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

Re: [PATCH v4 14/24] xen/domctl: wrap pci-subset iommu-related domctl op with CONFIG_MGMT_HYPERCALLS


  • To: Penny Zheng <Penny.Zheng@xxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Tue, 25 Nov 2025 16:37:01 +0100
  • Autocrypt: addr=jbeulich@xxxxxxxx; keydata= xsDiBFk3nEQRBADAEaSw6zC/EJkiwGPXbWtPxl2xCdSoeepS07jW8UgcHNurfHvUzogEq5xk hu507c3BarVjyWCJOylMNR98Yd8VqD9UfmX0Hb8/BrA+Hl6/DB/eqGptrf4BSRwcZQM32aZK 7Pj2XbGWIUrZrd70x1eAP9QE3P79Y2oLrsCgbZJfEwCgvz9JjGmQqQkRiTVzlZVCJYcyGGsD /0tbFCzD2h20ahe8rC1gbb3K3qk+LpBtvjBu1RY9drYk0NymiGbJWZgab6t1jM7sk2vuf0Py O9Hf9XBmK0uE9IgMaiCpc32XV9oASz6UJebwkX+zF2jG5I1BfnO9g7KlotcA/v5ClMjgo6Gl MDY4HxoSRu3i1cqqSDtVlt+AOVBJBACrZcnHAUSuCXBPy0jOlBhxPqRWv6ND4c9PH1xjQ3NP nxJuMBS8rnNg22uyfAgmBKNLpLgAGVRMZGaGoJObGf72s6TeIqKJo/LtggAS9qAUiuKVnygo 3wjfkS9A3DRO+SpU7JqWdsveeIQyeyEJ/8PTowmSQLakF+3fote9ybzd880fSmFuIEJldWxp Y2ggPGpiZXVsaWNoQHN1c2UuY29tPsJgBBMRAgAgBQJZN5xEAhsDBgsJCAcDAgQVAggDBBYC AwECHgECF4AACgkQoDSui/t3IH4J+wCfQ5jHdEjCRHj23O/5ttg9r9OIruwAn3103WUITZee e7Sbg12UgcQ5lv7SzsFNBFk3nEQQCACCuTjCjFOUdi5Nm244F+78kLghRcin/awv+IrTcIWF hUpSs1Y91iQQ7KItirz5uwCPlwejSJDQJLIS+QtJHaXDXeV6NI0Uef1hP20+y8qydDiVkv6l IreXjTb7DvksRgJNvCkWtYnlS3mYvQ9NzS9PhyALWbXnH6sIJd2O9lKS1Mrfq+y0IXCP10eS FFGg+Av3IQeFatkJAyju0PPthyTqxSI4lZYuJVPknzgaeuJv/2NccrPvmeDg6Coe7ZIeQ8Yj t0ARxu2xytAkkLCel1Lz1WLmwLstV30g80nkgZf/wr+/BXJW/oIvRlonUkxv+IbBM3dX2OV8 AmRv1ySWPTP7AAMFB/9PQK/VtlNUJvg8GXj9ootzrteGfVZVVT4XBJkfwBcpC/XcPzldjv+3 HYudvpdNK3lLujXeA5fLOH+Z/G9WBc5pFVSMocI71I8bT8lIAzreg0WvkWg5V2WZsUMlnDL9 mpwIGFhlbM3gfDMs7MPMu8YQRFVdUvtSpaAs8OFfGQ0ia3LGZcjA6Ik2+xcqscEJzNH+qh8V m5jjp28yZgaqTaRbg3M/+MTbMpicpZuqF4rnB0AQD12/3BNWDR6bmh+EkYSMcEIpQmBM51qM EKYTQGybRCjpnKHGOxG0rfFY1085mBDZCH5Kx0cl0HVJuQKC+dV2ZY5AqjcKwAxpE75MLFkr wkkEGBECAAkFAlk3nEQCGwwACgkQoDSui/t3IH7nnwCfcJWUDUFKdCsBH/E5d+0ZnMQi+G0A nAuWpQkjM1ASeQwSHEeAWPgskBQL
  • Cc: ray.huang@xxxxxxx, grygorii_strashko@xxxxxxxx, Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>, Jason Andryuk <jason.andryuk@xxxxxxx>, "Daniel P. Smith" <dpsmith@xxxxxxxxxxxxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx
  • Delivery-date: Tue, 25 Nov 2025 15:37:15 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 21.11.2025 11:57, Penny Zheng wrote:
>  xen/drivers/passthrough/amd/pci_amd_iommu.c | 20 ++++----
>  xen/drivers/passthrough/pci.c               | 52 +++++++++++----------
>  xen/drivers/passthrough/vtd/iommu.c         |  6 ++-
>  xen/include/xsm/dummy.h                     |  6 ++-
>  xen/include/xsm/xsm.h                       | 12 +++--
>  xen/xsm/dummy.c                             |  6 ++-
>  xen/xsm/flask/hooks.c                       | 12 +++--
>  7 files changed, 68 insertions(+), 46 deletions(-)

