Re: Error when building kernel

"Bruce Dubbs" ([email protected] via alfs-discuss Mailing List) <[email protected]>
Newsgroups gmane.linux.lfs.automated
Message-ID <[email protected]>
On 9/17/23 15:47, Pierre Labastie ([email protected] via alfs-discuss Mailing 
List) wrote:
> On Sun, 2023-09-17 at 15:11 -0500, Bruce Dubbs wrote:
>> On 9/17/23 14:23, Pierre Labastie ([email protected] via alfs-discuss
>> Mailing
>> List) wrote:
>>> On Sun, 2023-09-17 at 13:13 -0500, Bruce Dubbs wrote:
>>>> On 9/17/23 02:10, Dragan ([email protected] via alfs-discuss Mailing List)
>>>> wrote:
>>>>> On 17/09/2023 08:04, Pierre Labastie ([email protected] via alfs-
>>>>> discuss
>>>>> Mailing List) wrote:
>>>>>> But error 124 is most likely due to a timeout, which most likely means
>>>>>> your
>>>>>> kernel config is not up to date, and "make oldconfig" was waiting for
>>>>>> some
>>>>>> input...
>>>>>> Previously, the build would have been stuck forever. Now it times out
>>>>>> after a
>>>>>> minute.
>>>>>>
>>>>>> Pierre
>>>>>
>>>>> Thanks. You were right!
>>>>>
>>>>> I have wrongly assumed that I will get option to run make oldconfig in
>>>>> the
>>>>> shell and
>>>>> then resume build of kernel!
>>>>>
>>>>>        If the kernel-config file is not up to date (missing options or
>>>>>        whatever), the kernel "make" will hang, waiting for input. So
>>>>>        run "make oldconfig" explicitly, with a timeout.
>>>>>            Fixes: #1736
>>>>
>>>> I'll note that I never try to build the kernel automatically.  There are
>>>> just
>>>> too
>>>> many issues with 'make oldconfig' unless you are using a known good config
>>>> file for
>>>> the current kernel version.
>>>
>>> You can prepare a kernel config before launching jhalfs, with the same
>>> kernel
>>> version as the one in lfs. Then there are not so many problems. But there
>>> may be
>>> some. Here is an example I ran into today:
>>> My host is debian 12, which has gcc-12. This version of gcc does not support
>>> creating debug sections compressed with zstd. So this choice is not
>>> proposed.
>>> But lfs has gcc 13, which does support that. So make oldconfig proposes this
>>> option as (NEW), then stops waiting for the user choice... Note that for
>>> some
>>> reason my kernel configs for VM's enable debug symbols in the kernel (a
>>> remnant
>>> of when I had to debug the kernel a few years ago). This problem does not
>>> exist
>>> if debugging is turned of.
>>>
>>> With the new behavior of timing out if the build system is stuck waiting for
>>> input, you are given a chance to enter chroot, configure the kernel again,
>>> save
>>> the .config to /sources/kernel-config, exit chroot, then restart make.
>>
>> That's a good point.  I still think the convenience of building the kernel in
>> jhalfs
>> vs manually building it in chroot after running jhalfs is minimal.  Just my
>> opinion.
> 
> Well, doing it automatically prevents typos (sometimes annoying to detect) when
> installing the files into /boot...

True.  That's a trade off.  I generally just copy/paste from a rendering of the book.

Not sure.  Does jhalfs update grub.cfg?  I sorta doubt it.
Does it use a separate partition for /boot if desired?

   -- Bruce

-- 
http://lists.linuxfromscratch.org/sympa/info/alfs-discuss
Unsubscribe: See the above information page
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.