[ANNOUNCE] makedumpfile 1.8.0
YAMAZAKI MASAMITSU(山崎 真光)
m-yamazaki2 at nec.com
Wed Nov 11 02:46:38 PST 2026
On 2026/11/01 6:19, Leonidas Spyropoulos wrote:
> On 30/10/2026 08:20, YAMAZAKI MASAMITSU(山崎 真光) wrote:
>> Hi,
>>
>> We're pleased to announce the release of makedumpfile 1.8.0.
>> Thank you everyone for your help to maintain the tool.
>>
>> Download:
>> The latest makedumpfile can be downloaded from the following page.
>> https://github.com/makedumpfile/makedumpfile/releases
>>
>> New features:
>> - Support for kernels up to v7.3 (x86_64)
>> - Extension module support based on BTF and kallsyms
>> - Add 5-level page table support for 4K pages with 52-bit VA (arm64)
>>
>> Commits since 1.7.9:
>> ae6c64c [v1.8.0] Update version (YAMAZAKI Masamitsu)
>> cecce00 [PATCH] Derive the mem_section mem_map mask from SIZE(page) (Nikunj Kela)
>> 505d15b [PATCH] README: Remove email addresses that will be gone and update information (Kazuhito Hagio)
>> 6772ca7 [PATCH v2 7/9] Add PG_INCLUDE_HEAD extension return status (Stephen Brennan)
>> f50904d [PATCH v2 6/9] Simplify arguments for page checks (Stephen Brennan)
>> fa9e082 [PATCH v2 5/9] Move page checks into makedumpfile.h (Stephen Brennan)
>> 53f8cea [PATCH v2 4/9] Introduce a stat for pages retained by extension (Stephen Brennan)
>> 8194fae [PATCH v2 3/9] Share page information with extension callbacks (Stephen Brennan)
>> 2c053fb [PATCH v2 2/9] Honor CFLAGS in extensions/Makefile (Stephen Brennan)
>> 82fc0fd [PATCH v2 1/9] Do not call extensions for tail pages (Stephen Brennan)
>> 755160b [PATCH 6/6] Ensure e_phentsize does not exceed buf size (Tao Liu)
>> 80f6586 [PATCH 5/6] Ensure get_smram_cpu_state() does not read beyond smram (Tao Liu)
>> 69585d6 [PATCH 4/6] Add an out-of-bounds check for info->valid_pages[] (Tao Liu)
>> d96b213 [PATCH 3/6] Prevent a multiply overflow in cache_init() (Tao Liu)
>> 93d65a2 [PATCH 2/6] Add an out-of-bounds check for si->block_table[] (Tao Liu)
>> e63708a [PATCH 1/6] Ensure page_desc does not exceed page_size (Tao Liu)
>> fe9858f [PATCH] Avoid glibc lazy-unwind machinery in parallel teardown (Maxi Saparov)
>> b7587f1 [PATCH] arm64: Add 5-level page table support for 4K pages with 52-bit VA (Xiaolei Wang)
>> 7da05f7 [PATCH v6 09/10] Doc: Add --extension option to makedumpfile manual (Tao Liu)
>> 55ae865 [PATCH v6 08/10] Add erase_sample extension as an data erase reference (Tao Liu)
>> ee345d5 [PATCH v6 07/10] Add sample extension as an example reference (Tao Liu)
>> 8c91a53 [PATCH v6 06/10] Add makedumpfile extensions support (Tao Liu)
>> cc6cd4d [PATCH v6 05/10] Implement kernel module's btf resolving (Tao Liu)
>> 8ee42c4 [PATCH v6 04/10] Implement kernel module's kallsyms resolving (Tao Liu)
>> 0a256c3 [PATCH v6 03/10] Implement kernel btf resolving (Tao Liu)
>> fd380d7 [PATCH v6 02/10] Implement kernel kallsyms resolving (Tao Liu)
>> fdf62f3 [PATCH v6 01/10] Reserve sections for makedumpfile and extenions (Tao Liu)
>> fd52214 [PATCH] Fix error and simplify eppic_scripts/proc_names_3_10_to_4_8.c (Kazuhito Hagio)
>> f50c015 [PATCH] Fix wrong exclusion of compound tail pages on Linux 7.1 and later (Kazuhito Hagio)
>> 863e2c4 [PATCH v2 2/2] riscv64: Fix huge page translation in vtop_riscv64() (Rui Qi)
>> 998b137 [PATCH v2 1/2] riscv64: Fix vaddr_to_paddr for vmalloc and vmemmap regions (Rui Qi)
>> 3fbe2cb [PATCH] Mark start of 1.8.0 development phase with version 1.7.9++ (Kazuhito Hagio)
>>
>> Description of makedumpfile:
>> The makedumpfile is a tool for creating a dumpfile from /proc/vmcore
>> with filtering out unnecessary pages for analysis and compressing the
>> remaining pages, to shorten the size of the dumpfile and the time of
>> creating it.
