Re: git: dc6c193bb4a6 - main - loader.efi: Add xz decompression

From: Bjoern A. Zeeb <bz_at_freebsd.org>
Date: Sun, 12 Jul 2026 01:17:54 UTC
On Sat, 11 Jul 2026, Warner Losh wrote:

> On Sat, Jul 11, 2026 at 2:47 PM Bjoern A. Zeeb <bz@freebsd.org> wrote:
>
>> On Sat, 11 Jul 2026, Warner Losh wrote:
>>
>>> The branch main has been updated by imp:
>>>
>>> URL:
>> https://cgit.FreeBSD.org/src/commit/?id=dc6c193bb4a67e2e9ee97d1e3ac3e2950bd16979
>>>
>>> commit dc6c193bb4a67e2e9ee97d1e3ac3e2950bd16979
>>> Author:     Warner Losh <imp@FreeBSD.org>
>>> AuthorDate: 2026-07-11 14:44:40 +0000
>>> Commit:     Warner Losh <imp@FreeBSD.org>
>>> CommitDate: 2026-07-11 19:28:47 +0000
>>>
>>>    loader.efi: Add xz decompression
>>>
>>>    We can now decompress .xz compressed memory disks, like
>>>    FreeBSD-15.1-RELEASE-amd64-disc1.iso.xz
>>
>> Thank you!  Helps me a lot!
>>
>> We need to make sure the .xz is correctly padded if needed:
>> see 4bf34c597c44a2393ac5ec483e9cecac128d8202
>>
>
> Earlier versions of my xz stuff had an off by one error that would trigger
> that warning. But I fixed that so we wouldn't lose the last byte. If there's
> still a partial sector at the end, it should be truncated, no?

I used to pass a foo.xz file in the past in as an md_image and the foo.xz
wasn't properly padded as xz doesn't pad unlike gzip and bzip -- at least
that was my understanding, so that file was corrupt when trying to read it.
That's what triggered the above commit by kib.

What I realize now is that we may now transparently uncompress the xz first
and then pass the uncompressed version to md.  Is there a way to prevent that?

For an mfs_root I would love to have the transparent uncompressions but couldn't
so far and that case should be fixed now and will defintiively save me cycles,
but for a general purpose other "file" I many not want it to be decompressed
just stay there as a read-only copy as-is.

/bz

-- 
Bjoern A. Zeeb                                                     r15:7