Do not count buffer objects that have no backing pages, including imported
buffers where pages are set by VM faults triggered by userspace or pinned
by other drivers. Instead, return information about actual memory used by
the NPU.
Counting imported buffers results in incorrect calculations when
the same pages are counted multiple times, giving overly high
results.
Fixes: 7bfc9fa99580 ("accel/ivpu: Expose NPU memory utilization info in sysfs")
Signed-off-by: Karol Wachowski <[email protected]>
---
drivers/accel/ivpu/ivpu_sysfs.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/accel/ivpu/ivpu_sysfs.c b/drivers/accel/ivpu/ivpu_sysfs.c
index 268ab7744a8b..d250a10caca9 100644
--- a/drivers/accel/ivpu/ivpu_sysfs.c
+++ b/drivers/accel/ivpu/ivpu_sysfs.c
@@ -63,7 +63,8 @@ npu_memory_utilization_show(struct device *dev, struct
device_attribute *attr, c
mutex_lock(&vdev->bo_list_lock);
list_for_each_entry(bo, &vdev->bo_list, bo_list_node)
- total_npu_memory += bo->base.base.size;
+ if (ivpu_bo_is_resident(bo))
+ total_npu_memory += ivpu_bo_size(bo);
mutex_unlock(&vdev->bo_list_lock);
return sysfs_emit(buf, "%lld\n", total_npu_memory);
--
2.43.0
