Re: [PATCH v2 2/2] kernel.h: Disable -Wreturn-stack-address for _THIS_IP_

[Nick Desaulniers]

Another strange error from 0-day bot.  The source file does not make
use of my added macros, so not sure how they could generate such a
warning.
On Tue, Jul 31, 2018 at 6:54 AM kbuild test robot <lkp@xxxxxxxxx> wrote:
>
> Hi Nick,
>
> Thank you for the patch! Yet something to improve:
>
> [auto build test ERROR on linus/master]
> [also build test ERROR on v4.18-rc7 next-20180727]
> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
>
> url:    https://github.com/0day-ci/linux/commits/Nick-Desaulniers/compiler-clang-h-Add-CLANG_VERSION-and-__diag-macros/20180731-161932
> config: x86_64-randconfig-s1-07312048 (attached as .config)
> compiler: gcc-6 (Debian 6.4.0-9) 6.4.0 20171026
> reproduce:
>         # save the attached .config to linux build tree
>         make ARCH=x86_64
>
> All errors (new ones prefixed by >>):
>
>    drivers/gpu//drm/i915/i915_gem.c: In function '__i915_gem_object_set_pages':
> >> drivers/gpu//drm/i915/i915_gem.c:2697:1: error: expected expression before '#pragma'
>      GEM_BUG_ON(!HAS_PAGE_SIZES(i915, obj->mm.page_sizes.sg));
>     ^~~
>
> vim +2697 drivers/gpu//drm/i915/i915_gem.c
>
> 03ac84f18 Chris Wilson 2016-10-28  2658
> 03ac84f18 Chris Wilson 2016-10-28  2659  void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
> a5c081662 Matthew Auld 2017-10-06  2660                                  struct sg_table *pages,
> 84e8978e6 Matthew Auld 2017-10-09  2661                                  unsigned int sg_page_sizes)
> 03ac84f18 Chris Wilson 2016-10-28  2662  {
> a5c081662 Matthew Auld 2017-10-06  2663         struct drm_i915_private *i915 = to_i915(obj->base.dev);
> a5c081662 Matthew Auld 2017-10-06  2664         unsigned long supported = INTEL_INFO(i915)->page_sizes;
> a5c081662 Matthew Auld 2017-10-06  2665         int i;
> a5c081662 Matthew Auld 2017-10-06  2666
> 1233e2db1 Chris Wilson 2016-10-28  2667         lockdep_assert_held(&obj->mm.lock);
> 03ac84f18 Chris Wilson 2016-10-28  2668
> 03ac84f18 Chris Wilson 2016-10-28  2669         obj->mm.get_page.sg_pos = pages->sgl;
> 03ac84f18 Chris Wilson 2016-10-28  2670         obj->mm.get_page.sg_idx = 0;
> 03ac84f18 Chris Wilson 2016-10-28  2671
> 03ac84f18 Chris Wilson 2016-10-28  2672         obj->mm.pages = pages;
> 2c3a3f44d Chris Wilson 2016-11-04  2673
> 2c3a3f44d Chris Wilson 2016-11-04  2674         if (i915_gem_object_is_tiled(obj) &&
> f2123818f Chris Wilson 2017-10-16  2675             i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) {
> 2c3a3f44d Chris Wilson 2016-11-04  2676                 GEM_BUG_ON(obj->mm.quirked);
> 2c3a3f44d Chris Wilson 2016-11-04  2677                 __i915_gem_object_pin_pages(obj);
> 2c3a3f44d Chris Wilson 2016-11-04  2678                 obj->mm.quirked = true;
> 2c3a3f44d Chris Wilson 2016-11-04  2679         }
> a5c081662 Matthew Auld 2017-10-06  2680
> 84e8978e6 Matthew Auld 2017-10-09  2681         GEM_BUG_ON(!sg_page_sizes);
> 84e8978e6 Matthew Auld 2017-10-09  2682         obj->mm.page_sizes.phys = sg_page_sizes;
> a5c081662 Matthew Auld 2017-10-06  2683
> a5c081662 Matthew Auld 2017-10-06  2684         /*
> 84e8978e6 Matthew Auld 2017-10-09  2685          * Calculate the supported page-sizes which fit into the given
> 84e8978e6 Matthew Auld 2017-10-09  2686          * sg_page_sizes. This will give us the page-sizes which we may be able
> 84e8978e6 Matthew Auld 2017-10-09  2687          * to use opportunistically when later inserting into the GTT. For
> 84e8978e6 Matthew Auld 2017-10-09  2688          * example if phys=2G, then in theory we should be able to use 1G, 2M,
> 84e8978e6 Matthew Auld 2017-10-09  2689          * 64K or 4K pages, although in practice this will depend on a number of
> 84e8978e6 Matthew Auld 2017-10-09  2690          * other factors.
> a5c081662 Matthew Auld 2017-10-06  2691          */
> a5c081662 Matthew Auld 2017-10-06  2692         obj->mm.page_sizes.sg = 0;
> a5c081662 Matthew Auld 2017-10-06  2693         for_each_set_bit(i, &supported, ilog2(I915_GTT_MAX_PAGE_SIZE) + 1) {
> a5c081662 Matthew Auld 2017-10-06  2694                 if (obj->mm.page_sizes.phys & ~0u << i)
> a5c081662 Matthew Auld 2017-10-06  2695                         obj->mm.page_sizes.sg |= BIT(i);
> a5c081662 Matthew Auld 2017-10-06  2696         }
> a5c081662 Matthew Auld 2017-10-06 @2697         GEM_BUG_ON(!HAS_PAGE_SIZES(i915, obj->mm.page_sizes.sg));
> f2123818f Chris Wilson 2017-10-16  2698
> f2123818f Chris Wilson 2017-10-16  2699         spin_lock(&i915->mm.obj_lock);
> f2123818f Chris Wilson 2017-10-16  2700         list_add(&obj->mm.link, &i915->mm.unbound_list);
> f2123818f Chris Wilson 2017-10-16  2701         spin_unlock(&i915->mm.obj_lock);
> 03ac84f18 Chris Wilson 2016-10-28  2702  }
> 03ac84f18 Chris Wilson 2016-10-28  2703
>
> :::::: The code at line 2697 was first introduced by commit
> :::::: a5c08166265adc172a4cbde8ed26a1a96ce77fb7 drm/i915: introduce page_size members
>
> :::::: TO: Matthew Auld <matthew.auld@xxxxxxxxx>
> :::::: CC: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
>
> ---
> 0-DAY kernel test infrastructure                Open Source Technology Center
> https://lists.01.org/pipermail/kbuild-all                   Intel Corporation



-- 
Thanks,
~Nick Desaulniers