NetBSD Problem Report #58171

From www@netbsd.org  Wed Apr 17 22:08:18 2024
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))
	(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 2EE5A1A9238
	for <gnats-bugs@gnats.NetBSD.org>; Wed, 17 Apr 2024 22:08:18 +0000 (UTC)
Message-Id: <20240417220816.D10631A923A@mollari.NetBSD.org>
Date: Wed, 17 Apr 2024 22:08:16 +0000 (UTC)
From: campbell+netbsd@mumble.net
Reply-To: campbell+netbsd@mumble.net
To: gnats-bugs@NetBSD.org
Subject: pckbport is not MP-safe
X-Send-Pr-Version: www-1.0

>Number:         58171
>Category:       kern
>Synopsis:       pckbport is not MP-safe
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Apr 17 22:10:00 +0000 2024
>Last-Modified:  Thu Apr 18 12:00:02 +0000 2024
>Originator:     Taylor R Campbell
>Release:        current, 10, 9, 8
>Organization:
The NetBS/2 Foundation
>Environment:
>Description:
pckbport and all associated drivers -- pckbd(4), pms(4), the sub-drivers for alps/elantech/synaptics -- are not MP-safe: they run with the kernel lock.

This is bad for interactivity when _any other_ legacy kernel-locked subsystem hogs _any_ CPU.

(Note: wscons/wskbd/wsmouse/wsevent also have the same problem, but let's use a separate PR to track those.)
>How-To-Repeat:
find something that hogs the kernel lock for longer than a few milliseconds and try to type or mouse around while using it
>Fix:
Yes, please!

>Audit-Trail:
From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/58171: pckbport is not MP-safe
Date: Thu, 18 Apr 2024 11:56:33 -0000 (UTC)

 campbell+netbsd@mumble.net writes:

 >pckbport and all associated drivers -- pckbd(4), pms(4), the sub-drivers for alps/elantech/synaptics -- are not MP-safe: they run with the kernel lock.

 ISA interrupts are not MP-safe.

 pckbd might benefit from the simplicity of KERNEL_LOCK as it should
 work for DDB.

Home	PR Database Search