Re: [PATCH v3] OvmfPkg/ResetVector: Depend on PCD values of the page tables.

[Laszlo Ersek <lersek@redhat.com>]

Three comments:

On 11/03/16 22:41, Marvin Häuser wrote:
> Currently, the value of the page tables' address is hard-coded in the
> ResetVector. This patch replaces these values with a PCD dependency.
> 
> A check for the size has been added to alert the developer to rewrite
> the ASM according to the new size, if it has been changed.
> 
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Marvin Haeuser <Marvin.Haeuser@outlook.com>
> ---
>  OvmfPkg/ResetVector/Ia32/PageTables64.asm | 23 ++++++++++----------
>  OvmfPkg/ResetVector/ResetVector.inf       |  5 +++++
>  OvmfPkg/ResetVector/ResetVector.nasmb     |  6 +++++
>  3 files changed, 23 insertions(+), 11 deletions(-)
> 
> diff --git a/OvmfPkg/ResetVector/Ia32/PageTables64.asm b/OvmfPkg/ResetVector/Ia32/PageTables64.asm
> index b5a4cf8d7187..0b95a7fa9a86 100644
> --- a/OvmfPkg/ResetVector/Ia32/PageTables64.asm
> +++ b/OvmfPkg/ResetVector/Ia32/PageTables64.asm
> @@ -44,10 +44,11 @@ BITS    32
>  SetCr3ForPageTables64:
>  
>      ;
> -    ; For OVMF, build some initial page tables at 0x800000-0x806000.
> +    ; For OVMF, build some initial page tables at
> +    ; PcdOvmfSecPageTablesBase - (PcdOvmfSecPageTablesBase + 0x6000).
>      ;
> -    ; This range should match with PcdOvmfSecPageTablesBase and
> -    ; PcdOvmfSecPageTablesSize which are declared in the FDF files.
> +    ; This range should match with PcdOvmfSecPageTablesSize which is
> +    ; declared in the FDF files.
>      ;
>      ; At the end of PEI, the pages tables will be rebuilt into a
>      ; more permanent location by DxeIpl.
> @@ -56,21 +57,21 @@ SetCr3ForPageTables64:
>      mov     ecx, 6 * 0x1000 / 4
>      xor     eax, eax
>  clearPageTablesMemoryLoop:
> -    mov     dword[ecx * 4 + 0x800000 - 4], eax
> +    mov     dword[ecx * 4 + PT_ADDR (0 - 4)], eax

I think this is incorrect (or dubious at least); the NASM-level
replacement text for this will be

  (0x800000 + (0 - 4))

and I strongly dislike the (0 - 4) subexpression. The type in which NASM
evaluates (0 - 4) is not specified, and I'd rather not risk it being
UINT64, for example.

>      loop    clearPageTablesMemoryLoop
>  
>      ;
>      ; Top level Page Directory Pointers (1 * 512GB entry)
>      ;
> -    mov     dword[0x800000], 0x801000 + PAGE_PDP_ATTR
> +    mov     dword[PT_ADDR (0)], PT_ADDR (0x1000) + PAGE_PDP_ATTR
>  
>      ;
>      ; Next level Page Directory Pointers (4 * 1GB entries => 4GB)
>      ;
> -    mov     dword[0x801000], 0x802000 + PAGE_PDP_ATTR
> -    mov     dword[0x801008], 0x803000 + PAGE_PDP_ATTR
> -    mov     dword[0x801010], 0x804000 + PAGE_PDP_ATTR
> -    mov     dword[0x801018], 0x805000 + PAGE_PDP_ATTR
> +    mov     dword[PT_ADDR (0x1000)], PT_ADDR (0x2000) + PAGE_PDP_ATTR
> +    mov     dword[PT_ADDR (0x1008)], PT_ADDR (0x3000) + PAGE_PDP_ATTR
> +    mov     dword[PT_ADDR (0x1010)], PT_ADDR (0x4000) + PAGE_PDP_ATTR
> +    mov     dword[PT_ADDR (0x1018)], PT_ADDR (0x5000) + PAGE_PDP_ATTR
>  
>      ;
>      ; Page Table Entries (2048 * 2MB entries => 4GB)
> @@ -81,13 +82,13 @@ pageTableEntriesLoop:
>      dec     eax
>      shl     eax, 21
>      add     eax, PAGE_2M_PDE_ATTR
> -    mov     [ecx * 8 + 0x802000 - 8], eax
> +    mov     [ecx * 8 + PT_ADDR (0x2000 - 8)], eax
>      loop    pageTableEntriesLoop
>  
>      ;
>      ; Set CR3 now that the paging structures are available
>      ;
> -    mov     eax, 0x800000
> +    mov     eax, PT_ADDR (0)
>      mov     cr3, eax
>  
>      OneTimeCallRet SetCr3ForPageTables64
> diff --git a/OvmfPkg/ResetVector/ResetVector.inf b/OvmfPkg/ResetVector/ResetVector.inf
> index 46610d243ecf..d1e5d4d9bdea 100644
> --- a/OvmfPkg/ResetVector/ResetVector.inf
> +++ b/OvmfPkg/ResetVector/ResetVector.inf
> @@ -29,9 +29,14 @@ [Sources]
>    ResetVector.nasmb
>  
>  [Packages]
> +  OvmfPkg/OvmfPkg.dec
>    MdePkg/MdePkg.dec
>    UefiCpuPkg/UefiCpuPkg.dec
>  
>  [BuildOptions]
>     *_*_IA32_NASMB_FLAGS = -I$(WORKSPACE)/UefiCpuPkg/ResetVector/Vtf0/
>     *_*_X64_NASMB_FLAGS = -I$(WORKSPACE)/UefiCpuPkg/ResetVector/Vtf0/
> +
> +[Pcd]
> +  gUefiOvmfPkgTokenSpaceGuid.PcdOvmfSecPageTablesBase
> +  gUefiOvmfPkgTokenSpaceGuid.PcdOvmfSecPageTablesSize
> diff --git a/OvmfPkg/ResetVector/ResetVector.nasmb b/OvmfPkg/ResetVector/ResetVector.nasmb
> index 31ac06ae4a8c..4bf8c97d8158 100644
> --- a/OvmfPkg/ResetVector/ResetVector.nasmb
> +++ b/OvmfPkg/ResetVector/ResetVector.nasmb
> @@ -53,6 +53,12 @@
>  %include "Ia32/SearchForSecEntry.asm"
>  
>  %ifdef ARCH_X64
> +  %if (FixedPcdGet32 (PcdOvmfSecPageTablesSize) != 0x6000)
> +    %error "This implementation inherently depends on PcdOvmfSecPageTablesSize"
> +  %endif
> +
> +  #include <AutoGen.h>

I think this order is wrong. The C preprocessor won't fully expand
FixedPcdGet32 (PcdOvmfSecPageTablesSize) I believe, because AutoGen.h is
included after the use of the macro. In turn NASM won't find a plain
numeric constant on the LHS of the inequality.

> +  %define PT_ADDR(Offset) FixedPcdGet32 (PcdOvmfSecPageTablesBase)

This misses the Offset macro param in the replacement text.

>  %include "Ia32/Flat32ToFlat64.asm"
>  %include "Ia32/PageTables64.asm"
>  %endif
> 

Marvin, please send a v4, or else please respond with your R-b to "my"
v3 (and then we can commit that series, if Jordan agrees) -- either
works for me.

Thanks
Laszlo