Hi Jan,
On 29.06.2016 10:57, Jan Kiszka wrote:
On 2016-06-28 20:00, Valentine Sinitsyn wrote:
Hi all,
When interrupt remapping support was introduced first in Jailhouse for
VT-d, we gained union x86_msi_vector.remap and related struct in union
ioapic_redir_entry. Both are in vendor-neutral x86 code.
Do I get correctly that in fact the layout of this structure is VT-d
specific? On AMD, interrupt remapping tables are indexed with 10 lowest
bits of MSI data (Sect 2.2.5.1 in the spec), and we don't seem to touch
this part in Jailhouse. You may find relevant code in the Linux kernel
at [1]
Is it true, or am I missing something?
True, x86_msi_vector.remap is only relevant in the context of VT-d. Feel
tree to rename it to vtd_remap or intel_remap and add an amd_remap
variant (if needed).
OK, thanks. I don't think I need a new field for AMD, but I'll probably
introduce a set of two vendor-specific functions which fuse the index
iommu_map_interrupt() returns back to x86_msi or ioapic_redir_entry.
Valentine
--
You received this message because you are subscribed to the Google Groups
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
