> On May 30, 2016, at 12:29 PM, Vladimir Olovyannikov
> <vladimir.olovyanni...@broadcom.com> wrote:
>
>> -----Original Message-----
>> From: af...@apple.com [mailto:af...@apple.com]
>> Sent: Monday, May 30, 2016 11:53 AM
>> To: Vladimir Olovyannikov
>> Cc: edk2-de...@ml01.01.org
>> Subject: Re: [edk2] EfiConvertPointer/AllocatePages for an address
> return
>> EFI_NOT FOUND
>>
>>
>>> On May 30, 2016, at 11:38 AM, Vladimir Olovyannikov
>> <vladimir.olovyanni...@broadcom.com> wrote:
>>>
>>> Hi Andrew,
>>> Please see below.
>>> The region of interest is marked as "Reserved".
>>>
>>> Thanks,
>>> Vladimir
>>>
>>>> -----Original Message-----
>>>> From: af...@apple.com [mailto:af...@apple.com]
>>>> Sent: Monday, May 30, 2016 11:32 AM
>>>> To: Vladimir Olovyannikov
>>>> Cc: edk2-de...@ml01.01.org
>>>> Subject: Re: [edk2] EfiConvertPointer/AllocatePages for an address
>>> return
>>>> EFI_NOT FOUND
>>>>
>>>>
>>>>> On May 30, 2016, at 10:37 AM, Vladimir Olovyannikov
>>>> <vladimir.olovyanni...@broadcom.com> wrote:
>>>>>
>>>>>
>>>>> Hello All,
>>>>>
>>>>> I faced an issue with my UEFI implementation (Armv8 64 bit
> platform).
>>>>> I need to convert a device physical address to the virtual on
>>>>> VirtualAddressChange event.
>>>>> EfiConvertPointer always returns EFI_NOT_FOUND.
>>>>>
>>>>> I tried to allocate just an address:
>>>>> gBS->AllocatePages(AllocateAddress, EfiRuntimeServicesCode, 1,
>>>>> &MyAddress); // Allocate one page
>>>>> where MyAddress is 0x65042000.
>>>>> Which also returns EFI_NOT_FOUND.
>>>>>
>>>>> I declared this device's memory in the Platform memory lib as
>>>>>
>>>>> // Mydevice
>>>>> // 1MB @0x65000000 (0x65000000 - 0x650fffff)
>>>>> MMUTable[++Index].PhysicalBase = MY_DEV_BASE;
>>>>> MMUTable[Index].VirtualBase = MY_DEV_BASE;
>>>>> MMUTable[Index].Length = MY_DEV_SZ; // SIZE_1MB
>>>>> MMUTable[Index].Attributes =
>>>>> ARM_MEMORY_REGION_ATTRIBUTE_DEVICE;
>>>>>
>>>>> I can access this device perfectly from BootServices (read/write)
> with
>>> MMU
>>>>> (1:1) on.
>>>>> But it is nowhere in virtual table map which
>>>>> AllocatePages()/EfiConvertPointer() eventually use.
>>>>>
>>>>> Please let me know what I am missing here?
>>>>> I have declared this memory as a Runtime with
>>>>>
>>>>> Status = gDS->AddMemorySpace (
>>>>> EfiGcdMemoryTypeMemoryMappedIo,
>>>>> AlignedAddr, SIZE_64KB,
>>>>> EFI_MEMORY_UC | EFI_MEMORY_RUNTIME
>>>>> );
>>>>> if (!EFI_ERROR (Status)) {
>>>>> Status = gDS->SetMemorySpaceAttributes (
>>>>> AlignedAddr,
>>>>> SIZE_64KB,
>>>>> EFI_MEMORY_UC | EFI_MEMORY_RUNTIME
>>>>> );
>>>>> }
>>>>>
>>>>
>>>> If you boot to the UEFI shell and dump the memory map what to do you
>>> see?
>>>
>>
>> The attribute is wrong, so it was not allocate correctly.
>>
>> #define EFI_MEMORY_RUNTIME 0x8000000000000000
>>
>> Did you check all the return Status from your gDS calls? You can read up
>> about them in the PI spec from the UEF Forum website.
>>
>> If you do a gDS->GetMemorySpaceDescriptor() for the AlignedAddr does it
>> succeed? If yes I think you want to get the descriptor and then set the
>> memory space attribute.