>> https://github.com/makedumpfile/makedumpfile
>>
>> Thanks,
>> Masa
>
> Hello Masa,
>
> As the Arch Linux maintainer of makedumpfile I have to raise the signing chain once more, as I did for 1.7.1 [1] and 1.7.7 [2].
>
> The 1.8.0 tag is signed by a key we have not seen before:
> 3E690EAEADB42DC30966523B970314D9A9FCC2EC
> (tagger: YAMAZAKI Masamitsu <m-yamazaki2 at nec.com>)
>
> 1.7.8 was signed by Masa's previous key, 4D1A09161401A7E773BA02A0A56B268DCE35849. From what I understand nothing connects the two keys, the new key is not published as a pubkey* tag, it is not on keys.openpgp.org nor keyserver.ubuntu.com, and it is not certified by any earlier key. The README FAQ (003) still says pubkey4 is the current key.
>
> Previously the handover for 1.7.8 was done correctly. The pubkey5 tag holding Masa's key was signed by Kazu's key E95D17F5, so anyone could verify that the existing maintainer had vouched for the new one. Without it, the only thing tying 3E690EAE... to the project is push access to GitHub.
>
> Could you please do the following:
> 1. Sign the new key with the old one (F4D1A091... or Kazu's E95D17F5),
> and publish it as a pubkey6 tag signed by the old key, e.g.:
> gpg -u F4D1A09161401A7E773BA02A0A56B268DCE35849 \
> --sign-key 3E690EAEADB42DC30966523B970314D9A9FCC2EC
> gpg --armor --export 3E690EAEADB42DC30966523B970314D9A9FCC2EC > new.asc
> git tag -s -u F4D1A09161401A7E773BA02A0A56B268DCE35849 \
> -m "makedumpfile release key: YAMAZAKI Masamitsu <m-yamazaki2 at nec.com>" \
> pubkey6 $(git hash-object -w new.asc)
> It would also help to upload the certified key to keys.openpgp.org.
> 2. Update README FAQ 003 to list the current key fingerprint(s).
> 3. Please do this before Kazu's nec.com addresses go away (commit
> 505d15b) as his key is the other link in the chain.
>
> If the old key is no longer available, please tell us.
>
> We documented our expectations for upstream signing in RFC 0046 [3].
> Under that policy we cannot ship 1.8.0 until a chain of trust for the current gpg key is established again. If keeping the chain is too much overhead, please inform us and would then stop verifying signatures and rely on the git tag and checksums only. Our preference of course is to keep the verification in place.
>
> Thank you,
>
> [1] https://www.mail-archive.com/kexec@lists.infradead.org/msg28322.html
> [2] https://www.mail-archive.com/kexec@lists.infradead.org/msg36050.html
> [3] https://gitlab.archlinux.org/archlinux/rfcs/-/blob/main/rfcs/0046-upstream-package-sources.md
Hello Leonidas,
My e-mail address has changed, so I apologize for inconvenience. I was delayed
because I ran into trouble configuring gpg.
I created pubkey6 following your method. There are no problems compared to
before pubkey5. I have signed there new key(m-yamazaki2) with the old
key (yamazaki-msmt). Additionally, I have updated the `README` file.
Regards,
Masa
More information about the kexec
mailing list