK Ltd.  (Company #02937203, SL9 0BG)

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.