linux-toradex.git/arch/x86/platform, branch v4.4.7

Merge branch 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2015-11-04T05:33:18+00:00

Pull x86 platform changes from Ingo Molnar:
 "Misc updates to the Intel MID and SGI UV platforms"

* 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/intel-mid: Make intel_mid_ops static
  arch/x86/intel-mid: Use kmemdup rather than duplicating its implementation
  x86/platform/uv: Implement simple dump failover if kdump fails
  x86/platform/uv: Insert per_cpu accessor function on uv_hub_nmi

efi: Use correct type for struct efi_memory_map::phys_map

2015-10-28T11:28:06+00:00

We have been getting away with using a void* for the physical
address of the UEFI memory map, since, even on 32-bit platforms
with 64-bit physical addresses, no truncation takes place if the
memory map has been allocated by the firmware (which only uses
1:1 virtually addressable memory), which is usually the case.

However, commit:

  0f96a99dab36 ("efi: Add "efi_fake_mem" boot option")

adds code that clones and modifies the UEFI memory map, and the
clone may live above 4 GB on 32-bit platforms.

This means our use of void* for struct efi_memory_map::phys_map has
graduated from 'incorrect but working' to 'incorrect and
broken', and we need to fix it.

So redefine struct efi_memory_map::phys_map as phys_addr_t, and
get rid of a bunch of casts that are now unneeded.

Signed-off-by: Ard Biesheuvel 
Reviewed-by: Matt Fleming 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: izumi.taku@jp.fujitsu.com
Cc: kamezawa.hiroyu@jp.fujitsu.com
Cc: linux-efi@vger.kernel.org
Cc: matt.fleming@intel.com
Link: http://lkml.kernel.org/r/1445593697-1342-1-git-send-email-ard.biesheuvel@linaro.org
Signed-off-by: Ingo Molnar

Merge tag 'efi-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi into core/efi

2015-10-14T14:51:34+00:00

Pull v4.4 EFI updates from Matt Fleming:

  - Make the EFI System Resource Table (ESRT) driver explicitly
    non-modular by ripping out the module_* code since Kconfig doesn't
    allow it to be built as a module anyway. (Paul Gortmaker)

  - Make the x86 efi=debug kernel parameter, which enables EFI debug
    code and output, generic and usable by arm64. (Leif Lindholm)

  - Add support to the x86 EFI boot stub for 64-bit Graphics Output
    Protocol frame buffer addresses. (Matt Fleming)

  - Detect when the UEFI v2.5 EFI_PROPERTIES_TABLE feature is enabled
    in the firmware and set an efi.flags bit so the kernel knows when
    it can apply more strict runtime mapping attributes - Ard Biesheuvel

  - Auto-load the efi-pstore module on EFI systems, just like we
    currently do for the efivars module. (Ben Hutchings)

  - Add "efi_fake_mem" kernel parameter which allows the system's EFI
    memory map to be updated with additional attributes for specific
    memory ranges. This is useful for testing the kernel code that handles
    the EFI_MEMORY_MORE_RELIABLE memmap bit even if your firmware
    doesn't include support. (Taku Izumi)

Note: there is a semantic conflict between the following two commits:

  8a53554e12e9 ("x86/efi: Fix multiple GOP device support")
  ae2ee627dc87 ("efifb: Add support for 64-bit frame buffer addresses")

I fixed up the interaction in the merge commit, changing the type of
current_fb_base from u32 to u64.

Signed-off-by: Ingo Molnar

Merge branch 'x86/urgent' into core/efi, to pick up a pending EFI fix

2015-10-14T14:05:18+00:00

Signed-off-by: Ingo Molnar

x86/efi: Rename print_efi_memmap() to efi_print_memmap()

2015-10-12T13:20:08+00:00

This patch renames print_efi_memmap() to efi_print_memmap() and
make it global function so that we can invoke it outside of
arch/x86/platform/efi/efi.c

Signed-off-by: Taku Izumi 
Cc: Tony Luck 
Cc: Xishi Qiu 
Cc: Kamezawa Hiroyuki 
Cc: Ard Biesheuvel 
Signed-off-by: Matt Fleming

efi/x86: Move efi=debug option parsing to core

2015-10-12T13:20:05+00:00

fed6cefe3b6e ("x86/efi: Add a "debug" option to the efi= cmdline")
adds the DBG flag, but does so for x86 only. Move this early param
parsing to core code.

Signed-off-by: Leif Lindholm 
Acked-by: Ard Biesheuvel 
Tested-by: Ard Biesheuvel 
Cc: Mark Salter 
Cc: Borislav Petkov 
Signed-off-by: Matt Fleming

x86/intel-mid: Make intel_mid_ops static

2015-10-11T19:03:12+00:00

The following warning is issued on unfixed code.

arch/x86/platform/intel-mid/intel-mid.c:64:22: warning: symbol 'intel_mid_ops' was not declared. Should it be static?

Signed-off-by: Andy Shevchenko 
Link: http://lkml.kernel.org/r/1444400741-98669-1-git-send-email-andriy.shevchenko@linux.intel.com
Signed-off-by: Thomas Gleixner

x86/efi: Fix boot crash by mapping EFI memmap entries bottom-up at runtime, instead of top-down

2015-10-01T10:51:28+00:00

Beginning with UEFI v2.5 EFI_PROPERTIES_TABLE was introduced
that signals that the firmware PE/COFF loader supports splitting
code and data sections of PE/COFF images into separate EFI
memory map entries. This allows the kernel to map those regions
with strict memory protections, e.g. EFI_MEMORY_RO for code,
EFI_MEMORY_XP for data, etc.