> OK. I think my attempt to build a resource descriptor hob for this device
> was wrong.
> ResourceAttributes = (
> EFI_RESOURCE_ATTRIBUTE_PRESENT |
> EFI_RESOURCE_ATTRIBUTE_INITIALIZED |
> EFI_RESOURCE_ATTRIBUTE_UNCACHEABLE );
> BuildResourceDescriptorHob (EFI_RESOURCE_MEMORY_MAPPED_IO_PORT,
> ResourceAttributes,
> PcdGet32(PcdPlatformResetAddr) &
> ~(SIZE_64KB - 1),
> SIZE_64KB);
> This is what provided the "Reserved".
>
> When I removed it memmap gave me 65002000 as "Mmio" and ConvertPointer
> succeeded.
>
> One thing which is not quite clear to me:
> Should I not declare this device in the Viirtual memory table for the
> platform and just use AddMemorySpace instead ?
>
I was thinking the PI spec would be clear on what to do. I've not had a chance
to reread that section.
>From looking at the Quark code:
~/work/src/edk2(master)>cd QuarkSocPkg; git grep -W GetMemorySpaceDescriptor
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c=371=QNCInitializeResource (
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-372- )
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-373-{
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-374- EFI_PHYSICAL_ADDRESS
BaseAddress;
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-375- EFI_STATUS
Status;
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-376- UINT64
ExtraRegionLength;
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-377- EFI_GCD_MEMORY_SPACE_DESCRIPTOR
Descriptor;
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-378- UINTN
Index;
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-379-
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-380- // Report TSEG range
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-381- // This range maybe has been
reportted in PEI phase via Resource Hob.
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-382- //
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-383- QNCGetTSEGMemoryRange
(&BaseAddress, &ExtraRegionLength);
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-384- if (ExtraRegionLength != 0) {
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c:385: Status =
gDS->GetMemorySpaceDescriptor (BaseAddress, &Descriptor);
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-386- if (Status == EFI_NOT_FOUND) {
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-387- Status = gDS->AddMemorySpace (
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-388-
EfiGcdMemoryTypeReserved,
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-389- BaseAddress,
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-390-
ExtraRegionLength,
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-391- EFI_MEMORY_UC
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-392- );
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-393- }
QuarkNorthCluster/QNCInit/Dxe/QNCInit.c-394- }
So if you construct the HOB then you should just be able to do
gDS->GetMemorySpaceDescriptor(), you only need to gDS->AddMemorySpace() if it
does not exist.
Thanks,
Andrew Fish
> Thank you,
> Vladimir
>>
>> Thanks,
>>
>> Andrew Fish
>>
>>> Shell> memmap
>>> Type Start End # Pages
> Attributes
>>> Available 0000000080000000-0000000084FFFFFF 0000000000005000
>>> 000000000000000F
>>> Available 00000000850F0000-00000000B986FFFF 0000000000034780
>>> 000000000000000F
>>> RT_Data 00000000B9870000-00000000B98DFFFF 0000000000000070
>>> 800000000000000F
>>> Available 00000000B98E0000-00000000B98E5FFF 0000000000000006
>>> 000000000000000F
>>> LoaderCode 00000000B98E6000-00000000B9A12FFF 000000000000012D
>>> 000000000000000F
>>> BS_Code 00000000B9A13000-00000000B9B8FFFF 000000000000017D
>>> 000000000000000F
