On 6/12/24 9:50 PM, Jason Gunthorpe wrote:
On Wed, Jun 12, 2024 at 10:19:46AM -0300, Jason Gunthorpe wrote:
I prefer not to mess the definition of user API data and the kernel
driver implementation. The kernel driver may change in the future, but
the user API will remain stable for a long time.
sure it remains stable for reasonable reason. Here we defined some
fields but they are even not used and checked in the kernel. IMHO it
suggests redundant definition. If there is value to keep them, do we
need to at least verify them same as the completion record?
They are not here for the kernel, they are here for userspace.
A single HWPT and a single fault queue can be attached to multiple
devices we need to return the dev_id so that userspace can know which
device initiated the PRI. Same with PASID.
The only way we could remove them is if we are sure that no vIOMMU
requires RID or PASID in the virtual fault queue PRI fault message.. I
don't think that is true?
Cookie is not a replacement, cookie is an opaque value for the kernel
to use to match a response to a request.
Oh I got this wrong, the above is the response, yeah we can ditch
everything but the cookie, and code right?
struct iommu_hwpt_page_response {
__u32 cookie;
__u32 code;
};
What is the workflow here? We expect the VMM will take the vIOMMU
response and match it to a request cookie for the kernel to complete?
The workflow in the host kernel involves looking up the iopf_group using
the cookie value in the response queue and then responding to the
iopf_group with the code. Therefore, from the host kernel's perspective,
it is acceptable if the user only passes the cookie and code in the
response message.
Since you both agreed that the response message could be simplified, I
will implement the above structure in the new version.
Best regards,
baolu
