Bug#1142571: linux-modules-7.1.4+deb14-loong64: Unable to boot due to unloadable modules

Hi,

> 2026年7月24日 05:40，Miao Wang <[email protected]> 写道：
> 
> Hi,
> 
>> 2026年7月24日 05:02，Salvatore Bonaccorso <[email protected]> 写道：
>> 
>> Hi Miao,
>> 
>> On Thu, Jul 23, 2026 at 06:58:29PM +0800, Miao Wang wrote:
>>> Hi,
>>> 
>>>> 2026年7月23日 00:14，Miao Wang <[email protected]> 写道：
>>>> 
>>>> Control: tags -1 - moreinfo
>>>> 
>>>>> 2026年7月23日 00:10，Salvatore Bonaccorso <[email protected]> 写道：
>>>>> 
>>>>> Control: tags -1 + moreinfo
>>>>> 
>>>>> On Wed, Jul 22, 2026 at 07:50:40AM +0800, Miao Wang wrote:
>>>>>> Package: linux-modules-7.1.4+deb14-loong64
>>>>>> Version: 7.1.4-1
>>>>>> Severity: serious
>>>>>> Control: block -1 by 1142570
>>>>>> 
>>>>>> Dear maintainer,
>>>>>> 
>>>>>> The kernel modules in 7.1.4+deb14 for loong64 is unable to be loaded by the
>>>>>> kernel. The error in dmesg is:
>>>>>> 
>>>>>> kmod: xxxx: Unsupport relocation type 102, please add its support.
>>>>>> 
>>>>>> It is because a bug introduced in binutils, which inserts unexpected relocation
>>>>>> entries into ld -r outputs. This has been fixed by upstream.
>>>>> 
>>>>> So this is an issue in binutils only or what change has been applied
>>>>> to Linux? Is it applied in mainline and as well queued for 7.1.5?
>>>> 
>>>> This is an issue in binutils only. No changes is needed to be done
>>>> in Linux and nothing can be done in Linux to work around this.
>>>> 
>>> 
>>> I saw you have reassigned the bug to binutils. I believe that although the
>>> problem lies in binutils, the compiled binary packages are seriously
>>> broken, preventing the system from booting. I wonder whether we should keep
>>> a bug open for linux to prevent the problematic version from migrating
>>> from unstable to testing according to the policy and convention.
>> 
>> Yes I did so because the issue is not in src:linux. This would not
>> prevent in fact src:linux to migrate (which is a pity in itself since
>> we really want 7.1.4 to go to testing). But I talked to the release
>> team and they added a block hint so linux won't migrate until the
>> block lifted.
>> 
>> Can you please respond here once binutils is fixed?
>> 
> 
> Many thanks for your quick action. I have opened #1142570 to record this.
> I have seen patches for this in upstream, but there seems to be some
> confusion on this.

The bug in binutils has been fixed by doko in 2.47-1.

Cheers,

Miao Wang