Re: [regression] microcode files missing in initramfs imgages from dracut (was Re: [PATCH] x86: Clean up remaining references to CONFIG_MICROCODE_AMD)

From: Antonio Feijoo <antonio.feijoo@suse.com>
To: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Linux regressions mailing list <regressions@lists.linux.dev>,
	lukas.bulwahn@gmail.com, dave.hansen@linux.intel.com,
	hpa@zytor.com, kernel-janitors@vger.kernel.org,
	linux-kernel@vger.kernel.org, mingo@redhat.com,
	tglx@linutronix.de, x86@kernel.org
Subject: Re: [regression] microcode files missing in initramfs imgages from dracut (was Re: [PATCH] x86: Clean up remaining references to CONFIG_MICROCODE_AMD)
Date: Fri, 24 Nov 2023 14:15:02 +0100	[thread overview]
Message-ID: <081ddbc7-8e3a-41c2-b361-3a03dfb3af12@suse.com> (raw)
In-Reply-To: <20231124121500.GAZWCTxML94BSYaQf3@fat_crate.local>

On 24/11/2023 13.15, Borislav Petkov wrote:
> On Fri, Nov 24, 2023 at 09:49:27AM +0100, Antonio Feijoo wrote:
>> As Linus said, the `check_kernel_config` stuff was implemented in 2014 and this
>> is not the only kernel config option that it's being checked by dracut
>> (CONFIG_ACPI_TABLE_UPGRADE, CONFIG_ACPI_INITRD_TABLE_OVERRIDE, CONFIG_RD_ZSTD),
>> although I agree that it's fragile if something changes. But adding in CC the
>> initramfs list (like you did), would be enough to prepare a simple fix in time.
> 
> Right, how about we give you a more reliable way to check functionality
> built into the kernel instead of grepping the .config file?
> 
> Like the ELF note thing, for example:
> 
> https://lore.kernel.org/r/20231122132419.GBZV4BA399sG2JRFAJ@fat_crate.local
> 
> The current thing is not ABI and will break everytime we change
> something and even if you fix it on time, older dracuts will still be
> broken.

The problem I see is having to add a new patch with a new note every time a
user space application requires new information to query. And also new dracuts
will be broken with older kernels that do not contain this info.
But (from a user space application point of view) if you (the kernel devs) are
ok with this approach, I don't see why we can at least get some info from there.

> 
>> The only problem I see in your patch is that we should also remove the
>> `--early-microcode` option, and dracut will fail if someone pass an option
>> available since 2013 (5f2c30d9bcd614d546d5c55c6897e33f88b9ab90) that would not
>> be recognized now (and by failing, I mean it will not build an initramfs if an
>> unrecognized option is passed).
> 
> Ah ok, --early-microcode becomes a no-op with my change. Sure.
> 
>> Please, submit it to https://github.com/dracutdevs/dracut, so more people can
>> see it and discuss it. Thank you.
> 
> I presume I should read this first:
> 
> https://github.com/dracutdevs/dracut/blob/master/docs/HACKING.md
> 
> and send a github pull request?

Yes, that would be enough.

> Anything else I need to pay attention to when sending dracut patches?

Just follow the Conventional Commit style for the commit messages, but that's
also specified in the HACKING.md doc.

> Or is there also an old school mailing list where I can send the patch
> to?
> 
> :-)

Unfortunately no. All the development process was moved to github.

> 
> Thx.
> 

Thank you,

Best regards.

-- 
Antonio Álvarez Feijoo
System Boot and Init
SUSE