Hi Min,

Thank you for this patch.

Please find my feedback inline marked [SAMI].

Regards,

Sami Mujawar


On 02/11/2021 02:50 AM, Min Xu wrote:
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=3625

DxeTpmMeasurementLib supports TPM based measurement in DXE phase.
After CcMeasurementProtocol is introduced, CC based measurement needs
to be supported in DxeTpmMeasurementLib as well.

In TpmMeasureAndLogData, CC based measurement will be first called.
If it failed, TPM based measurement will be called sequentially.
Currently there is an assumption that CC based measurement and
TPM based measurement won't be exist at the same time.If the
assumption is not true in the future, we will revisit here then.

Cc: Michael D Kinney <[email protected]>
Cc: Liming Gao <[email protected]>
Cc: Zhiguang Liu <[email protected]>
Cc: Jiewen Yao <[email protected]>
Cc: Jian J Wang <[email protected]>
Cc: Sami Mujawar <[email protected]>
Cc: Gerd Hoffmann <[email protected]>
Signed-off-by: Min Xu <[email protected]>
---
  .../DxeTpmMeasurementLib.c                    | 91 ++++++++++++++++++-
  .../DxeTpmMeasurementLib.inf                  |  9 +-
  2 files changed, 92 insertions(+), 8 deletions(-)

diff --git a/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.c 
b/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.c
index 061136ee7860..2ddb9033a0d5 100644
--- a/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.c
+++ b/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.c
@@ -1,5 +1,6 @@
  /** @file
-  This library is used by other modules to measure data to TPM.
+  This library is used by other modules to measure data to TPM and Confidential
+  Computing (CC) measure registers.
Copyright (c) 2012 - 2018, Intel Corporation. All rights reserved. <BR>
  SPDX-License-Identifier: BSD-2-Clause-Patent
@@ -19,8 +20,7 @@ SPDX-License-Identifier: BSD-2-Clause-Patent
#include <Guid/Acpi.h>
  #include <IndustryStandard/Acpi.h>
-
-
+#include <Protocol/CcMeasurement.h>
/**
    Tpm12 measure and log data, and extend the measurement result into a 
specific PCR.
@@ -149,6 +149,73 @@ Tpm20MeasureAndLogData (
    return Status;
  }
+/**
+  Cc measure and log data, and extend the measurement result into a
+  specific CC MR.
+
+  @param[in]  PcrIndex         PCR Index.
+  @param[in]  EventType        Event type.
+  @param[in]  EventLog         Measurement event log.
+  @param[in]  LogLen           Event log length in bytes.
+  @param[in]  HashData         The start of the data buffer to be hashed, 
extended.
+  @param[in]  HashDataLen      The length, in bytes, of the buffer referenced 
by HashData
+
+  @retval EFI_SUCCESS           Operation completed successfully.
+  @retval EFI_UNSUPPORTED       Tdx device not available.
+  @retval EFI_OUT_OF_RESOURCES  Out of memory.
+  @retval EFI_DEVICE_ERROR      The operation was unsuccessful.
+**/
+EFI_STATUS
+EFIAPI
+CcMeasureAndLogData (
+  IN UINT32             PcrIndex,
+  IN UINT32             EventType,
+  IN VOID               *EventLog,
+  IN UINT32             LogLen,
+  IN VOID               *HashData,
+  IN UINT64             HashDataLen
+  )
+{
+  EFI_STATUS                    Status;
+  EFI_CC_MEASUREMENT_PROTOCOL  *CcProtocol;
+  EFI_CC_EVENT                 *EfiCcEvent;
+  UINT32                        MrIndex;
[SAMI] Same comment as in patch 2/3. Is it possible to use the typedef for the measurment register index here, please?
+
+  Status = gBS->LocateProtocol (&gEfiCcMeasurementProtocolGuid, NULL, (VOID **) 
&CcProtocol);
+  if (EFI_ERROR (Status)) {
+    return Status;
+  }
+
+  Status = CcProtocol->MapPcrToMrIndex (CcProtocol, PcrIndex, &MrIndex);
+  if (EFI_ERROR (Status)) {
+    return EFI_INVALID_PARAMETER;
[SAMI] Is it possible to return the error code returned by CcProtocol->MapPcrToMrIndex(), please?
+  }
+
+  EfiCcEvent = (EFI_CC_EVENT *) AllocateZeroPool (LogLen + sizeof 
(EFI_CC_EVENT));
+  if(EfiCcEvent == NULL) {
+    return EFI_OUT_OF_RESOURCES;
+  }
+
+  EfiCcEvent->Size = (UINT32) LogLen + sizeof (EFI_CC_EVENT) - sizeof 
(EfiCcEvent->Event);
+  EfiCcEvent->Header.HeaderSize    = sizeof (EFI_CC_EVENT_HEADER);
+  EfiCcEvent->Header.HeaderVersion = EFI_CC_EVENT_HEADER_VERSION;
+  EfiCcEvent->Header.MrIndex       = MrIndex;
+  EfiCcEvent->Header.EventType     = EventType;
+  CopyMem (&EfiCcEvent->Event[0], EventLog, LogLen);
+
+  Status = CcProtocol->HashLogExtendEvent (
+                           CcProtocol,
+                           0,
+                           (EFI_PHYSICAL_ADDRESS) (UINTN) HashData,
+                           HashDataLen,
+                           EfiCcEvent
+                           );
+  FreePool (EfiCcEvent);
+
+  return Status;
+}
+
+
  /**
    Tpm measure and log data, and extend the measurement result into a specific 
PCR.
@@ -178,9 +245,9 @@ TpmMeasureAndLogData (
    EFI_STATUS  Status;
//
-  // Try to measure using Tpm20 protocol
+  // Try to measure using Cc measurement protocol
    //
-  Status = Tpm20MeasureAndLogData(
+  Status = CcMeasureAndLogData (
               PcrIndex,
               EventType,
               EventLog,
@@ -189,6 +256,20 @@ TpmMeasureAndLogData (
               HashDataLen
               );
+ if (EFI_ERROR (Status)) {
+    //
+    // Try to measure using Tpm20 protocol
+    //
+    Status = Tpm20MeasureAndLogData(
+               PcrIndex,
+               EventType,
+               EventLog,
+               LogLen,
+               HashData,
+               HashDataLen
+               );
+  }
[SAMI] Apologies, I missed this in my previous review. I think the behaviour if both the TCG2 and CC measurement protocols are installed would be inconsistent between DxeTpmMeasurementLib and DxeTpm2MeasureBootLib. The main difference being in the later, the
TCG2 protocol takes precedence for extending the measurement.
I think it would be good to modify DxeTpm2MeasureBootLib so that the CC measurement protocol is used if both protocols are installed. What do you think?

