Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262899AbUCKARt (ORCPT ); Wed, 10 Mar 2004 19:17:49 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262912AbUCKARs (ORCPT ); Wed, 10 Mar 2004 19:17:48 -0500 Received: from smtp02.web.de ([217.72.192.151]:7174 "EHLO smtp.web.de") by vger.kernel.org with ESMTP id S262899AbUCKAQv (ORCPT ); Wed, 10 Mar 2004 19:16:51 -0500 From: Thomas Schlichter To: "Prakash K. Cheemplavam" Subject: Re: ACPI PM Timer vs. C1 halt issue Date: Thu, 11 Mar 2004 01:15:34 +0100 User-Agent: KMail/1.5.4 Cc: Len Brown , john stultz , lkml , Andrew Morton , Linus Torvalds , Dominik Brodowski References: <404E38B7.5080008@gmx.de> <1078956711.2557.72.camel@dhcppc4> <404F9B5B.6010207@gmx.de> In-Reply-To: <404F9B5B.6010207@gmx.de> MIME-Version: 1.0 Content-Type: Multipart/Mixed; boundary="Boundary-00=_o+6TAqhw4WLGQPo" Message-Id: <200403110115.36257.thomas.schlichter@web.de> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2818 Lines: 85 --Boundary-00=_o+6TAqhw4WLGQPo Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Hi! As your problem may be similar to mine, (http://marc.theaimsgroup.com/?l=3Dlinux-kernel&m=3D107840458123059&w=3D2) you may try to boot with ACPI PM timer enabled but with the additional boot= =20 option 'noapic'. If this also cools down your processor, you maybe should t= ry=20 the attached patch.... Thomas Am Mittwoch, 10. M=E4rz 2004 23:48 schrieb Prakash K. Cheemplavam: > [snip] > > > That said, I don't know what caused the regression you describe. > > Perhaps you can clarfiy the minimal changes necessary to switch between > > correct and incorrect behaviour? > > Simply use tsc as timesource instead of PM Timer (ie not compiling in > the latter) and all is cool again (now idle temp is as low as 40=B0C!). > ACPI as such is included. I attached my current config. I guess one > could also use a kernel boot parameter to switch between those timers. > > I don't know whether it is actual a problem with C1 or whatever. I just > described my assumption I made. If I disable CPU disconnect, my CPU runs > hot again even with tsc timer. With ACPI PM timer disconnect makes no > difference (hot everytime...). > > Cheers, > > Prakash --Boundary-00=_o+6TAqhw4WLGQPo Content-Type: text/x-diff; charset="iso-8859-1"; name="fix-8259-timer-ack.diff" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="fix-8259-timer-ack.diff" --- linux-2.6.4-rc2-mm1/arch/i386/kernel/io_apic.c.orig 2004-03-08 16:29:14.000000000 +0100 +++ linux-2.6.4-rc2-mm1/arch/i386/kernel/io_apic.c 2004-03-08 17:26:40.000000000 +0100 @@ -2181,7 +2181,7 @@ static inline void check_timer(void) */ apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); init_8259A(1); - if (nmi_watchdog == NMI_IO_APIC && !APIC_INTEGRATED(ver)) + if (nmi_watchdog == NMI_IO_APIC || APIC_INTEGRATED(ver)) timer_ack = 1; else timer_ack = !cpu_has_tsc; @@ -2202,7 +2202,7 @@ static inline void check_timer(void) disable_8259A_irq(0); setup_nmi(); enable_8259A_irq(0); - if (check_nmi_watchdog() < 0); + if (check_nmi_watchdog() < 0) timer_ack = !cpu_has_tsc; } return; @@ -2226,7 +2226,7 @@ static inline void check_timer(void) add_pin_to_irq(0, 0, pin2); if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); - if (check_nmi_watchdog() < 0); + if (check_nmi_watchdog() < 0) timer_ack = !cpu_has_tsc; } return; --Boundary-00=_o+6TAqhw4WLGQPo-- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/