[PATCH] efi: fdt: avoid FDT manipulation after ExitBootServices()

[Ard Biesheuvel]

On 1 February 2017 at 22:59, Jeffrey Hugo <jhugo at codeaurora.org> wrote:
> On 2/1/2017 2:08 PM, Jeffrey Hugo wrote:
>>
>> On 2/1/2017 10:45 AM, Ard Biesheuvel wrote:
>>>
>>> Some AArch64 UEFI implementations disable the MMU in ExitBootServices(),
>>> after which unaligned accesses to RAM are no longer supported.
>>>
>>> Commit abfb7b686a3e ("efi/libstub/arm*: Pass latest memory map to the
>>> kernel") fixed an issue in the memory map handling of the stub FDT code,
>>> but inadvertently created an issue with such firmwares, by moving some
>>> of the FDT manipulation to after the invocation of ExitBootServices().
>>> Given that the stub's libfdt implementation uses the ordinary,
>>> accelerated
>>> string functions, which rely on hardware handling of unaligned accesses,
>>> manipulating the FDT with the MMU off may result in alignment faults.
>>>
>>> So fix the situation by moving the update_fdt_memmap() call into the
>>> callback function invoked by efi_exit_boot_services() right before it
>>> calls the ExitBootServices() UEFI service (which is arguably a better
>>> place for it anyway)
>>>
>>> Note that disabling the MMU in ExitBootServices() is not compliant with
>>> the UEFI spec, and carries great risk due to the fact that switching from
>>> cached to uncached memory accesses halfway through compiler generated
>>> code
>>> (i.e., involving a stack) can never be done in a way that is
>>> architecturally
>>> safe.
>>>
>>> Cc: <stable at vger.kernel.org>
>>> Fixes: abfb7b686a3e ("efi/libstub/arm*: Pass latest memory map to the
>>> kernel")
>>> Tested-by: Riku Voipio <riku.voipio at linaro.org>
>>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel at linaro.org>
>>
>>
>> NACK, please.  This causes a regression on my platform, in the form of
>> an assert in UEFI once ExitBootServices() is called, per initial
>> testing.  I'll do more testing to determine why.
>>
>
> Sorry, false alarm.  The assert appears to have been the result of a bad
> tree and a bad target configuration, and was not reproduced on another
> setup, nor did the assert make sense in context with this change.
>

No worries. Thanks for taking the time to test the patch.

-- 
Ard.