When we detect consecutive allocation of pages swap them to avoid accidentally freeing them as huge page.
v2: use swap
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index c82d94cbbabc..b6c5148607e9 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -921,6 +921,10 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags, return -ENOMEM; }
+ /* Swap the pages if we detect consecutive order */ + if (i && pages[i - 1] == p - 1) + swap(p, pages[i - 1]); + pages[i++] = p; --npages; }
When we detect consecutive allocation of pages swap them to avoid accidentally freeing them as huge page.
v2: use swap
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index b6c5148607e9..29652b2615a8 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -953,8 +953,12 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags, r = ttm_page_pool_get_pages(pool, &plist, flags, cstate, npages - count, 0);
- list_for_each_entry(p, &plist, lru) + list_for_each_entry(p, &plist, lru) { + /* Swap the pages if we detect consecutive order */ + if (count && pages[count - 1] == p - 1) + swap(p, pages[count - 1]); pages[count++] = p; + }
if (r) { /* If there is any pages in the list put them back to
-----Original Message----- From: Christian König [mailto:ckoenig.leichtzumerken@gmail.com] Sent: Monday, December 04, 2017 8:13 PM To: dri-devel@lists.freedesktop.org; amd-gfx@lists.freedesktop.org; He, Roger Hongbo.He@amd.com; ckoenig.leichtzumerken@gmail.com Subject: [PATCH 1/2] drm/ttm: swap consecutive allocated cached pages v2
When we detect consecutive allocation of pages swap them to avoid accidentally freeing them as huge page.
v2: use swap
Signed-off-by: Christian König christian.koenig@amd.com --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index c82d94cbbabc..b6c5148607e9 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -921,6 +921,10 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags, return -ENOMEM; }
How about add first index as below?
first = i; //init first here while (npages) { p = alloc_page(gfp_flags); if (!p) { pr_debug("Unable to allocate page\n"); return -ENOMEM; } /* Swap the pages if we detect consecutive order */ if (i > first && pages[i - 1] == p - 1) //change to (i > first), because I worry if npages is 513, the first regular page is exactly consecutive with last page of huge page. Then swap them is not what I want here. swap(p, pages[i - 1]);
pages[i++] = p; --npages; }
pages[i++] = p; --npages; } -- 2.11.0
dri-devel@lists.freedesktop.org
-
Christian König -
He, Roger