Date: Thu, 9 Jul 1998 08:34:40 -0700 (PDT) From: gallatin@cs.duke.edu To: freebsd-gnats-submit@FreeBSD.ORG Subject: i386/7227: pmap_setdevram() disabled in wrong place Message-ID: <199807091534.IAA29258@hub.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 7227
>Category: i386
>Synopsis: pmap_setdevram() disabled in wrong place
>Confidential: no
>Severity: non-critical
>Priority: medium
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Thu Jul 9 08:40:00 PDT 1998
>Last-Modified:
>Originator: Andrew Gallatin
>Organization:
Duke University, Department of Computer Science
>Release: 3.0-CURRENT
>Environment:
FreeBSD grasshopper.cs.duke.edu 3.0-CURRENT FreeBSD 3.0-CURRENT #4: Wed Jul 8 19:58:49 EDT 1998 gallatin@grasshopper.cs.duke.edu:/usr/src/sys/compile/GRASSHOPPER i386
>Description:
The dangerous call to pmap_setdevram() which enables write combining
on video cards in P6/PII machines is already disabled in vga_probe().
Others may want to enable write combining on other devices (like Myrinet
cards) -- the immediate return in pmap_setdevram() should go away.
(broken out of i386/7201)
>How-To-Repeat:
Try enabling WC on a memory mapped PCI device on a PII ;-)
>Fix:
RCS file: /scratch/freebsd-cvs/src/sys/i386/i386/pmap.c,v
retrieving revision 1.202
diff -c -r1.202 pmap.c
*** pmap.c 1998/05/21 07:47:34 1.202
--- pmap.c 1998/07/09 15:27:37
***************
*** 505,511 ****
unsigned long long base;
unsigned long long mask;
- return;
if (cpu != CPU_686)
return;
--- 505,510 ----
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199807091534.IAA29258>