linux-kernel - Potential heapify performance issue in dm-vdo

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <aC3riPZ6GV/dlr7Y@visitorckw-System-Product-Name>
Date: Wed, 21 May 2025 23:04:40 +0800
From: Kuan-Wei Chiu <visitorckw@...il.com>
To: Matthew Sakai <msakai@...hat.com>
Cc: dm-devel@...ts.linux.dev, linux-kernel@...r.kernel.org,
	jserv@...s.ncku.edu.tw
Subject: Potential heapify performance issue in dm-vdo

Hi Matthew,

Recently noticed that the current heapify method in min_heap.h may
degrade performance when the heap contains many compare-equal elements,
compared to the previous version.

In detail, the new heapify reduces the number of comparisons by about
50% when all elements are distinct. However, when all elements are
equal, the comparison count can degrade from O(1) to O(log n).

I don't have enough domain knowledge of dm-vdo, so I'd like to ask
whether it uses heaps with many compare-equal elements. If so, I'll
work on fixing the issue.

Regards,
Kuan-Wei