NetBSD Problem Report #54767

From elo@foobox.net  Sat Dec 14 21:41:29 2019
Return-Path: <elo@foobox.net>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.2 with cipher ECDHE-RSA-AES256-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 D166F7A174
	for <gnats-bugs@gnats.NetBSD.org>; Sat, 14 Dec 2019 21:41:29 +0000 (UTC)
Message-Id: <20191214214125.86DF1821F@marmite.localnet>
Date: Sat, 14 Dec 2019 21:41:25 +0000 (GMT)
From: elo@foobox.net
Reply-To: elo@foobox.net
To: gnats-bugs@NetBSD.org
Subject: Wrong mask constant in VESA BIOS Extensions code
X-Send-Pr-Version: 3.95

>Number:         54767
>Category:       kern
>Synopsis:       Wrong mask constant in VESA BIOS Extensions code
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Dec 14 21:45:00 +0000 2019
>Closed-Date:    Sun Dec 15 10:09:23 +0000 2019
>Last-Modified:  Sun Dec 15 10:09:23 +0000 2019
>Originator:     elo
>Release:        NetBSD 9.0_BETA
>Organization:
>Environment:
System: NetBSD marmite.localnet 9.0_BETA NetBSD 9.0_BETA (BLUEBELL) #3: Sun Nov 3 02:10:56 GMT 2019 elo@marmite.localnet:/usr/obj/sys/arch/amd64/compile/BLUEBELL amd64
Architecture: x86_64
Machine: amd64
>Description:
	In the file 'src/sys/arch/i386/stand/lib/vbe.c' is a mask
	constant with transposed digits. According to the 'VESA BIOS
	EXTENSION (VBE) Core Functions Standard, Version 3.0', dated
	16 September 1998, found at
	http://www.petesqbsite.com/sections/tutorials/tuts/vbe3.pdf,
	in the ModeAttributes field of the ModeInfoBlock structure for
	VBE Function 01h (see page 32), it is bit 7 that indicates
	whether linear frame-buffer mode is available, but because
	the mask is wrong (0x08 instead of 0x80), it is in fact the
	graphics/text-mode flag being tested.
>How-To-Repeat:
	The error was found by code inspection. To my knowledge, it
	has never caused a problem in practice.
>Fix:
	Apply the following patch.

--- a/usr/src/sys/arch/i386/stand/lib/vbe.c	2019-08-21 04:14:10.759601198 +0100
+++ b/usr/src/sys/arch/i386/stand/lib/vbe.c	2019-08-21 04:14:10.800315268 +0100
@@ -50,7 +50,7 @@
 {
 	if ((mi->ModeAttributes & 0x01) == 0)
 		return 0;	/* mode not supported by hardware */
-	if ((mi->ModeAttributes & 0x08) == 0)
+	if ((mi->ModeAttributes & 0x80) == 0)
 		return 0;	/* linear fb not available */
 	if ((mi->ModeAttributes & 0x10) == 0)
 		return 0;	/* text mode */

>Release-Note:

>Audit-Trail:
From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/54767: Wrong mask constant in VESA BIOS Extensions code
Date: Sun, 15 Dec 2019 01:12:14 -0000 (UTC)

 elo@foobox.net writes:

 >	in the ModeAttributes field of the ModeInfoBlock structure for
 >	VBE Function 01h (see page 32), it is bit 7 that indicates
 >	whether linear frame-buffer mode is available, but because
 >	the mask is wrong (0x08 instead of 0x80), it is in fact the
 >	graphics/text-mode flag being tested.

 0x08 is a test for color mode, ruling out monochrome modes. I would
 guess that there is no support for that, and we are actually missing
 two lines like:

 Index: vbe.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/i386/stand/lib/vbe.c,v
 retrieving revision 1.9
 diff -p -u -r1.9 vbe.c
 --- vbe.c       24 Jan 2017 11:09:14 -0000      1.9
 +++ vbe.c       15 Dec 2019 01:10:26 -0000
 @@ -51,6 +51,8 @@ vbe_mode_is_supported(struct modeinfoblo
         if ((mi->ModeAttributes & 0x01) == 0)
                 return 0;       /* mode not supported by hardware */
         if ((mi->ModeAttributes & 0x08) == 0)
 +               return 0;       /* only color modes are supported */
 +       if ((mi->ModeAttributes & 0x80) == 0)
                 return 0;       /* linear fb not available */
         if ((mi->ModeAttributes & 0x10) == 0)
                 return 0;       /* text mode */

 -- 
 -- 
                                 Michael van Elst
 Internet: mlelstv@serpens.de
                                 "A potential Snark may lurk in every tree."

From: "Christos Zoulas" <christos@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/54767 CVS commit: src/sys/arch/i386/stand/lib
Date: Sat, 14 Dec 2019 22:38:18 -0500

 Module Name:	src
 Committed By:	christos
 Date:		Sun Dec 15 03:38:17 UTC 2019

 Modified Files:
 	src/sys/arch/i386/stand/lib: vbe.c

 Log Message:
 PR/54767: elo: fix incorrect test (mlelstv)
 Add symbolic constants and reference to the standard.


 To generate a diff of this commit:
 cvs rdiff -u -r1.9 -r1.10 src/sys/arch/i386/stand/lib/vbe.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

State-Changed-From-To: open->closed
State-Changed-By: wiz@NetBSD.org
State-Changed-When: Sun, 15 Dec 2019 10:09:23 +0000
State-Changed-Why:
Fixed by mlelstv & christos. Thanks!


>Unformatted:

Home	PR Database Search