@@ -94,6 +94,7 @@ struct iommu_table {
unsigned long it_size; /* Size of iommu table in entries */
unsigned long it_indirect_levels;
unsigned long it_level_size;
+ unsigned long it_allocated_size;
unsigned long it_offset; /* Offset into global table */
unsigned long it_base; /* mapped address of tce table */
unsigned long it_index; /* which iommu table this is */
@@ -159,6 +160,10 @@ struct iommu_table_group_ops {
void (*set_ownership)(struct iommu_table_group *table_group,
bool enable);
+ unsigned long (*get_table_size)(
+ __u32 page_shift,
+ __u64 window_size,
+ __u32 levels);
long (*create_table)(struct iommu_table_group *table_group,
int num,
__u32 page_shift,
@@ -1373,6 +1373,57 @@ static void pnv_free_tce_table(unsigned long addr, unsigned long size,
free_pages(addr, get_order(size << 3));
}
+static unsigned long pnv_get_tce_table_size(unsigned shift, unsigned levels,
+ unsigned long *left)
+{
+ unsigned long ret, chunk = 1UL << shift, i;
+
+ ret = chunk;
+
+ if (!*left)
+ return 0;
+
+ --levels;
+ if (!levels) {
+ /* This is last level, actual TCEs */
+ *left -= min(*left, chunk);
+ return chunk;
+ }
+
+ for (i = 0; i < (chunk >> 3); ++i) {
+ ret += pnv_get_tce_table_size(shift, levels, left);
+ if (!*left)
+ break;
+ }
+
+ return ret;
+}
+
+static unsigned long pnv_ioda2_get_table_size(__u32 page_shift,
+ __u64 window_size, __u32 levels)
+{
+ unsigned long tce_table_size, shift, ret;
+
+ if (!levels || (levels > POWERNV_IOMMU_MAX_LEVELS))
+ return -EINVAL;
+
+ if ((window_size > memory_hotplug_max()) || !is_power_of_2(window_size))
+ return -EINVAL;
+
+ tce_table_size = (window_size >> page_shift) * 8;
+ tce_table_size = max(0x1000UL, tce_table_size);
+
+ /* Allocate TCE table */
+ shift = ROUND_UP(ilog2(window_size) - page_shift, levels) / levels;
+ shift += 3;
+ shift = max_t(unsigned, shift, IOMMU_PAGE_SHIFT_4K);
+
+ ret = tce_table_size; /* tbl->it_userspace */
+ ret += pnv_get_tce_table_size(shift, levels, &tce_table_size);
+
+ return ret;
+}
+
static __be64 *pnv_alloc_tce_table(int nid,
unsigned shift, unsigned levels, unsigned long *left)
{
@@ -1452,6 +1503,8 @@ static long pnv_pci_ioda2_create_table(struct iommu_table_group *table_group,
return -ENOMEM;
tbl->it_indirect_levels = levels - 1;
+ tbl->it_allocated_size = pnv_ioda2_get_table_size(page_shift,
+ window_size, levels);
/* Setup linux iommu table */
pnv_pci_setup_iommu_table(tbl, addr, tce_table_size,
@@ -1681,6 +1734,7 @@ static long pnv_pci_ioda2_create_table_with_uas(
static struct iommu_table_group_ops pnv_pci_ioda2_ops = {
.set_ownership = pnv_ioda2_set_ownership,
+ .get_table_size = pnv_ioda2_get_table_size,
.create_table = pnv_pci_ioda2_create_table_with_uas,
.set_window = pnv_pci_ioda2_set_window,
.unset_window = pnv_pci_ioda2_unset_window,
This adds a way for the IOMMU user to know how much a new table will use so it can be accounted in the locked_vm limit before allocation happens. This stores the allocated table size in pnv_pci_ioda2_create_table() so the locked_vm counter can be updated correctly when a table is being disposed. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> --- arch/powerpc/include/asm/iommu.h | 5 +++ arch/powerpc/platforms/powernv/pci-ioda.c | 54 +++++++++++++++++++++++++++++++ 2 files changed, 59 insertions(+)