linux-stable/arch/x86/pci
Matt Fleming 65694c5aad x86/PCI: Map PCI setup data with ioremap() so it can be in highmem
f9a37be0f0 ("x86: Use PCI setup data") added support for using PCI ROM
images from setup_data.  This used phys_to_virt(), which is not valid for
highmem addresses, and can cause a crash when booting a 32-bit kernel via
the EFI boot stub.

pcibios_add_device() assumes that the physical addresses stored in
setup_data are accessible via the direct kernel mapping, and that calling
phys_to_virt() is valid.  This isn't guaranteed to be true on x86 where the
direct mapping range is much smaller than on x86-64.

Calling phys_to_virt() on a highmem address results in the following:

 BUG: unable to handle kernel paging request at 39a3c198
 IP: [<c262be0f>] pcibios_add_device+0x2f/0x90
 ...
 Call Trace:
  [<c2370c73>] pci_device_add+0xe3/0x130
  [<c274640b>] pci_scan_single_device+0x8b/0xb0
  [<c2370d08>] pci_scan_slot+0x48/0x100
  [<c2371904>] pci_scan_child_bus+0x24/0xc0
  [<c262a7b0>] pci_acpi_scan_root+0x2c0/0x490
  [<c23b7203>] acpi_pci_root_add+0x312/0x42f
  ...

The solution is to use ioremap() instead of phys_to_virt() to map the
setup data into the kernel address space.

[bhelgaas: changelog]
Tested-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Matthew Garrett <mjg59@srcf.ucam.org>
Cc: Seth Forshee <seth.forshee@canonical.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: stable@vger.kernel.org	# v3.8+
2013-06-05 10:50:04 -06:00
..
acpi.c PCI changes for the v3.9 merge window: 2013-02-25 21:18:18 -08:00
amd_bus.c x86/PCI: put busn resource in pci_root_info for native host bridge drivers 2012-06-13 15:42:24 -06:00
broadcom_bus.c x86/PCI: dynamically allocate pci_root_info for native host bridge drivers 2012-04-30 14:52:43 -06:00
bus_numa.c X86: drivers: remove __dev* attributes. 2013-01-03 15:57:04 -08:00
bus_numa.h x86/PCI: put busn resource in pci_root_info for native host bridge drivers 2012-06-13 15:42:24 -06:00
ce4100.c x86/ce4100: Fix PCI configuration register access for devices without interrupts 2012-10-30 10:16:47 +01:00
common.c x86/PCI: Map PCI setup data with ioremap() so it can be in highmem 2013-06-05 10:50:04 -06:00
direct.c x86: constify PCI raw ops structures 2011-10-14 09:05:28 -07:00
early.c x86/PCI: remove early PCI pr_debug statements 2009-11-24 15:25:19 -08:00
fixup.c X86: drivers: remove __dev* attributes. 2013-01-03 15:57:04 -08:00
i386.c x86/PCI: Implement pcibios_resource_survey_bus() 2013-01-07 15:58:48 -07:00
init.c x86, olpc: Use pci subarch init for OLPC 2010-02-25 19:26:23 -08:00
irq.c x86/PCI: irq and pci_ids patch for Intel Panther Point DeviceIDs 2011-05-10 15:43:36 -07:00
legacy.c PCI changes for the v3.9 merge window: 2013-02-25 21:18:18 -08:00
Makefile x86/PCI: Add NumaChip remote PCI support 2012-12-07 14:24:32 -07:00
mmconfig-shared.c x86, uv, uv3: Update ACPI Check to include SGI UV3 2013-02-11 17:17:44 -08:00
mmconfig_32.c X86: drivers: remove __dev* attributes. 2013-01-03 15:57:04 -08:00
mmconfig_64.c X86: drivers: remove __dev* attributes. 2013-01-03 15:57:04 -08:00
mrst.c Revert "x86/pci/mrst: Use configuration mechanism 1 for 00:00.0, 00:02.0, 00:03.0" 2013-05-20 10:20:21 -06:00
numachip.c x86/PCI: Add NumaChip remote PCI support 2012-12-07 14:24:32 -07:00
numaq_32.c PCI changes for the v3.9 merge window: 2013-02-25 21:18:18 -08:00
olpc.c x86: constify PCI raw ops structures 2011-10-14 09:05:28 -07:00
pcbios.c X86: drivers: remove __dev* attributes. 2013-01-03 15:57:04 -08:00
sta2x11-fixup.c x86, platform: Initial support for sta2x11 I/O hub 2012-04-12 11:10:30 -07:00
visws.c PCI: Provide a default pcibios_update_irq() 2012-09-18 17:28:21 -06:00
xen.c xen/pci: Used cached MSI-X capability offset 2013-05-10 09:10:38 -04:00