With this diffstat and there being quite a few HAS_PCI under
xen/drivers/passthrough/arm/, what's the (PCI) deal there?
> @@ -772,14 +774,16 @@ static const struct iommu_ops __initconst_cf_clobber 
> _iommu_ops = {
>      .quarantine_init = amd_iommu_quarantine_init,
>      .add_device = amd_iommu_add_device,
>      .remove_device = amd_iommu_remove_device,
> -    .assign_device  = amd_iommu_assign_device,
>      .teardown = amd_iommu_domain_destroy,
>      .clear_root_pgtable = amd_iommu_clear_root_pgtable,
>      .map_page = amd_iommu_map_page,
>      .unmap_page = amd_iommu_unmap_page,
>      .iotlb_flush = amd_iommu_flush_iotlb_pages,
> +#ifdef CONFIG_MGMT_HYPERCALLS
> +    .assign_device  = amd_iommu_assign_device,
>      .reassign_device = reassign_device,
>      .get_device_group_id = amd_iommu_group_id,
> +#endif

You don't zap the hooks themselves, i.e. they end up being NULL now in
the (still only hypothetical, provided the Kconfig change will be adjusted)
case of MGMT_HYPERCALLS=n. I understand the former two hooks are still
needed for DT, but at least .get_device_group_id should be properly dealt
with in xen/iommu.h right away, imo. This would then also clarify already
here that that's the plan for the other two hooks as well.

> --- a/xen/include/xsm/dummy.h
> +++ b/xen/include/xsm/dummy.h
> @@ -407,7 +407,8 @@ static XSM_INLINE int cf_check xsm_get_vnumainfo(
>      return xsm_default_action(action, current->domain, d);
>  }
>  
> -#if defined(CONFIG_HAS_PASSTHROUGH) && defined(CONFIG_HAS_PCI)
> +#if defined(CONFIG_HAS_PASSTHROUGH) && defined(CONFIG_MGMT_HYPERCALLS)
> +#ifdef CONFIG_HAS_PCI

Why the separate #ifdef? Can't that be folded with the #if? Are there further
changes to be put inside the outer #if? (Applies again further down as well.)

>  static XSM_INLINE int cf_check xsm_get_device_group(
>      XSM_DEFAULT_ARG uint32_t machine_bdf)
>  {
> @@ -429,7 +430,8 @@ static XSM_INLINE int cf_check xsm_deassign_device(
>      return xsm_default_action(action, current->domain, d);
>  }
>  
> -#endif /* HAS_PASSTHROUGH && HAS_PCI */
> +#endif /* CONFIG_HAS_PCI */
> +#endif /* HAS_PASSTHROUGH && MGMT_HYPERCALLS */
>  
>  #if defined(CONFIG_HAS_PASSTHROUGH) && 
> defined(CONFIG_HAS_DEVICE_TREE_DISCOVERY)
>  static XSM_INLINE int cf_check xsm_assign_dtdevice(

The DT counterpart, otoh, is separate anyway.

Jan



 


Rackspace

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