On 6/7/20 7:15 AM, Syed Nayyar Waris wrote:
> On Fri, Jun 5, 2020 at 5:54 PM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
>> On Fri, Jun 05, 2020 at 02:12:54AM +0530, Syed Nayyar Waris wrote:
>>> On Sun, May 31, 2020 at 12:50 AM kbuild test robot <lkp@intel.com> wrote:
>>>>>> WARNING: modpost: lib/test_bitmap.o(.data+0xe80): Section mismatch in reference from the variable clump_test_data to the variable .init.rodata:clump_exp1
>>>> The variable clump_test_data references
>>>> the variable __initconst clump_exp1
>>>> If the reference is valid then annotate the
>>>> variable with or __refdata (see linux/init.h) or name the variable:
>>>>
>>>> --
>>>>>> WARNING: modpost: lib/test_bitmap.o(.data+0xec8): Section mismatch in reference from the variable clump_test_data to the variable .init.rodata:clump_exp2
>>>> The variable clump_test_data references
>>>> the variable __initconst clump_exp2
>>>> If the reference is valid then annotate the
>>>> variable with or __refdata (see linux/init.h) or name the variable:
>>>>
>>>> --
>>>>>> WARNING: modpost: lib/test_bitmap.o(.data+0xf10): Section mismatch in reference from the variable clump_test_data to the variable .init.rodata:clump_exp3
>>>> The variable clump_test_data references
>>>> the variable __initconst clump_exp3
>>>> If the reference is valid then annotate the
>>>> variable with or __refdata (see linux/init.h) or name the variable:
>>>>
>>>> --
>>>>>> WARNING: modpost: lib/test_bitmap.o(.data+0xf58): Section mismatch in reference from the variable clump_test_data to the variable .init.rodata:clump_exp4
>>>> The variable clump_test_data references
>>>> the variable __initconst clump_exp4
>>>> If the reference is valid then annotate the
>>>> variable with or __refdata (see linux/init.h) or name the variable:
>>> I am unable to reproduce the compilation warning.
>> You have to enable section mismatch checker.
>>
>>> I ran the command:
>>> make W=1 C=1 ARCH=x86_64 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'  lib/
>>>
>>> But the compilation warning didn't show up. Can anyone please point to me
>>> what I am doing wrong here? How shall I reproduce the warning? Thanks !
>> You put some data into init section of the object, while you are trying to
>> access it from non-init one. It's easy-to-fix issue.
>>
>> --
>> With Best Regards,
>> Andy Shevchenko
> Thanks! I have made code changes for the above warning. Actually I am
> still unable to reproduce the compilation warning. But I believe the
> following code fix will fix the compilation warning:
>
> In file lib/test_bitmap.c
>
> @@ -692,7 +692,7 @@ struct clump_test_data_params {
>          unsigned long const *exp;
>   };
>
> -struct clump_test_data_params clump_test_data[] =
> +static struct clump_test_data_params clump_test_data[] __initdata =
>                                          { {{0}, 2, 0, 64, 8, clump_exp1},
>                                          {{0}, 8, 2, 240, 24, clump_exp2},
>                                          {{0}, 8, 10, 240, 30, clump_exp3},
>
>
>
> Let me know if I should submit a new patchset (v8) for
> 'for_each_set_clump' including above code fix.
>
> Just to share how I attempted to reproduce the warning (but unsuccessful):
>
> Step 1: Use the config file in attachment. Download, extract, rename
> file to .config at the root of the tree.
> Step 2: '$ make lib/'
> No warning reproduced after above step 2.
> Step 3: '$ make W=1 C=1 ARCH=x86_64 CF='-fdiagnostic-prefix
> -D__CHECK_ENDIAN__' lib/'
> After step 3 I got error in build:
> scripts/kconfig/conf  --syncconfig Kconfig
>    CHECK   scripts/mod/empty.c
> No such file: asan-globals=1
> scripts/Makefile.build:266: recipe for target 'scripts/mod/empty.o' failed
> make[1]: *** [scripts/mod/empty.o] Error 1
> Makefile:1147: recipe for target 'prepare0' failed
> make: *** [prepare0] Error 2
>
> The command in above step 3 was mentioned in the bot mail.
>
> Regards
> Syed Nayyar Waris
>

Hi Syed Nayyar Waris,

We can reproduce the warning with the steps in original report,
you may need to build the whole kernel instead of the 'lib'.

Best Regards,
Rong Chen