NetBSD Problem Report #59101

From www@netbsd.org  Mon Feb 24 21:59:36 2025
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
	 key-exchange X25519 server-signature RSA-PSS (2048 bits)
	 client-signature RSA-PSS (2048 bits))
	(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 6B49F1A923A
	for <gnats-bugs@gnats.NetBSD.org>; Mon, 24 Feb 2025 21:59:36 +0000 (UTC)
Message-Id: <20250224215935.684321A923D@mollari.NetBSD.org>
Date: Mon, 24 Feb 2025 21:59:35 +0000 (UTC)
From: campbell+netbsd@mumble.net
Reply-To: campbell+netbsd@mumble.net
To: gnats-bugs@NetBSD.org
Subject: cpuset.c bugs
X-Send-Pr-Version: www-1.0

>Number:         59101
>Category:       lib
>Synopsis:       cpuset.c bugs
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    lib-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Feb 24 22:00:00 +0000 2025
>Last-Modified:  Mon Feb 24 22:15:01 +0000 2025
>Originator:     Taylor R Campbell
>Release:        current, 10, 9, ...
>Organization:
The NetCPU Foundaset
>Environment:
>Description:
1. common/lib/libc/sys/cpuset.c is all under #ifndef _KERNEL, so it has no reason to live there; it should live instead in lib/libc/

2. missing memory barriers
>How-To-Repeat:
code inspection
>Fix:
1. move from common/lib/libc/ to lib/libc/

2. change `if (cpuset_size == 0) { ...; cpuset_size = CPUSET_SIZE(cpuset_nentries); }' to `if (atomic_load_acquire(&cpuset_size) == 0) { ...; atomic_store_release(&cpuset_size, CPUSET_SIZE(cpuset_nentries)); }' (except we don't have atomic_store_release or atomic_load_acquire exposed in userland yet)

>Audit-Trail:
From: Thomas Klausner <wiz@NetBSD.org>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: lib/59101: cpuset.c bugs
Date: Mon, 24 Feb 2025 23:10:51 +0100

 Not sure if it belongs in this bug report, but I think the API could
 do with a cpuset_count() or similar that returns how many CPUs there
 even are, so you can iterate over all of them easily. (No,
 cpuset_size() returns something different.)
  Thomas

Home	PR Database Search