Date: Thu, 19 Feb 2009 13:10:38 +0200 From: Andriy Gapon <avg@icyb.net.ua> To: Rui Paulo <rpaulo@FreeBSD.org> Cc: freebsd-acpi@FreeBSD.org Subject: Re: piix4 and C2: patch for BRLD_EN_BM/ACPI_BITREG_BUS_MASTER_RLD Message-ID: <499D3E2E.7090500@icyb.net.ua> In-Reply-To: <66885316-6F5B-4C56-BFE0-DBB86E050C99@freebsd.org> References: <498708FD.9030404@icyb.net.ua> <66885316-6F5B-4C56-BFE0-DBB86E050C99@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
on 02/02/2009 17:50 Rui Paulo said the following: > > On 2 Feb 2009, at 14:53, Andriy Gapon wrote: > >> >> Another patch for acpi_cpu code related to PIIX4 chipset. >> >> We do not support C3 with this chipset and as such we never touch >> ACPI_BITREG_BUS_MASTER_RLD when running on it. It's also possible >> (probable) that we never reset ACPI_BITREG_BUS_MASTER_STATUS (because of >> CPU_QUIRK_NO_BM_CTRL). >> So, if BIOS configures ACPI_BITREG_BUS_MASTER_RLD to 1 and >> ACPI_BITREG_BUS_MASTER_STATUS gets set to 1, we would get a situation >> where the system would always immediately break out of C2 state should >> we ever attempt to enter it. >> >> Actually, ACPI spec says that bus master activity need not break out of >> C2 state (only interrupts must do that). >> >> Given all of the above, it seems to be safe and desired to reset >> ACPI_BITREG_BUS_MASTER_RLD to zero with this chipset. >> >> The patch is successfully tested on a real system that exhibited the >> above behavior (immediate backout from cx_lowest of C2 to C1). > > > > Looks ok to me. I no longer have a PIIX4 system to test it, though. So, > if no one raises any objection, I'll commit it to HEAD in a few days. > > Andriy, if I forget, please ping me :-) Rui, may I commit this one now? :) -- Andriy Gapon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?499D3E2E.7090500>