> On Jan 10, 2026, at 9:10 PM, Enji Cooper (yaneurabeya)
> <[email protected]> wrote:
>
> Hi Olivier,
>
>> On Jan 9, 2026, at 8:44 AM, Olivier Certner <[email protected]> wrote:
>>
>> Hi Enji,
>>
>>> The branch main has been updated by ngie:
>>>
>>> URL:
>>> https://cgit.FreeBSD.org/src/commit/?id=f224591746bdaf14ad5f63de4738a3146cc2f55f
>>>
>>> commit f224591746bdaf14ad5f63de4738a3146cc2f55f
>>> Author: Enji Cooper <[email protected]>
>>> AuthorDate: 2026-01-04 08:27:57 +0000
>>> Commit: Enji Cooper <[email protected]>
>>> CommitDate: 2026-01-06 01:43:04 +0000
>>>
>>> Add ASMC_DEBUG make option
>>
>> This commit broke the i386 build. I've just committed a fix. Also, I've
>> restored 'device vt_efifb' in NOTES (under 'x86/' instead of 'amd64/'), as
>> AFAIK, NOTES should list all possible devices/options.
>
> re: vt_efifb: what I did was removed the duplicate entry, since it was
> already in the x86 one:
> https://cgit.freebsd.org/src/tree/sys/x86/conf/NOTES#n332 .
Admittedly, my commit message re: it already being in GENERIC was wrong. It
being in x86/NOTES is what mattered.
>> Please make sure you also MFC the two following commits (67599eef01f5
>> ("sys/x86/NOTES: Add vt_efifb"), fde9fe18219f i386 ("Fix kernel compilation
>> after introduction of ASMC_DEBUG option")) along with the original one.
>
> Sure — although the correct fix would be to not build the driver at all on
> i386, since it doesn’t support 32-bit CPUs. Apple sold very few i386-based
> Intel chips before they transitioned to amd64-only ones (it was only a matter
> of 8-9 months IIRC) and asmc(4) doesn’t support any of the 32-bit chips (I
> checked).
Cheers,
-Enji
