drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c
authorZhao Liu <zhao1.liu@intel.com>
Sun, 3 Dec 2023 13:29:42 +0000 (21:29 +0800)
committerTvrtko Ursulin <tvrtko.ursulin@intel.com>
Fri, 15 Dec 2023 09:34:29 +0000 (09:34 +0000)
The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1], and this patch converts the call from
kmap_atomic() to kmap_local_page().

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption (the preemption is
disabled for !PREEMPT_RT case, otherwise it only disables migration).

With kmap_local_page(), we can avoid the often unwanted side effect of
unnecessary page faults or preemption disables.

In drm/i915/gem/selftests/huge_pages.c, function __cpu_check_shmem()
mainly uses mapping to flush cache and check the value. There're
2 reasons why __cpu_check_shmem() doesn't need to disable pagefaults
and preemption for mapping:

1. The flush operation is safe. Function __cpu_check_shmem() calls
drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush. Since
CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in
drm_clflush_virt_range(), the flush operation is global.

2. Any context switch caused by preemption or page faults (page fault
may cause sleep) doesn't affect the validity of local mapping.

Therefore, __cpu_check_shmem() is a function where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231203132947.2328805-5-zhao1.liu@linux.intel.com
drivers/gpu/drm/i915/gem/selftests/huge_pages.c

index 6b9f6cf50bf6b24b9bb5804b123db70adeb206d8..c9e6d77abab075e490ee434e0a50c3e9235b1400 100644 (file)
@@ -1082,7 +1082,7 @@ __cpu_check_shmem(struct drm_i915_gem_object *obj, u32 dword, u32 val)
                goto err_unlock;
 
        for (n = 0; n < obj->base.size >> PAGE_SHIFT; ++n) {
-               u32 *ptr = kmap_atomic(i915_gem_object_get_page(obj, n));
+               u32 *ptr = kmap_local_page(i915_gem_object_get_page(obj, n));
 
                if (needs_flush & CLFLUSH_BEFORE)
                        drm_clflush_virt_range(ptr, PAGE_SIZE);
@@ -1090,12 +1090,12 @@ __cpu_check_shmem(struct drm_i915_gem_object *obj, u32 dword, u32 val)
                if (ptr[dword] != val) {
                        pr_err("n=%lu ptr[%u]=%u, val=%u\n",
                               n, dword, ptr[dword], val);
-                       kunmap_atomic(ptr);
+                       kunmap_local(ptr);
                        err = -EINVAL;
                        break;
                }
 
-               kunmap_atomic(ptr);
+               kunmap_local(ptr);
        }
 
        i915_gem_object_finish_access(obj);