Re: linux-next: build failure after merge of the bitmap tree

[Yury Norov <yury.norov@gmail.com>]

On Tue, Sep 27, 2022 at 12:53:48AM +0100, broonie@kernel.org wrote:
> Hi all,
> 
> After merging the bitmap tree, today's linux-next build (x86
> allmodconfig) failed like this:

Hmm, this weird. I checked the next-20220923, and the drivers' code
mentioned in the log differs from what I see, and looks correct.
bitmap_weight() definition hasn't been changed in bitmap-for-next
patches.

Allmodconfig build looks good to me.

Check what I see in next-20220923 below.

Thanks,
Yury
 
> /tmp/next/build/drivers/net/ethernet/marvell/octeontx2/af/rvu_npc_hash.c:492:3: note: in expansion of macro 'dev_err'
>   492 |   dev_err(rvu->dev, "%s: No space in id bitmap (%lu)\n",
>       |   ^~~~~~~
> /tmp/next/build/drivers/net/ethernet/marvell/octeontx2/af/rvu_npc_hash.c:492:51: note: format string is defined here
>   492 |   dev_err(rvu->dev, "%s: No space in id bitmap (%lu)\n",
>       |                                                 ~~^
>       |                                                   |
>       |                                                   long unsigned int
>       |                                                 %u

                dev_err(rvu->dev, "%s: No space in id bitmap (%d)\n",
                        __func__, bitmap_weight(table->id_bmap, table->tot_ids));

> /tmp/next/build/drivers/net/ethernet/marvell/octeontx2/af/rvu_npc_hash.c:525:22: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'unsigned int' [-Werror=format=]
>   525 |   dev_info(rvu->dev, "%s: No space in exact cam table, weight=%lu\n", __func__,
>       |                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

                dev_info(rvu->dev, "%s: No space in exact cam table, weight=%u\n", __func__,
                         bitmap_weight(table->cam_table.bmap, table->cam_table.depth));jj

> /tmp/next/build/drivers/gpu/drm/i915/gt/intel_sseu.c:867:39: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'unsigned int' [-Werror=format=]
>   867 |   seq_printf(m, "  %s Geometry DSS: %lu\n", type,
>       |                                     ~~^
>       |                                       |
>       |                                       long unsigned int
>       |                                     %u
>   868 |       bitmap_weight(sseu->geometry_subslice_mask.xehp,
>       |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>       |       |
>       |       unsigned int
>   869 |       XEHP_BITMAP_BITS(sseu->geometry_subslice_mask)));

                seq_printf(m, "  %s Geometry DSS: %u\n", type,
                           bitmap_weight(sseu->geometry_subslice_mask.xehp,
                                         XEHP_BITMAP_BITS(sseu->geometry_subslice_mask)));


>       |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> /tmp/next/build/drivers/gpu/drm/i915/gt/intel_sseu.c:870:38: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'unsigned int' [-Werror=format=]
>   870 |   seq_printf(m, "  %s Compute DSS: %lu\n", type,
>       |                                    ~~^
>       |                                      |
>       |                                      long unsigned int
>       |                                    %u
>   871 |       bitmap_weight(sseu->compute_subslice_mask.xehp,
>       |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>       |       |
>       |       unsigned int
>   872 |       XEHP_BITMAP_BITS(sseu->compute_subslice_mask)));

                           bitmap_weight(sseu->compute_subslice_mask.xehp,
                                         XEHP_BITMAP_BITS(sseu->compute_subslice_mask)));

>       |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> cc1: all warnings being treated as errors
> 
> These look like they have probably been there for a while but I've used
> a slightly different compiler version to Stephen so they're showing up
> now once the bitmap tree is merged.  I will have a proper look tomorrow
> hopefully but for now I've dropped the bitmap tree as it's getting very
> late and it is likely I'll not finish the -next run today at all.