Re: [built failure] aicdb.h / __udivdi3

[Randy Dunlap]

On 9/23/26 9:10 AM, David Laight wrote:
> On Wed, 23 Sep 2026 15:41:59 +0200
> Manuel Ebner <manuelebnerli@xxxxxxxxxxx> wrote:
> 
>> On Tue, 2026-09-22 at 12:25 -0700, Randy Dunlap wrote:
>>> Hi--
>>>
>>> On 9/22/26 11:13 AM, Manuel Ebner wrote:  
>>>> Hi,
>>>>
>>>> I did some built testing with randconf and got some build errors.
>>>>
>>>> First error is:
>>>>   
>>>>> $ make KCFLAGS=-Werror -j8 > make_output.txt
>>>>> aicasm_symbol.c:47:10: fatal error: aicdb.h: No such file or directory
>>>>>    47 | #include "aicdb.h"
>>>>>       |          ^~~~~~~~~  
>>>>
>>>> This could be fixed by installing libdb-devel.
>>>> But than I got this error with the same configuration:
>>>>   
>>>>> $ make KCFLAGS=-Werror -j8 > make_output.txt
>>>>> date
>>>>> drivers/scsi/aic7xxx/aicasm/aicasm: 880 instructions used
>>>>> drivers/scsi/aic7xxx/aicasm/aicasm: 25 of 37 register definitions used
>>>>> /usr/bin/ld.bfd: fs/exfat/super.o: in function `exfat_set_vol_flags.isra.0':
>>>>> super.c:(.text+0x622): undefined reference to `__udivdi3'
>>>>> make[2]: *** [scripts/Makefile.vmlinux:72: vmlinux.unstripped] Error 1
>>>>> make[1]: *** [/home/manuel/skripts/next/Makefile:1390: vmlinux] Error 2
>>>>> make: *** [Makefile:248: __sub-make] Error 2  
>>>>
>>>> What can I do about it?
>>>> Is it me/ my system or the kernel?  
>>>
>>> For the fs/exfat/ build error:
>>>
>>> What is the target (x86_64, i386, arm64, arm, etc.)?  
>>
>> My system is x86_64.
>> But i guess this means the target is 32 bit:
>>> CONFIG_X86_32=y  
>> The rest of the .config is below.
>>
>>> What is the compiler and its version?  
>>
>> gcc (GCC) 16.2.1 20260819 (Red Hat 16.2.1-2)
>>
>>> That error usually means that the target is a 32-bit CPU
>>> and the code in question is doing an unsupported divide or
>>> modulus function.  
>>
>> Can the dependencies in the .config be improved to to make this issue impossible?
>>
>>> It possibly should be using a kernel macro from <linux/math.h>
>>> or other headers.  
>>
>> I figured I didn't successfully checkout next but this shouldn't change this error.
> 
> I just tried to build the .config from below against a 'quite recent' linux-next.
> If fails for a couple of unrelated reasons - I edited the config to fix them.
> But there isn't a divide in exfat_set_vol_flags().
> I can't see one in the source either.
> The usual 'next stage' is to disassemble the .o file and work out which
> source line is responsible.

I also can build exfat fs successfully with this .config file.
And as David did, I had a few other build issues aside from exfat fs.

-- 
~Randy