From owner-freebsd-stable Tue Mar 20 21:19: 4 2001 Delivered-To: freebsd-stable@freebsd.org Received: from laptop.os2warp.org (laptop.os2warp.org [209.136.194.228]) by hub.freebsd.org (Postfix) with ESMTP id 93A8F37B719 for ; Tue, 20 Mar 2001 21:19:00 -0800 (PST) (envelope-from lambert@laptop.os2warp.org) Received: by laptop.os2warp.org (Postfix, from userid 1000) id 5CC169C06; Tue, 20 Mar 2001 23:19:19 -0600 (CST) Date: Tue, 20 Mar 2001 23:19:19 -0600 From: Scott Lambert To: FreeBSD-STABLE@FreeBSD.org Subject: microuptime() went backwards Message-ID: <20010320231919.A23244@laptop.os2warp.org> Reply-To: FreeBSD-STABLE@FreeBSD.org Mail-Followup-To: Scott Lambert , FreeBSD-STABLE@FreeBSD.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG I have just encounterred the microuptime() went backwards issue on my dual PPro 200 box. I have not changed the operating system at all. What I have done is change the video card (an ATI Expert 98, PCI), and compiled XFree86-4.0.3 from ports. I had not used X on this box before. I then installed X, ran xf86cfg, and exitted xf86cfg. At this point I had 5 microuptime errors all during the same second. I then tried to run X but it did not work. I had one microuptime error. I ran xf86cfg again and created 6 more microuptime errors. Again all errors happened in the same second. I am rebuilding this history from the log files. I did not pay a lot of attention to the error messages at the time because I knew I had seen the errors mentioned on the mailling lists and assumed I could fix it later. Two minutes later I have another 5 messages in one second. I think I editted the XF86Config file by hand in those two minutes then popped back into xf86cfg to check some things. After that it gets fuzzy. It seems that they only happenned while playing with xf86cfg or xvidtune. Most of the errors in my log showed up during the time that I think I was playing with xvidtune. During those times there were many more than 5 microuptime errors per second. Or maybe it just happens while actively doing something in X. I have spent the rest of the day building GNOME from ports. It was still building when I left work. I did use X after the errors end but I was simply runinig top in an xterm or installing xscreensaver in an xterm. (xscreensaver didn't want to work, I'm waiting for GNOME to use the GUI config rather than read the manual tonight. It was time for dinner anyway.) Now, I've searched the mailing lists and found two recommendations that don't seem to apply to me, and thoughts that microuptime errors may be related to certain Abit mobo's or even AMD processors. I'm posting this for data that it is not necessarily limited to the newer AMD related hardware and that APM may not be directly related. APM has never been compiled into the kernel on this machine. The sysctl variable is set to what was reccomended: $ sysctl kern.timecounter.method kern.timecounter.method: 0 Could the video card be busmastering too long and not allowing the kernel to get timer readings on time? ie. a bus contention issue? Thanks for listening, -- Scott Lambert lambert@os2warp.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message