Re: [PATCH v5] ARM64: dts: meson-gx: Add firmware reserved memory zones

[Kevin Hilman]

Andreas FÃrber <afaerber@xxxxxxx> writes:

> Am 19.01.2017 um 01:20 schrieb Andreas FÃrber:
>> Hi,
>> 
>> Am 18.01.2017 um 17:50 schrieb Neil Armstrong:
>>> The Amlogic Meson GXBB/GXL/GXM secure monitor uses part of the memory space,
>>> this patch adds these reserved zones.
>>>
>>> Without such reserved memory zones, running the following stress command :
>>> $ stress-ng --vm 16 --vm-bytes 128M --timeout 10s
>>> multiple times:
>>>
>>> Could lead to the following kernel crashes :
>>> [   46.937975] Bad mode in Error handler detected on CPU1, code 0xbf000000 -- SError
>>> ...
>>> [   47.058536] Internal error: Attempting to execute userspace memory: 8600000f [#3] PREEMPT SMP
>>> ...
>>> Instead of the OOM killer.
>>>
>> 
>> I miss a Fixes: or Cc: here for the backport you desired. To have it
>> fixed back to my very introduction:
>> 
>> Fixes: 4f24eda8401f ("ARM64: dts: Prepare configs for Amlogic Meson GXBaby")
>> 
>> People backporting it would need to handle the meson-{gx => gxbb}.dtsi
>> transition for 4.9 down to 4.6, which seems fairly straightforward.
>> 
>>> Signed-off-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx>
>>> ---
>>>  arch/arm64/boot/dts/amlogic/meson-gx.dtsi | 18 ++++++++++++++++++
>>>  1 file changed, 18 insertions(+)
>>>
>>> Changes since v4 at [5]:
>>> - Move start of ddr memory to reserved-memory node
>>> - Drop memory node move
>>> - Fix typo in sizes
>>>
>>> Changes since resent v2 at [4]:
>>> - Fix invalid comment of useable memory attributes
>>>
>>> Changes since original v2 at [3]:
>>> - Typo in commit 2GiB -> 1GiB, 4GiB -> 2GiB
>>>
>>> Changes since v2 at [2]:
>>> - Moved all memory node out of dtsi
>>> - Added comment about useable memory
>>> - Fixed comment about secmon reserved zone
>>>
>>> Changes since v1 at [1] :
>>> - Renamed reg into linux,usable-memory to ovveride u-boot memory
>>> - only kept secmon memory zone
>>>
>>> [1] http://lkml.kernel.org/r/20161212101801.28491-1-narmstrong@xxxxxxxxxxxx
>>> [2] http://lkml.kernel.org/r/1483105232-6242-1-git-send-email-narmstrong@xxxxxxxxxxxx
>>> [3] http://lkml.kernel.org/r/1484128128-22454-1-git-send-email-narmstrong@xxxxxxxxxxxx
>>> [4] http://lkml.kernel.org/r/1484128540-22662-1-git-send-email-narmstrong@xxxxxxxxxxxx
>>> [5] http://lkml.kernel.org/r/1484129414-23325-1-git-send-email-narmstrong@xxxxxxxxxxxx
>>>
>>> diff --git a/arch/arm64/boot/dts/amlogic/meson-gx.dtsi b/arch/arm64/boot/dts/amlogic/meson-gx.dtsi
>>> index eada0b5..63d52b7 100644
>>> --- a/arch/arm64/boot/dts/amlogic/meson-gx.dtsi
>>> +++ b/arch/arm64/boot/dts/amlogic/meson-gx.dtsi
>>> @@ -55,6 +55,24 @@
>>>  	#address-cells = <2>;
>>>  	#size-cells = <2>;
>>>  
>>> +	reserved-memory {
>>> +		#address-cells = <2>;
>>> +		#size-cells = <2>;
>>> +		ranges;
>>> +
>>> +		/* 16 MiB reserved for Hardware ROM Firmware */
>>> +		hwrom: hwrom {
>> 
>> Both sub-nodes get a label that is unused, but reserved-memory itself
>> does not (my v4 remark). Intentional?
>> 
>>> +			reg = <0x0 0x0 0x0 0x1000000>;
>>> +			no-map;
>>> +		};
>>> +
>>> +		/* 2 MiB reserved for ARM Trusted Firmware (BL31) */
>>> +		secmon: secmon {
>> 
>> I note that this .dtsi further down has a node /firmware/secure-monitor
>> with label sm.
>> a) Is there any naming convention such as secmon_mem to adopt here to
>> avoid mixups with sm?
>> b) Should this secmon node be referenced in the secure-monitor node via
>> memory-node = <&secmon>; to model their connection, thereby giving the
>> label a use? Or should we maybe merge the two nodes by moving the
>> compatible string here?
>> 
>> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/Documentation/devicetree/bindings/reserved-memory/reserved-memory.txt
>
> Answering my own question: the example labels use _reserved suffix.
>
>>> +			reg = <0x0 0x10000000 0x0 0x200000>;
>
> And since we use a reg property here, the node name should get a unit
> address to avoid future dtc warnings/errors. Ditto for hwrom.

OK, I added Fixes:, your Reviewed-by, added the _reserved suffix and
unit address and applied to v4.10/fixes.

Update patch below for reference.

Other cleanups/fixups (like adding a phandle from secure monitor) can be
done as add-ons, as they are not strictly related to this fix.

Kevin