Unfortunately, an unwritten requirement of this new feature is
that the regions need to be mapped with the same offsets
relative to each other as observed in the EFI memory map. If
this is not done crashes like this may occur,

  BUG: unable to handle kernel paging request at fffffffefe6086dd
  IP: [] 0xfffffffefe6086dd
  Call Trace:
   [] efi_call+0x7e/0x100
   [] ? virt_efi_set_variable+0x61/0x90
   [] efi_delete_dummy_variable+0x63/0x70
   [] efi_enter_virtual_mode+0x383/0x392
   [] start_kernel+0x38a/0x417
   [] x86_64_start_reservations+0x2a/0x2c
   [] x86_64_start_kernel+0xeb/0xef

Here 0xfffffffefe6086dd refers to an address the firmware
expects to be mapped but which the OS never claimed was mapped.
The issue is that included in these regions are relative
addresses to other regions which were emitted by the firmware
toolchain before the "splitting" of sections occurred at
runtime.

Needless to say, we don't satisfy this unwritten requirement on
x86_64 and instead map the EFI memory map entries in reverse
order. The above crash is almost certainly triggerable with any
kernel newer than v3.13 because that's when we rewrote the EFI
runtime region mapping code, in commit d2f7cbe7b26a ("x86/efi:
Runtime services virtual mapping"). For kernel versions before
v3.13 things may work by pure luck depending on the
fragmentation of the kernel virtual address space at the time we
map the EFI regions.

Instead of mapping the EFI memory map entries in reverse order,
where entry N has a higher virtual address than entry N+1, map
them in the same order as they appear in the EFI memory map to
preserve this relative offset between regions.

This patch has been kept as small as possible with the intention
that it should be applied aggressively to stable and
distribution kernels. It is very much a bugfix rather than
support for a new feature, since when EFI_PROPERTIES_TABLE is
enabled we must map things as outlined above to even boot - we
have no way of asking the firmware not to split the code/data
regions.

In fact, this patch doesn't even make use of the more strict
memory protections available in UEFI v2.5. That will come later.

Suggested-by: Ard Biesheuvel 
Reported-by: Ard Biesheuvel 
Signed-off-by: Matt Fleming 
Cc: 
Cc: Borislav Petkov 
Cc: Chun-Yi 
Cc: Dave Young 
Cc: H. Peter Anvin 
Cc: James Bottomley 
Cc: Lee, Chun-Yi 
Cc: Leif Lindholm 
Cc: Linus Torvalds 
Cc: Matthew Garrett 
Cc: Mike Galbraith 
Cc: Peter Jones 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: linux-kernel@vger.kernel.org
Link: http://lkml.kernel.org/r/1443218539-7610-2-git-send-email-matt@codeblueprint.co.uk
Signed-off-by: Ingo Molnar

arch/x86/intel-mid: Use kmemdup rather than duplicating its implementation

2015-09-17T21:38:59+00:00

The patch was generated using fixed coccinelle semantic patch
scripts/coccinelle/api/memdup.cocci [1].

[1]: http://permalink.gmane.org/gmane.linux.kernel/2014320

Signed-off-by: Andrzej Hajda 
Cc: Bartlomiej Zolnierkiewicz 
Cc: Marek Szyprowski 
Cc: Julia Lawall 
Link: http://lkml.kernel.org/r/1438934377-4922-9-git-send-email-a.hajda@samsung.com
Signed-off-by: Thomas Gleixner

x86/platform/uv: Implement simple dump failover if kdump fails

2015-09-14T13:31:30+00:00

The ability to trigger a kdump using the system NMI command
was added by

    commit 12ba6c990fab ("x86/UV: Add kdump to UV NMI handler")
    Author: Mike Travis 
    Date:   Mon Sep 23 16:25:03 2013 -0500

This is useful because when kdump is working the information
gathered is more informative than the original per CPU stack
traces or "dump" option.  However a number of things can go
wrong with kdump and then the stack traces are more useful than
nothing.

The two most common reasons for kdump to not be available are:

  1) if a problem occurs during boot before the kdump service is
     started, or
  2) the kdump daemon failed to start.

In either case the call to crash_kexec() returns unexpectedly.

When this happens uv_nmi_kdump() also sets the
uv_nmi_kexec_failed flag which causes the slave CPU's to also
return to the NMI handler. Upon this unexpected return to the
NMI handler, the NMI handler will revert to the "dump" action
which uses show_regs() to obtain a process trace dump for all
the CPU's.

Other minor changes:
    The "dump" action now generates both the show_regs() stack trace
    and show instruction pointer information.  Whereas the "ips"
    action only shows instruction pointers for non-idle CPU's.  This
    is more like an abbreviated "ps" display.

    Change printk(KERN_DEFAULT...) --> pr_info()

Signed-off-by: Mike Travis 
Signed-off-by: George Beshers 
Cc: Alex Thorlton 
Cc: Dimitri Sivanich 
Cc: Hedi Berriche 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Russ Anderson 
Cc: Thomas Gleixner 
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar