public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Min Xu" <min.m.xu@intel.com>
To: "devel@edk2.groups.io" <devel@edk2.groups.io>,
	"kraxel@redhat.com" <kraxel@redhat.com>
Cc: Ard Biesheuvel <ardb+tianocore@kernel.org>,
	"Justen, Jordan L" <jordan.l.justen@intel.com>,
	Brijesh Singh <brijesh.singh@amd.com>,
	"Aktas, Erdem" <erdemaktas@google.com>,
	James Bottomley <jejb@linux.ibm.com>,
	"Yao, Jiewen" <jiewen.yao@intel.com>,
	Tom Lendacky <thomas.lendacky@amd.com>
Subject: Re: [edk2-devel] [PATCH V4 22/31] OvmfPkg: Update PlatformPei to support TDX
Date: Thu, 20 Jan 2022 09:07:22 +0000	[thread overview]
Message-ID: <PH0PR11MB50645F8FBDC046E237AAAFE8C55A9@PH0PR11MB5064.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20211215085305.qb6akg5nlxio4m2h@sirius.home.kraxel.org>

Hi
> > +/**
> > +  Check if it is Tdx guest
> > +
> > +  @retval    TRUE   It is Tdx guest
> > +  @retval    FALSE  It is not Tdx guest
> > +**/
> > +BOOLEAN
> > +PlatformPeiIsTdxGuest (
> > +  VOID
> > +  )
> > +{
> > +  CONFIDENTIAL_COMPUTING_WORK_AREA_HEADER
> *CcWorkAreaHeader;
> > +
> > +  CcWorkAreaHeader =
> (CONFIDENTIAL_COMPUTING_WORK_AREA_HEADER
> > +*)FixedPcdGet32 (PcdOvmfWorkAreaBase);
> > +  return (CcWorkAreaHeader != NULL && CcWorkAreaHeader->GuestType
> ==
> > +GUEST_TYPE_INTEL_TDX); }
> 
> fwcfg patch adds a similar function.  Can't we have a single function
> somewhere, or have a #define for this check, instead of doing this
> cut+paste programming?
There is TdIsEnabled() in BaseLib. So PlatformPeiIsTdxGuest is deprecated in the next version.
> 
> > +VOID
> > +EFIAPI
> > +DEBUG_HOBLIST (
> > +  IN CONST VOID  *HobStart
> > +  )
> > +{
> > +  EFI_PEI_HOB_POINTERS  Hob;
> > +
> > +  Hob.Raw = (UINT8 *)HobStart;
> > +  //
> > +  // Parse the HOB list until end of list or matching type is found.
> > +  //
> > +  while (!END_OF_HOB_LIST (Hob)) {
> > +    DEBUG ((DEBUG_INFO, "HOB(%p) : %x %x\n", Hob, Hob.Header-
> >HobType, Hob.Header->HobLength));
> > +    switch (Hob.Header->HobType) {
> > +      case EFI_HOB_TYPE_RESOURCE_DESCRIPTOR:
> > +        DEBUG ((
> > +          DEBUG_INFO,
> > +          "\t: %x %x %llx %llx\n",
> > +          Hob.ResourceDescriptor->ResourceType,
> > +          Hob.ResourceDescriptor->ResourceAttribute,
> > +          Hob.ResourceDescriptor->PhysicalStart,
> > +          Hob.ResourceDescriptor->ResourceLength
> > +          ));
> > +
> > +        break;
> > +      case EFI_HOB_TYPE_MEMORY_ALLOCATION:
> > +        DEBUG ((
> > +          DEBUG_INFO,
> > +          "\t: %llx %llx %x\n",
> > +          Hob.MemoryAllocation->AllocDescriptor.MemoryBaseAddress,
> > +          Hob.MemoryAllocation->AllocDescriptor.MemoryLength,
> > +          Hob.MemoryAllocation->AllocDescriptor.MemoryType
> > +          ));
> > +        break;
> > +      default:
> > +        break;
> > +    }
> > +
> > +    Hob.Raw = GET_NEXT_HOB (Hob);
> > +  }
> > +}
> 
> Likewise, I've seen this before in another patch of this series.
This will be deprecated in the next version.
> 
> > diff --git a/OvmfPkg/PlatformPei/MemDetect.c
> > b/OvmfPkg/PlatformPei/MemDetect.c index 934d5c196570..9227fa260ccd
> > 100644
> > --- a/OvmfPkg/PlatformPei/MemDetect.c
> > +++ b/OvmfPkg/PlatformPei/MemDetect.c
> > @@ -36,6 +36,7 @@ Module Name:
> >  #include <Library/MtrrLib.h>
> >  #include <Library/QemuFwCfgLib.h>
> >  #include <Library/QemuFwCfgSimpleParserLib.h>
> > +#include <Library/TdxLib.h>
> >
> >  #include "Platform.h"
> >  #include "Cmos.h"
> > @@ -556,7 +557,19 @@ AddressWidthInitialization (
> >      mPhysMemAddressWidth = 36;
> >    }
> >
> > + #if defined (MDE_CPU_X64)
> > +  if (PlatformPeiIsTdxGuest ()) {
> > +    if (TdSharedPageMask () == (1ULL << 47)) {
> > +      mPhysMemAddressWidth = 48;
> > +    } else {
> > +      mPhysMemAddressWidth = 52;
> > +    }
> > +  }
> > +
> > +  ASSERT (mPhysMemAddressWidth <= 52); #else
> 
> Making this TDX-specific looks wrong to me.  5-level paging exists outside TDX
> too.
> 
> Given we don't support 5-level paging (yet) I think we can just drop this and
> revisit in case 5-level paging support is added in the future.
> 
> > +  UINT32   Pml5Entries;
> 
> Same here.
mPhysMemAddressWidth indicates the physical memory address width. Tdx guest supports GPAW of 52 or 48. So I think here mPhysMemAddresWidth is correct.
Even mPhysMemAddress is 52, we can still use 4-level paging. So I am going to update GetPeiMemoryCap() like below:

  if (mPhysMemAddressWidth <= 39) {
    Pml4Entries = 1;
    PdpEntries = 1 << (mPhysMemAddressWidth - 30);
    ASSERT (PdpEntries <= 0x200);
  } else {
    if (TdIsEnabled ()) {           <-- If it is td guest, then Pml4Entries is set to 0x200.
      Pml4Entries = 0x200;
    } else {
      Pml4Entries = 1 << (mPhysMemAddressWidth - 39);
    }
    ASSERT (Pml4Entries <= 0x200);
    PdpEntries = 512;
  }

Thanks
Min

  reply	other threads:[~2022-01-20  9:07 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-13 12:56 [PATCH V4 00/31] Enable Intel TDX in OvmfPkg (Config-A) Min Xu
2021-12-13 12:56 ` [PATCH V4 01/31] MdePkg: Add Tdx.h Min Xu
2021-12-13 12:56 ` [PATCH V4 02/31] MdePkg: Introduce basic Tdx functions in BaseLib Min Xu
2021-12-15  6:33   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 03/31] MdePkg: Add TdxLib to wrap Tdx operations Min Xu
2021-12-15  6:44   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 04/31] UefiCpuPkg: Extend VmgExitLibNull to handle #VE exception Min Xu
2021-12-13 12:56 ` [PATCH V4 05/31] OvmfPkg: Extend VmgExitLib " Min Xu
2021-12-15  6:56   ` [edk2-devel] " Gerd Hoffmann
2022-01-20  6:34     ` Min Xu
2021-12-13 12:56 ` [PATCH V4 06/31] UefiCpuPkg/CpuExceptionHandler: Add base support for the " Min Xu
2021-12-13 12:56 ` [PATCH V4 07/31] MdePkg: Add helper functions for Tdx guest in BaseIoLibIntrinsic Min Xu
2021-12-13 12:56 ` [PATCH V4 08/31] MdePkg: Support mmio " Min Xu
2021-12-15  7:02   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 09/31] MdePkg: Support IoFifo " Min Xu
2021-12-15  7:18   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 10/31] MdePkg: Support IoRead/IoWrite " Min Xu
2021-12-15  7:18   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 11/31] UefiCpuPkg: Support TDX in BaseXApicX2ApicLib Min Xu
2021-12-13 12:56 ` [PATCH V4 12/31] MdePkg: Add macro to check SEV / TDX guest Min Xu
2021-12-13 12:56 ` [PATCH V4 13/31] UefiCpuPkg: Enable Tdx support in MpInitLib Min Xu
2021-12-15  7:33   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 14/31] OvmfPkg: Add IntelTdx.h in OvmfPkg/Include/IndustryStandard Min Xu
2021-12-15  7:37   ` Gerd Hoffmann
2021-12-16  6:43     ` Min Xu
2021-12-13 12:56 ` [PATCH V4 15/31] OvmfPkg: Add TdxMailboxLib Min Xu
2021-12-15  7:47   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 16/31] MdePkg: Add EFI_RESOURCE_ATTRIBUTE_ENCRYPTED in PiHob.h Min Xu
2021-12-13 12:56 ` [PATCH V4 17/31] OvmfPkg: Update Sec to support Tdx Min Xu
2021-12-15  8:18   ` Gerd Hoffmann
2021-12-16  8:11     ` [edk2-devel] " Min Xu
2021-12-13 12:56 ` [PATCH V4 18/31] OvmfPkg: Check Tdx in QemuFwCfgPei to avoid DMA operation Min Xu
2021-12-15  8:19   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 19/31] MdeModulePkg: EFER should not be changed in TDX Min Xu
2021-12-13 12:56 ` [PATCH V4 20/31] MdeModulePkg: Add PcdTdxSharedBitMask Min Xu
2021-12-15  8:22   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 21/31] UefiCpuPkg: Update AddressEncMask in CpuPageTable Min Xu
2021-12-15  8:38   ` Gerd Hoffmann
2021-12-16  8:23     ` Min Xu
2021-12-13 12:56 ` [PATCH V4 22/31] OvmfPkg: Update PlatformPei to support TDX Min Xu
2021-12-15  8:53   ` Gerd Hoffmann
2022-01-20  9:07     ` Min Xu [this message]
2022-01-13 19:18   ` [edk2-devel] " Vishal Annapurve
2022-01-14  2:23     ` Min Xu
2022-01-14  2:37       ` Vishal Annapurve
2022-01-14  2:40         ` Min Xu
2021-12-13 12:56 ` [PATCH V4 23/31] OvmfPkg: Update AcpiPlatformDxe to alter MADT table Min Xu
2021-12-15  8:54   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 24/31] OvmfPkg: Add TdxDxe driver Min Xu
2021-12-15  9:05   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 25/31] OvmfPkg/BaseMemEncryptTdxLib: Add TDX helper library Min Xu
2021-12-15  9:16   ` Gerd Hoffmann
2022-01-21  2:54     ` Min Xu
2022-01-21  8:04       ` Gerd Hoffmann
2022-01-21  8:31         ` [edk2-devel] " Min Xu
2021-12-13 12:56 ` [PATCH V4 26/31] OvmfPkg/QemuFwCfgLib: Support Tdx in QemuFwCfgDxe Min Xu
2021-12-15  9:16   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 27/31] OvmfPkg: Update IoMmuDxe to support TDX Min Xu
2021-12-15  9:18   ` Gerd Hoffmann
2021-12-13 12:56 ` [PATCH V4 28/31] OvmfPkg: Rename XenTimerDxe to LocalApicTimerDxe Min Xu
2021-12-15  9:19   ` Gerd Hoffmann
2021-12-13 12:57 ` [PATCH V4 29/31] UefiCpuPkg: Setting initial-count register as the last step Min Xu
2021-12-15  9:20   ` Gerd Hoffmann
2021-12-13 12:57 ` [PATCH V4 30/31] OvmfPkg: Switch timer in build time for OvmfPkg Min Xu
2021-12-15  9:21   ` Gerd Hoffmann
2021-12-13 12:57 ` [PATCH V4 31/31] OvmfPkg: Move LocalApicTimerDxe to UefiCpuPkg Min Xu
2021-12-15  9:26   ` Gerd Hoffmann
2021-12-16  8:29     ` [edk2-devel] " Min Xu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=PH0PR11MB50645F8FBDC046E237AAAFE8C55A9@PH0PR11MB5064.namprd11.prod.outlook.com \
    --to=devel@edk2.groups.io \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox