On 3/2/2021 3:54 PM, Thomas Bogendoerfer wrote:
> On Mon, Mar 01, 2021 at 08:19:38PM -0800, Florian Fainelli wrote:
>> BMIPS is one of the few platforms that do change the exception base.
>> After commit 2dcb39645441 ("memblock: do not start bottom-up allocations
>> with kernel_end") we started seeing BMIPS boards fail to boot with the
>> built-in FDT being corrupted.
>>
>> Before the cited commit, early allocations would be in the [kernel_end,
>> RAM_END] range, but after commit they would be within [RAM_START +
>> PAGE_SIZE, RAM_END].
>>
>> The custom exception base handler that is installed by
>> bmips_ebase_setup() done for BMIPS5000 CPUs ends-up trampling on the
>> memory region allocated by unflatten_and_copy_device_tree() thus
>> corrupting the FDT used by the kernel.
>>
>> To fix this, we need to perform an early reservation of the custom
>> exception that is going to be installed and this needs to happen at
>> plat_mem_setup() time to ensure that unflatten_and_copy_device_tree()
>> finds a space that is suitable, away from reserved memory.
>>
>> Huge thanks to Serget for analysing and proposing a solution to this
>> issue.
>>
>> Fixes: Fixes: 2dcb39645441 ("memblock: do not start bottom-up allocations with kernel_end")
>> Debugged-by: Serge Semin <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx>
>> Reported-by: Kamal Dasu <kdasu.kdev@xxxxxxxxx>
>> Signed-off-by: Florian Fainelli <f.fainelli@xxxxxxxxx>
>> ---
>> Thomas,
>>
>> This is intended as a stop-gap solution for 5.12-rc1 and to be picked up
>> by the stable team for 5.11. We should find a safer way to avoid these
>> problems for 5.13 maybe.
>
> let's try to make it in one ago. Hwo about reserving vector space in
> cpu_probe, if it's known there and leave the rest to trap_init() ?
>
> Below patch got a quick test on IP22 (real hardware) and malta (qemu).
> Not sure, if I got all BMIPS parts correct, so please check/test.
Works for me here:
Tested-by: Florian Fainelli <f.fainelli@xxxxxxxxx>
Thanks!
> BTW. do we really need to EXPORT_SYMBOL ebase ?
It seems like MIPS KVM support can be built as a module which is why
ebase was exported to modules with
878edf014e29de38c49153aba20273fbc9ae31af ("MIPS: KVM: Restore host EBase
from ebase variable")?
--
Florian
