Date: Fri, 31 Aug 2001 21:15:33 +0200 From: Tor.Egge@fast.no To: jhb@FreeBSD.org Cc: freebsd-smp@FreeBSD.org, freebsd-hackers@FreeBSD.org, Patrick.Guelat@imp.ch, phk@FreeBSD.org, drussell@saturn-tech.com, roberto@eurocontrol.fr, atrn@zeta.org.au, sthaug@nethelp.no, kpielorz@tdx.co.uk, bkarp@icsi.berkeley.edu, bde@zeta.org.au, mb@imp.ch Subject: Re: Clock speedup on 4.X FreeBSD SMP and serverworks chipset Message-ID: <200108311915.VAA34938@midten.fast.no> In-Reply-To: Your message of "Fri, 31 Aug 2001 11:22:30 -0700 (PDT)" References: <XFMail.010831112230.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Hmm, does -current even need clkintr_pending anymore? What is its purpose? There is a problem if i8254_get_timecount locks clock_lock right before a rollover but reads the timer right afterwards. In that case, a too low value can be returned (resulting in 1/Hz too small time value) although the result will still be monotonous. For RELENG_4, which uses Xintr0, this problem is reduced by adding the clkintr_pending variable (cf. revision 1.134 of clock.c and revision 1.38 of apic_vector.s) Looking more at the code, I now see that clkintr_pending is never set in -current due to Xintr0 not being used (Xfastintr0 is now used). Using a fast interrupt for the clock interrupt gives some of the same reduction of the race window size as the use of clkintr_pending. - Tor Egge To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200108311915.VAA34938>