Initial IOMMU driver hooks.
This is intentially stripped back to initialization and a hook to the
driver when the VM's memory map is changing. This will work as the
baseline from which changes can be made to fit the needs of the drivers.
Change-Id: I904279f511e2d6e4b1c062fb49a2892042b79005
diff --git a/src/spci_architected_message.c b/src/spci_architected_message.c
index 7c9d2b7..2c3ff89 100644
--- a/src/spci_architected_message.c
+++ b/src/spci_architected_message.c
@@ -19,6 +19,7 @@
#include "hf/dlog.h"
#include "hf/spci_internal.h"
#include "hf/std.h"
+#include "hf/vm.h"
/**
* Obtain the next mode to apply to the two VMs.
@@ -316,8 +317,8 @@
* made to memory mappings.
*/
static bool spci_region_group_identity_map(
- struct mm_ptable *t, struct spci_memory_region *memory_region, int mode,
- struct mpool *ppool, bool commit)
+ struct vm_locked vm_locked, struct spci_memory_region *memory_region,
+ int mode, struct mpool *ppool, bool commit)
{
struct spci_memory_region_constituent *constituents =
spci_memory_region_get_constituents(memory_region);
@@ -331,10 +332,10 @@
paddr_t pa_end = pa_add(pa_begin, size);
if (commit) {
- mm_vm_identity_commit(t, pa_begin, pa_end, mode, ppool,
- NULL);
- } else if (!mm_vm_identity_prepare(t, pa_begin, pa_end, mode,
- ppool)) {
+ vm_identity_commit(vm_locked, pa_begin, pa_end, mode,
+ ppool, NULL);
+ } else if (!vm_identity_prepare(vm_locked, pa_begin, pa_end,
+ mode, ppool)) {
return false;
}
}
@@ -501,9 +502,9 @@
* sure the entire operation will succeed without exhausting the page
* pool.
*/
- if (!spci_region_group_identity_map(&from->ptable, memory_region,
+ if (!spci_region_group_identity_map(from_locked, memory_region,
from_mode, api_page_pool, false) ||
- !spci_region_group_identity_map(&to->ptable, memory_region, to_mode,
+ !spci_region_group_identity_map(to_locked, memory_region, to_mode,
api_page_pool, false)) {
/* TODO: partial defrag of failed range. */
ret = spci_error(SPCI_NO_MEMORY);
@@ -516,9 +517,8 @@
* already prepared above, but may free pages in the case that a whole
* block is being unmapped that was previously partially mapped.
*/
- CHECK(spci_region_group_identity_map(&from->ptable, memory_region,
- from_mode, &local_page_pool,
- true));
+ CHECK(spci_region_group_identity_map(
+ from_locked, memory_region, from_mode, &local_page_pool, true));
/* Clear the memory so no VM or device can see the previous contents. */
if ((memory_region->flags & SPCI_MEMORY_REGION_FLAG_CLEAR) &&
@@ -529,9 +529,9 @@
* `local_page_pool` by the call above, but will never allocate
* more pages than that so can never fail.
*/
- CHECK(spci_region_group_identity_map(
- &from->ptable, memory_region, orig_from_mode,
- &local_page_pool, true));
+ CHECK(spci_region_group_identity_map(from_locked, memory_region,
+ orig_from_mode,
+ &local_page_pool, true));
ret = spci_error(SPCI_NO_MEMORY);
goto out;
@@ -542,8 +542,8 @@
* won't allocate because the transaction was already prepared above, so
* it doesn't need to use the `local_page_pool`.
*/
- CHECK(spci_region_group_identity_map(&to->ptable, memory_region,
- to_mode, api_page_pool, true));
+ CHECK(spci_region_group_identity_map(to_locked, memory_region, to_mode,
+ api_page_pool, true));
ret = (struct spci_value){.func = SPCI_SUCCESS_32};