>>> RT_Data 00000000B9B90000-00000000B9BDFFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9BE0000-00000000B9C2FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Data 00000000B9C30000-00000000B9C7FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9C80000-00000000B9CCFFFF 0000000000000050
>>> 800000000000000F
>>> RT_Data 00000000B9CD0000-00000000B9D1FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9D20000-00000000B9D7FFFF 0000000000000060
>>> 800000000000000F
>>> ACPI_Recl 00000000B9D80000-00000000B9D9FFFF 0000000000000020
>>> 000000000000000F
>>> RT_Data 00000000B9DA0000-00000000B9DEFFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9DF0000-00000000B9E3FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Data 00000000B9E40000-00000000B9E8FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9E90000-00000000B9EDFFFF 0000000000000050
>>> 800000000000000F
>>> RT_Data 00000000B9EE0000-00000000B9F2FFFF 0000000000000050
>>> 800000000000000F
>>> RT_Code 00000000B9F30000-00000000B9F7FFFF 0000000000000050
>>> 800000000000000F
>>> Available 00000000B9F80000-00000000B9F80FFF 0000000000000001
>>> 000000000000000F
>>>
>>> Reserved 00000000B9F81000-00000000B9F84FFF 0000000000000004
>>> 000000000000000F
>>>
>>> BS_Code 00000000B9F85000-00000000B9F8BFFF 0000000000000007
>>> 000000000000000F
>>>
>>> Available 00000000B9F8C000-00000000BCA7AFFF 0000000000002AEF
>>> 000000000000000F
>>>
>>> BS_Data 00000000BCA7B000-00000000BEDBFFFF 0000000000002345
>>> 000000000000000F
>>>
>>> Available 00000000BEDC0000-00000000BEDC5FFF 0000000000000006
>>> 000000000000000F
>>>
>>> BS_Code 00000000BEDC6000-00000000BEF4FFFF 000000000000018A
>>> 000000000000000F
>>>
>>> Available 00000000BEF50000-00000000BEFBFFFF 0000000000000070
>>> 000000000000000F
>>>
>>> RT_Data 00000000BEFC0000-00000000BEFEFFFF 0000000000000030
>>> 800000000000000F
>>>
>>> Available 00000000BEFF0000-00000000BEFFEFFF 000000000000000F
>>> 000000000000000F
>>>
>>> BS_Data 00000000BEFFF000-00000000BEFFFFFF 0000000000000001
>>> 000000000000000F
>>>
>>> Available 00000000BF000000-00000000BF293FFF 0000000000000294
>>> 000000000000000F
>>>
>>> BS_Data 00000000BF294000-00000000BFFFFFFF 0000000000000D6C
>>> 000000000000000F
>>>
>>> Reserved 0000000065020000-000000006502FFFF 0000000000000010
>>> 0000000000000001
>>>
>>>
>>>
>>> Reserved : 20 Pages (81,920 Bytes)
>>>
>>> LoaderCode: 301 Pages (1,232,896 Bytes)
>>>
>>> LoaderData: 0 Pages (0 Bytes)
>>>
>>> BS_Code : 782 Pages (3,203,072 Bytes)
>>>
>>> BS_Data : 12,466 Pages (51,060,736 Bytes)
>>>
>>> RT_Code : 496 Pages (2,031,616 Bytes)
>>>
>>> RT_Data : 640 Pages (2,621,440 Bytes)
>>>
>>> ACPI_Recl : 32 Pages (131,072 Bytes)
>>>
>>> ACPI_NVS : 0 Pages (0 Bytes)
>>>
>>> MMIO : 0 Pages (0 Bytes)
>>>
>>> MMIO_Port : 0 Pages (0 Bytes)
>>>
>>> PalCode : 0 Pages (0 Bytes)
>>>
>>> Available : 247,183 Pages (1,012,461,568 Bytes)
>>>
>>> --------------
>>>
>>> Total Memory: 1,023 MB (1,072,742,400 Bytes)
>>>
>>>
>>>>
>>>> Thanks,
>>>>
>>>> Andrew Fish
>>>>
>>>>> Thank you,
>>>>> Vladimir
>>>>> _______________________________________________
>>>>> edk2-devel mailing list
>>>>> edk2-devel@lists.01.org
>>>>> https://urldefense.proofpoint.com/v2/url?u=https-
>>>> 3A__lists.01.org_mailman_listinfo_edk2-2Ddevel&d=CwICAg&c=Hw-
>>>>
>> EJUFt2_D9PK5csBJ29kRV40HqSDXWTLPyZ6W8u84&r=4sdzHKz0eU1vXqaUyS
>>>> VmyA&m=dj-
>>>>
>> PLSCQ4D53smC8eOprHa7aKtcLYkZDhDWnczTquDU&s=jb4N3rlGt6jI92m5sab
>>>> AKDmsB2T9jdtvYQhw_A91Xi4&e=
>>> _______________________________________________
>>> edk2-devel mailing list
>>> edk2-devel@lists.01.org
>>> https://lists.01.org/mailman/listinfo/edk2-devel
> _______________________________________________
> edk2-devel mailing list
> edk2-devel@lists.01.org
> https://lists.01.org/mailman/listinfo/edk2-devel
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
