[PATCH linux dev-4.7 1/3] ARM: dts: aspeed-bmc-opp-romulus: Enable mailbox and LPC control nodes
Joel Stanley
joel at jms.id.au
Thu Feb 16 16:11:15 AEDT 2017
On Thu, Feb 16, 2017 at 2:02 PM, Mine <mine260309 at gmail.com> wrote:
> Hi Andrew,
>
> On Thu, Feb 16, 2017 at 9:12 AM, Andrew Jeffery <andrew at aj.id.au> wrote:
>> Hi Lei,
>>
>> On Tue, 2017-02-14 at 15:25 +0800, Lei YU wrote:
>>> Also add the reserved-memory node needed by the LPC
>>> control driver.
>>>
>>> > Signed-off-by: Lei YU <mine260309 at gmail.com>
>>> ---
>>> arch/arm/boot/dts/aspeed-bmc-opp-romulus.dts | 22 ++++++++++++++++++++++
>>> 1 file changed, 22 insertions(+)
>>>
>>> diff --git a/arch/arm/boot/dts/aspeed-bmc-opp-romulus.dts b/arch/arm/boot/dts/aspeed-bmc-opp-romulus.dts
>>> index c23627e..f47eb63 100644
>>> --- a/arch/arm/boot/dts/aspeed-bmc-opp-romulus.dts
>>> +++ b/arch/arm/boot/dts/aspeed-bmc-opp-romulus.dts
>>> @@ -19,6 +19,18 @@
>>> > memory {
>>> > reg = <0x80000000 0x40000000>;
>>> > };
>>> +
>>> > + reserved-memory {
>>> > + #address-cells = <1>;
>>> > + #size-cells = <1>;
>>> > + ranges;
>>> +
>>> > > + flash_memory: region at 94000000 {
>>> > + no-map;
>>> + reg = <0x94000000 0x04000000>; /* 64M */
>>
>> So these values are also used in the zaius dts. Why are we reserving
>> 64M 128M down from the top of the RAM address space? Is the top 64M
>> meant to be reserved for anything as well? Why not reserve the top 64M?
The top 64MB is for the host VGA framebuffer.
$ python3 -c "print(((0x80000000 + 0x40000000) - (0x94000000 +
0x04000000)) /1024./1024.)"
640.0
There was a mistake made due to not knowing the size of Zaius' DRAM,
so things are 512MB lower than they should be.
> This is a good question, but honestly I just copied this part from Witherspoon,
> and this is just the same as Zaius.
> Maybe Joel has some comments on this?
The intent was to leave a 64MB window for the host framebuffer, and
then put the 64MB buffer for mbox below.
Can you please fix up this patch and also send one to fix Zaius?
Cheers,
Joel
More information about the openbmc
mailing list