Date: Fri, 4 Aug 2006 11:15:08 -0400 From: John Baldwin <jhb@freebsd.org> To: freebsd-mobile@freebsd.org, yraffah@savola.com Subject: Re: Interrupt Storms on irq:11 with Tecra A4 Message-ID: <200608041115.09320.jhb@freebsd.org> In-Reply-To: <44D24E6C.9050502@savola.com> References: <1152019643.704.42.camel@redevil.savola.com> <200608031114.51389.jhb@freebsd.org> <44D24E6C.9050502@savola.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 03 August 2006 15:28, Yousef Raffah wrote:
> John Baldwin wrote:
> > On Thursday 03 August 2006 03:47, Yousef Raffah wrote:
> >
> >> John Baldwin wrote:
> >>
> >>> On Wednesday 02 August 2006 01:56, Yousef Raffah wrote:
> >>>
> >>> Well, these two appear to be your problem due to your BIOS having
> >>> buggy ASL. You can work around it, but you'll need to figure out
> >>> the names of your link devices first. 'devinfo -v' with ACPI
> >>> enabled can help with that. For example, on my laptop:
> >>>
> >>> % devinfo -v | grep 'pci_link[67]'
> >>> pci_link6 pnpinfo _HID=PNP0C0F _UID=106 at handle=\_SB_.C002.C0F2
> >>> pci_link7 pnpinfo _HID=PNP0C0F _UID=107 at handle=\_SB_.C002.C0F3
> >>>
> >>>
> >>>
> >> Here is my devinfo
> >> devinfo -v | grep 'pci_link[67]'
> >> pci_link6 pnpinfo _HID=PNP0C0F _UID=7 at handle=\_SB_.PCI0.LPCB.LNKG
> >> pci_link7 pnpinfo _HID=PNP0C0F _UID=8 at handle=\_SB_.PCI0.LPCB.LNKH
> >>
> >> So I can see that pci_link6 is named LNKG and pci_link7 is LNKH
> >>
> >
> > Yep.
> >
> >
> >>> pci_link6 is called C0F2, and pci_link7 is called C0F3. You'll use
> >>> these names to override the IRQ for the link device by setting the
> >>> following variables in the loader:
> >>>
> >>> hw.pci.link.C0F2.irq=11
> >>> hw.pci.link.C0F3.irq=11
> >>>
> >>> (You'll have to replace C0F2 and C0F3 with the names for your link
> >>> devices. Probably LNKF and LNKG.)
> >>>
> >> Based on that, I should have the following in my /boot/loader.conf?
> >> hw.pci.link.LNKG.irq=11
> >> hw.pci.link.LNKH.irq=11
> >>
> >
> > Yep.
> >
> >
> >> Let us try the loader file first and see if we need the patch after that.
> >>
> >
> > Ok. I'm betting you will need the patch, but just want to make sure
first.
> >
> >
> Looks like I do need the patch as the interrupt storm is still there.
> Now how do I do that?
> What should I do with this code? I mean where to put it? is there a
> special way of
> patch file < your_patch ?
Just:
cd /usr/src/sys
patch < /path/to/patch_file
and then build a new kernel.
> Index: dev/acpica/acpi_pci_link.c
> ===================================================================
> RCS file: /usr/cvs/src/sys/dev/acpica/acpi_pci_link.c,v
> retrieving revision 1.53
> diff -u -r1.53 acpi_pci_link.c
> --- dev/acpica/acpi_pci_link.c 6 Jan 2006 16:14:32 -0000 1.53
> +++ dev/acpica/acpi_pci_link.c 2 Aug 2006 14:27:09 -0000
> @@ -941,18 +941,16 @@
> KASSERT(!PCI_INTERRUPT_VALID(link->l_irq),
> ("%s: link already has an IRQ", __func__));
>
> - /* Check for a tunable override and use it if it is valid. */
> + /* Check for a tunable override. */
> if (ACPI_SUCCESS(acpi_short_name(acpi_get_handle(dev), link_name,
> sizeof(link_name)))) {
> snprintf(tunable_buffer, sizeof(tunable_buffer),
> "hw.pci.link.%s.%d.irq", link_name, link->l_res_index);
> - if (getenv_int(tunable_buffer, &i) &&
> - PCI_INTERRUPT_VALID(i) && link_valid_irq(link, i))
> + if (getenv_int(tunable_buffer, &i) && PCI_INTERRUPT_VALID(i))
> return (i);
> snprintf(tunable_buffer, sizeof(tunable_buffer),
> "hw.pci.link.%s.irq", link_name);
> - if (getenv_int(tunable_buffer, &i) &&
> - PCI_INTERRUPT_VALID(i) && link_valid_irq(link, i))
> + if (getenv_int(tunable_buffer, &i) && PCI_INTERRUPT_VALID(i))
> return (i);
> }
>
>
>
>
>
>
> --
>
> Sincerely,
>
> Yousef Adnan Raffah
> Security Officer
> The Savola Group
>
> -------------------
> Aren't you using Firefox? Get it at http://www.getfirefox.com
>
>
>
--
John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200608041115.09320.jhb>