Date: Wed, 5 Aug 1998 16:47:47 +0930 From: Greg Lehey <grog@lemis.com> To: Brian Somers <brian@Awfulhak.org>, Mike Smith <mike@smith.net.au> Cc: Randy Philipp <cphilipp@pop500.gsfc.nasa.gov>, Doug White <dwhite@resnet.uoregon.edu>, freebsd-mobile@FreeBSD.ORG, Jeroen van den Boom <jeroen@ctxnl.com> Subject: Re: Neomagic chipset - is it a hardware problem ? Message-ID: <19980805164747.I6348@freebie.lemis.com> In-Reply-To: <199808050702.IAA12582@awfulhak.org>; from Brian Somers on Wed, Aug 05, 1998 at 08:02:33AM %2B0100 References: <199808041424.HAA00456@antipodes.cdrom.com> <199808050702.IAA12582@awfulhak.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday, 5 August 1998 at 8:02:33 +0100, Brian Somers wrote: > [Jeroen van den Boom from CTX cc'd - this may interest you!] > [Greg cc'd - there's a question/request near the end... :-)] >>> I've found out more about this flickering problem.... it only occurs >>> when the laptop is physically warm - the warmer, the more flickering. >>> If the machine sleeps for ~15 minutes, it'll wake up with a perfect >>> picture and that picture will start to deteriorate with the building >>> heat. If I switch the display off for 15 minutes it makes no >>> difference. >>> >>> This is all under load. If the machine is idle it lasts a lot >>> longer (the fan is more effective I guess....) >>> >>> Has anyone else seen this problem ? I'm pretty much convinced that >>> it's a hardware problem now, so I'll probably return the machine >>> (second time) nearer the end of the week. >> >> I've seen similar symptoms on a Toshiba using the C&T 65555. The part >> would misbehave so badly that the X server would crash. >> >> Windows worked OK on it though; my hypothesis was just that the X >> driver wasn't doing whatever power management stuff the Windows driver >> was, and so the chip was overheating. I could be way off on that >> though; I know nothing about the C&T architecture. > > It looks like you're at least partially correct Mike. I've done some > more tests with heating up, cooling down, restarting X, restarting > FreeBSD, suspend & resume and LCD off & on.... > > It seems that the screen corruption is due to a combination of the X > server initialisation (*not* doing something it should) and the heat > of the machine. If I start X and let the machine heat up, or start > X when the machine is already hot, I get the flickering. > > Once it's flickering, the display can be made crystal clear again by > either suspending 'till the machine cools down (as I said before) or > by simply switching the LCD off and on (there's a small switch for > detecting when the lid is down that switches the LCD off by default > and can be configured to make the laptop ``sleep''). If I ``sleep > 'till cool'', the problem re-occurs as the machine heats up again. OK. We've established that the problem relates to heat, at least in part to heat generated by the system/display. That doesn't happen with Microsoft, if I understand correctly. That makes it difficult to escape the conclusion that something in your setup is not working correctly. > I can only guess that the sleep/resume is being caught by the power > management stuff and X is re-initialising (wrongly) the display > whereas the LCD off/on event is completely transparent and the laptop > firmware knows how to make the screen work when hot..... once this > `tweak' has happened, the display is fine forever more - up until it > goes into text mode and back into graphics mode (ala X server). I'm not convinced that this is the problem. Presumably, under Microsoft you can run the thing for hours (or until the "system" crashes) without this problem occurring. > This is *very* peculiar, but I think it's now looking more like a > software thing. I'll continue to test things.... I'm now running > KDE 1.0 (window manager) and am using a screensaver - maybe this is > relevant. I *do* wish NeoMagic would just publish the damn specs :-( Well, I'd say it's an interaction between hardware and software. > The other ``weird'' thing is that nobody else seems to have reported > this problem, although I guess there aren't that many that have gone > to the trouble of manually merging the initialisation code, and even > fewer (if any) that have a CTX Cybernote. Greg, as you've got a > similar setup (although I doubt your machine is a CTX Cybernote), can > you try doing a ``make world'' and see if you get similar behaviour, > making sure that you *haven't* let the laptops LCD initialise itself > (X should be the last thing to have set up the display from text > mode) ? It would be much appreciated. Well, I don't have a NeoMagic chipset. I've just been passing this information around for people who do. I'll gladly do a make world (if this damn server would give me access), but I don't think it would do much. What I *do* think could be useful would be to extract the video registers under Microsoft and compare them to what gets written in there by X, if you have a Microsoft partition left on the laptop. If you don't have anything suitable, I can check what I have here. Greg -- See complete headers for address and phone numbers finger grog@lemis.com for PGP public key To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980805164747.I6348>