From: Jeremy Linton <jeremy.linton@arm.com>
To: Masahiro Yamada <masahiroy@kernel.org>
Cc: linux-kbuild@vger.kernel.org, nathan@kernel.org,
nicolas.schier@linux.dev, linux-kernel@vger.kernel.org,
Ard Biesheuvel <ardb@kernel.org>
Subject: Re: [PATCH] scripts: add zboot support to extract-vmlinux
Date: Tue, 8 Jul 2025 16:30:44 -0500 [thread overview]
Message-ID: <9ddd8122-a2d7-43f3-a0ec-d4d7a0450d2e@arm.com> (raw)
In-Reply-To: <CAK7LNAR=LaMQLqqhG86T9Vk6tZYOYzMsvfm_FtK8sBdL5CGY+A@mail.gmail.com>
Hi,
On 7/7/25 6:35 PM, Masahiro Yamada wrote:
> On Tue, Jul 8, 2025 at 6:29 AM Jeremy Linton <jeremy.linton@arm.com> wrote:
>>
>> Hi,
>>
>> Thanks for looking at this.
>>
>> On 6/24/25 12:56 PM, Masahiro Yamada wrote:
>>> On Tue, Jun 17, 2025 at 1:09 AM Jeremy Linton <jeremy.linton@arm.com> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Thanks for looking at this.
>>>>
>>>> On 6/7/25 11:04 AM, Masahiro Yamada wrote:
>>>>> On Fri, May 23, 2025 at 2:29 AM Jeremy Linton <jeremy.linton@arm.com> wrote:
>>>>>>
>>>>>> Zboot compressed kernel images are used for arm kernels on various
>>>>>> distros.
>>>>>
>>>>> Are you talking about arm 32 bit here?
>>>>> (arch/arm/boot/zImage)
>>>>
>>>> No, it should be arm64.
>>>>
>>>>>
>>>>>> extract-vmlinux fails with those kernels because the wrapped image is
>>>>>> another PE. While this could be a bit confusing, the tools primary
>>>>>> purpose of unwrapping and decompressing the contained vmlinux image
>>>>>> makes it the obvious place for this functionality.
>>>>>>
>>>>>> Add a 'file' check in check_vmlinux() that detects a contained PE
>>>>>> image before trying readelf. Recent file implementations output
>>>>>> something like:
>>>>>>
>>>>>> "Linux kernel ARM64 boot executable Image, little-endian, 4K pages"
>>>>>
>>>>> Are you talking about arm64 here?
>>>>>
>>>>> I am confused, as arm64 adopts a simple-compressed image.
>>>>
>>>> No, there is a CONFIG_EFI_ZBOOT, which is a EFI/PE image which self
>>>> decompresses a contained kernel similar to x86, but is for !x86 EFI
>>>> architectures. This patch extends this utility to work for those images
>>>> as well.
>>>
>>> The commit description does not explain why this is useful.
>>>
>>> Extracing vmlinux ELF is useful for debugging purposes.
>>
>> Right,
>>
>>>
>>> In this case, the extracted file is
>>> arch/arm64/boot/vmlinux.bin, which is just a (zero-padded) binary.
>>
>> $ file vmlinux.bin
>> vmlinux.bin: Linux kernel ARM64 boot executable Image, little-endian, 4K
>> pages
>> $ readpe -S vmlinux.bin
>> Sections
>> Section
>> Name: .text
>> Virtual Size: 0x2860000 (42336256 bytes)
>> Virtual Address: 0x10000
>> Size Of Raw Data: 0x2860000 (42336256 bytes)
>> Pointer To Raw Data: 0x10000
>> Number Of Relocations: 0
>> Characteristics: 0x60000020
>> Characteristic Names
>> IMAGE_SCN_CNT_CODE
>> IMAGE_SCN_MEM_EXECUTE
>> ...(trimming remainder of output)
>>
>> Its another complete PE boot image which can be used by UEFI/grub/etc as
>> well as any PE debugging and analysis utilities.
>>
>> So, this change effectively removes the zboot wrapper. The resulting
>> image is still useful for a certain amount of debugging (ex string
>> extraction, manually matching crash points, version checking, etc) in a
>> distro based environment where the user doesn't have a kernel build tree
>> handy. As well as any boot debugging, which could be caused by the ZBOOT
>> wrapper itself, and I'm sure a long list of other things.
>
> OK. Then, please add a little more context regarding
> how this is useful in the commit description of v2.
Sure,
>
> Please fix the first line of the description:
>
> used for arm kernels -> used for arm64 kernels
Right,
>
>
> Lastly, how about this implementation?
>
> check_vmlinux()
> {
> if file $1 | grep -q 'Linux kernel.*boot executable Image' ||
> readelf -h $1 > /dev/null 2>&1; then
> cat $1
> exit 0
> fi
> }
That works too.
>
>
> I used the -q option for grep instead of the redirection.
>
>
> Masahiro
next prev parent reply other threads:[~2025-07-08 21:30 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-22 17:29 Jeremy Linton
2025-05-22 23:10 ` Nathan Chancellor
2025-05-23 15:03 ` Jeremy Linton
2025-06-07 16:04 ` Masahiro Yamada
2025-06-16 16:07 ` Jeremy Linton
2025-06-24 17:56 ` Masahiro Yamada
2025-07-07 21:29 ` Jeremy Linton
2025-07-07 23:35 ` Masahiro Yamada
2025-07-08 21:30 ` Jeremy Linton [this message]
2025-07-07 21:31 ` Jeremy Linton
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=9ddd8122-a2d7-43f3-a0ec-d4d7a0450d2e@arm.com \
--to=jeremy.linton@arm.com \
--cc=ardb@kernel.org \
--cc=linux-kbuild@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=masahiroy@kernel.org \
--cc=nathan@kernel.org \
--cc=nicolas.schier@linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®