Date: Tue, 18 Apr 2000 13:33:36 -0600 From: Warner Losh <imp@village.org> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: Alan Clegg <abc@firehouse.net>, freebsd-current@FreeBSD.ORG Subject: Re: Patch to allow TSC with APM Message-ID: <200004181933.NAA93475@harmony.village.org> In-Reply-To: Your message of "Tue, 18 Apr 2000 20:11:33 %2B0200." <17397.956081493@critter.freebsd.dk> References: <17397.956081493@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <17397.956081493@critter.freebsd.dk> Poul-Henning Kamp writes: : If it is because the bios fiddles the frequency all the time we : should maybe understand the APM bios better. There were issues with some APM BIOSes not resetting this on resume, but I think that we properly store/restore the state now. : If it is because SMI interrupts "steal" time from us, then the : TSC is maybe better. This is an excellent theory since the problems are on newer laptops that use SMI a whole lot. But 6 seconds a minute is way too large to account for this unless the SMI interrupts are timed really really poorly. On a high precision system we have here that needs to run off the i8245, we've found that interrupts at the wrong time can cause "dropouts" of 1 click of time. Since these dropouts happen only once in 1000 samples and are eliminated by the outlier elimination mechanisms that we have in the software, I've not investigated this further. I guess what I'm saying is that this might be possible, but I don't know for sure if it is or not. Warner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200004181933.NAA93475>