Re: Continuing PCI and Yenta troubles in 2.6.13.1 and 2.6.14-rc1

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Sep 20, 2005 at 08:06:15PM +0200, Andreas Koch wrote:
> many thanks for the quick patch. I have applied it and your previous Yenta 
> patch (is that still necessary with the new patch?) to 2.6.13.2.

Yenta patch is still needed (it will be included in 2.6.13.3). 

> With pci=assign-busses, I don't see any obvious oops or slow-downs either.
> However: As you can see from the attached boot log (assign-busses-26132.log), 
> there are now problems with the USB ports 9-2 and 10-2, on  
> uhci_hcd 0000:05:02.1 and uhci_hcd 0000:05:03.0, respectively. This delays the 
> boot for about 5 seconds per error, but does continue. The system doesn't 
> hang afterwards, either (again, tested for roughly 15min). The lspci-vvx 
> output is attached as assign-busses-26132.lspci.

These errors most likely mean that there are no interrupts form PCI
devices in the docking station. I'd suppose that massively changed bus
numbers make ACPI confused wrt interrupt routing, but all relevant IRQs
are exactly the same as without "assign-busses", so I'm puzzled...

> I'll try using the patched 2.6.13.2 without pci=assign-busses in the next few 
> days and test whether other instabilities come up. If you want me to try 
> further patches for fixing the assign-busses option, I'm game.

At the moment I can only think of something like this: the PCIE-to-PCI bridge
(Intel 6702PXH) in your docking station also has IOxAPIC as PCI function #1,
but its config space is turned off by BIOS for some reason. However, its MMIO
registers are still present somewhere in PCI address space - we just don't
know where.
So there is a small possibility that we have accidentally stepped on them
and destroyed IRQ routing setup writing to MMIO registers of totally
unrelated device.

Please test this patch, though chances that it helps are quite small...
In either case lspci output will be useful.

Ivan.

--- 2.6.13.2/drivers/pci/quirks.c	Mon Aug 29 03:41:01 2005
+++ linux/drivers/pci/quirks.c	Wed Sep 21 16:58:01 2005
@@ -1298,6 +1298,12 @@ DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_IN
  */
 static void __devinit quirk_pcie_pxh(struct pci_dev *dev)
 {
+	u16 ctrl;
+
+	/* Enable config access to I/OxAPIC function. */
+	pci_read_config_word(dev, 0x40, &ctrl);
+	pci_write_config_word(dev, 0x40, ctrl | 0x2000);
+
 	disable_msi_mode(dev, pci_find_capability(dev, PCI_CAP_ID_MSI),
 					PCI_CAP_ID_MSI);
 	dev->no_msi = 1;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]
  Powered by Linux