From: Julia Lawall <[email protected]> The cleanup code at the end of the function should also be carried out when the function only partly completes its work.
A simplified version of the semantic match that finds this problem is: (http://coccinelle.lip6.fr/) // <smpl> @r exists@ local idexpression struct page ** x; expression ra,rr; position p1,p2; @@ x = drm_malloc_ab at p1(...) ... when != x = rr when != drm_free_large(x,...) when != if (...) { ... drm_free_large(x,...) ...} if(...) { ... when != x = ra when forall when != drm_free_large(x,...) \(return <+...x...+>; \| return at p2...; \) } @script:python@ p1 << r.p1; p2 << r.p2; @@ cocci.print_main("drm_malloc_ab",p1) cocci.print_secs("return",p2) // </smpl> Signed-off-by: Julia Lawall <julia at diku.dk> --- This is only compile-tested. It may be that processing all of the pages is not correct when they have been partially treated. drivers/gpu/drm/i915/i915_gem.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 7ce3f35..ef000e6 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -467,8 +467,10 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, page = read_cache_page_gfp(mapping, offset >> PAGE_SHIFT, GFP_HIGHUSER | __GFP_RECLAIMABLE); - if (IS_ERR(page)) - return PTR_ERR(page); + if (IS_ERR(page)) { + ret = PTR_ERR(page); + goto out; + } if (do_bit17_swizzling) { slow_shmem_bit17_copy(page,