There is another subtle difference in this patch. Consider the case where LocateProtocol(gEfiCcMeasurementProtocolGuid, ... ) succeeds but subsequently MapPcrToMrIndex() or HashLogExtendEvent() fail then Tpm20MeasureAndLogData() will be called. This is not the case with DxeTpm2MeasureBootLib.
I think it would be good to have a consistent behaviour.
 [/SAMI]
+
    if (EFI_ERROR (Status)) {
      //
      // Try to measure using Tpm1.2 protocol
diff --git a/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.inf 
b/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.inf
index 7d41bc41f95d..3af3d4e33b25 100644
--- a/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.inf
+++ b/SecurityPkg/Library/DxeTpmMeasurementLib/DxeTpmMeasurementLib.inf
@@ -1,5 +1,7 @@
  ## @file
-#  Provides TPM measurement functions for TPM1.2 and TPM 2.0
+#  Provides below measurement functions:
+#    1. TPM measurement functions for TPM1.2 and TPM 2.0
+#    2. Confidential Computing (CC) measurement functions
  #
  #  This library provides TpmMeasureAndLogData() to measure and log data, and
  #  extend the measurement result into a specific PCR.
@@ -40,5 +42,6 @@
    UefiBootServicesTableLib
[Protocols]
-  gEfiTcgProtocolGuid           ## SOMETIMES_CONSUMES
-  gEfiTcg2ProtocolGuid          ## SOMETIMES_CONSUMES
+  gEfiTcgProtocolGuid               ## SOMETIMES_CONSUMES
+  gEfiTcg2ProtocolGuid              ## SOMETIMES_CONSUMES
+  gEfiCcMeasurementProtocolGuid     ## SOMETIMES_CONSUMES



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#83114): https://edk2.groups.io/g/devel/message/83114
Mute This Topic: https://groups.io/mt/86758672/21656
Group Owner: [email protected]
Unsubscribe: https://edk2.groups.io/g/devel/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-


Reply via email to