On 01.10.19 18:36, Julien Grall wrote:
Hi Oleksandr,
Hi Julien
On 01/10/2019 16:25, Oleksandr wrote:
On 01.10.19 18:04, Julien Grall wrote:
Hi,
Hi Julien
I am reviving the thread. I think we need a patch similar to this
one for Xen 4.13. This is because generic are now used by Xen so
they should be hidden from the hardware domain.
Andrii, Oleksandr, can one of you look at it?
I will be able to look at it probably at the end of the week if there
is no urgency.
That's fine, I think we can make a case to add it in Xen 4.13.
Cheers,
On 21/01/2019 17:04, Andrii Anisov wrote:
From: Oleksandr Tyshchenko <oleksandr_tyshche...@epam.com>
We don't passthrough IOMMU device to DOM0 even if it is not used by
Xen. Therefore exposing the properties that describe relationship
between master devices and IOMMUs does not make any sense.
According to the:
1. Documentation/devicetree/bindings/iommu/iommu.txt
2. Documentation/devicetree/bindings/pci/pci-iommu.txt
Signed-off-by: Oleksandr Tyshchenko <oleksandr_tyshche...@epam.com>
Acked-by: Julien Grall <julien.gr...@arm.com>
---
xen/arch/arm/domain_build.c | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index d2c63a8..15a08d6 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -540,6 +540,16 @@ static int __init write_properties(struct
domain *d, struct kernel_info *kinfo,
continue;
}
+ /* Don't expose IOMMU specific properties to the guest */
+ if ( dt_property_name_is_equal(prop, "iommus") )
+ continue;
+
+ if ( dt_property_name_is_equal(prop, "iommu-map") )
+ continue;
+
+ if ( dt_property_name_is_equal(prop, "iommu-map-mask") )
+ continue;
+
res = fdt_property(kinfo->fdt, prop->name, prop_data,
prop_len);
if ( res )
Julien, are you happy to see this patch as is, or do you have some
comments regarding it?
I have some comments on the cover letter for this patch. Please see [1].
Thank you for having a look at the patch.
Cheers,
[1] <ed087980-a2b9-2fd4-7e84-446142e81...@arm.com>
Looking briefly, I found two main points regarding that patch. This is
how I understand them (please, correct me if I am wrong):
1. The IOMMU can be accessible by Dom0 (for example, if we pass
"iommu=disabled" to Xen command line or it is enabled, but there is not
suitable driver in Xen found).
There is no need to remove properties if Dom0 is already touching the IOMMU.
2. Generic IOMMU DT bindings is not used in Xen so far. There is no need
to remove properties.
As I understand, both points are not actual anymore and nothing to
modify in that patch, correct?
Because:
1. Giving the IOMMU to Dom0 is a bad idea.
2. Already supported.
--
Regards,
Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
