Date: Fri, 22 Feb 2008 17:53:03 +0000 From: Rui Paulo <rpaulo@fnop.net> To: Andriy Gapon <avg@icyb.net.ua> Cc: freebsd-hackers@freebsd.org, freebsd-acpi@freebsd.org Subject: Re: cx_lowest and CPU usage Message-ID: <77122846-92A6-47D2-B94C-3368AF167963@fnop.net> In-Reply-To: <47BEF2AA.900@icyb.net.ua> References: <479F0ED4.9030709@icyb.net.ua> <479F62D9.6080703@root.org> <47A33CCB.3090902@icyb.net.ua> <47B0C10F.6000109@icyb.net.ua> <47B4103A.6090902@icyb.net.ua> <uh7bq6jcveu.fsf@P142.sics.se> <47B4A103.7040801@icyb.net.ua> <47B4B31A.4020605@icyb.net.ua> <47B84E61.3060401@icyb.net.ua> <47BB375C.5010208@icyb.net.ua> <47BB4D5C.9000406@icyb.net.ua> <47BC7287.6000301@icyb.net.ua> <47BEF2AA.900@icyb.net.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
I'll handle this. Thanks!
On Feb 22, 2008, at 4:04 PM, Andriy Gapon wrote:
> on 20/02/2008 20:33 Andriy Gapon said the following:
>> on 19/02/2008 23:42 Andriy Gapon said the following:
>>> The last result most probably means that RTC IRQ was not the =20
>>> interrupt
>>> to wake CPU from sleeping state.
>>> The first possibility that comes to mind is that on this particular
>>> hardware RTC interrupt (IRQ8) is not able to wake the system from =20=
>>> C2 state.
>>
>> So it seems that this was true.
>> Here's a shortcut to the relevant info:
>> PIIX4E (FW82371EB) specification
>> DEVACTB =97 DEVICE ACTIVITY B (FUNCTION 3) pci register description
>> BRLD_EN_IRQ8, bit 5
>>
>> $ pciconf -r pci0:0:7:3 0x58
>> 03040c07
>
> Attached is a patch that fixes the issue for me (without any
> side-effects) and should not cause any harm for others.
>
> --=20
> Andriy Gapon
> --- acpi_cpu.c.orig 2008-02-21 21:08:16.000000000 +0200
> +++ acpi_cpu.c 2008-02-21 21:13:54.000000000 +0200
> @@ -113,6 +113,12 @@
> #define PCI_REVISION_B_STEP 1
> #define PCI_REVISION_4E 2
> #define PCI_REVISION_4M 3
> +#define PIIX4_DEVACTB_REG 0x58
> +#define PIIX4_BRLD_EN_IRQ0 (1<<0)
> +#define PIIX4_BRLD_EN_IRQ (1<<1)
> +#define PIIX4_BRLD_EN_IRQ8 (1<<5)
> +#define PIIX4_STOP_BREAK_MASK (PIIX4_BRLD_EN_IRQ0 | =20
> PIIX4_BRLD_EN_IRQ | PIIX4_BRLD_EN_IRQ8)
> +#define PIIX4_PCNTRL_BST_EN (1<<10)
>
> /* Platform hardware resource information. */
> static uint32_t cpu_smi_cmd; /* Value to write to =
SMI_CMD. */
> @@ -1004,6 +1010,7 @@
> acpi_cpu_quirks(void)
> {
> device_t acpi_dev;
> + uint32_t val;
>
> ACPI_FUNCTION_TRACE((char *)(uintptr_t)__func__);
>
> @@ -1052,12 +1059,25 @@
> * See erratum #18 ("C3 Power State/BMIDE and Type-F DMA
> * Livelock") from the January 2002 PIIX4 specification update.
> * Applies to all PIIX4 models.
> + *
> + * Also, make sure that all interrupts cause a "Stop Break"
> + * event to exit from C2 state.
> */
> + case PCI_REVISION_A_STEP:
> + case PCI_REVISION_B_STEP:
> case PCI_REVISION_4E:
> case PCI_REVISION_4M:
> cpu_quirks |=3D CPU_QUIRK_NO_C3;
> ACPI_DEBUG_PRINT((ACPI_DB_INFO,
> "acpi_cpu: working around PIIX4 bug, disabling C3\n"));
> +
> + val =3D pci_read_config(acpi_dev, PIIX4_DEVACTB_REG, 4);
> + if ((val & PIIX4_STOP_BREAK_MASK) !=3D =
PIIX4_STOP_BREAK_MASK) {
> + ACPI_DEBUG_PRINT((ACPI_DB_INFO,
> + "PIIX4: enabling IRQs to generate Stop Break\n"));
> + val |=3D PIIX4_STOP_BREAK_MASK;
> + pci_write_config(acpi_dev, PIIX4_DEVACTB_REG, val, 4);
> + }
> break;
> default:
> break;
> _______________________________________________
> freebsd-acpi@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-acpi
> To unsubscribe, send any mail to "freebsd-acpi-=20
> unsubscribe@freebsd.org"
--
Rui Paulo
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?77122846-92A6-47D2-B94C-3368AF167963>