* [edk2-devel] [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
@ 2023-10-23 13:04 ` Konstantin Aladyshev
2023-10-25 8:55 ` Attar, AbdulLateef (Abdul Lateef) via groups.io
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values Konstantin Aladyshev
` (15 subsequent siblings)
16 siblings, 1 reply; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:04 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Currently there is only a definition for the MCTP KCS HEADER structure.
Add definition for the MCTP KCS TRAILER structure as well.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Library/ManageabilityTransportMctpLib.h | 5 +++++
.../MctpProtocol/Common/MctpProtocolCommon.c | 14 +++++++-------
2 files changed, 12 insertions(+), 7 deletions(-)
diff --git a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
index 43bd142f4c..462e7436e6 100644
--- a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
+++ b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
@@ -39,6 +39,11 @@ typedef struct {
UINT8 DefiningBody; ///< Message type.
UINT8 ByteCount; ///< Byte count of payload.
} MANAGEABILITY_MCTP_KCS_HEADER;
+
+typedef struct {
+ UINT8 Pec; ///< MCTP over KCS Packet Error Code.
+} MANAGEABILITY_MCTP_KCS_TRAILER;
+
#define MCTP_KCS_NETFN_LUN 0xb0
#define DEFINING_BODY_DMTF_PRE_OS_WORKING_GROUP 0x01
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 1ad48efdc7..7576007f77 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -132,7 +132,7 @@ SetupMctpRequestTransportPacket (
MANAGEABILITY_MCTP_KCS_HEADER *MctpKcsHeader;
MCTP_TRANSPORT_HEADER *MctpTransportHeader;
MCTP_MESSAGE_HEADER *MctpMessageHeader;
- UINT8 *Pec;
+ MANAGEABILITY_MCTP_KCS_TRAILER *MctpKcsTrailer;
UINT8 *ThisPackage;
if ((PacketHeader == NULL) || (PacketHeaderSize == NULL) ||
@@ -151,8 +151,8 @@ SetupMctpRequestTransportPacket (
return EFI_OUT_OF_RESOURCES;
}
- Pec = (UINT8 *)AllocateZeroPool (sizeof (UINT8));
- if (Pec == NULL) {
+ MctpKcsTrailer = (MANAGEABILITY_MCTP_KCS_TRAILER *)AllocateZeroPool (sizeof (MANAGEABILITY_MCTP_KCS_TRAILER));
+ if (MctpKcsTrailer == NULL) {
DEBUG ((DEBUG_ERROR, "%a: Not enough resource for PEC.\n", __func__));
FreePool (MctpKcsHeader);
return EFI_OUT_OF_RESOURCES;
@@ -167,7 +167,7 @@ SetupMctpRequestTransportPacket (
if (ThisPackage == NULL) {
DEBUG ((DEBUG_ERROR, "%a: Not enough resource for package.\n", __func__));
FreePool (MctpKcsHeader);
- FreePool (Pec);
+ FreePool (MctpKcsTrailer);
return EFI_OUT_OF_RESOURCES;
}
@@ -193,14 +193,14 @@ SetupMctpRequestTransportPacket (
//
// Generate PEC follow SMBUS 2.0 specification.
- *Pec = HelperManageabilityGenerateCrc8 (MCTP_KCS_PACKET_ERROR_CODE_POLY, 0, ThisPackage, MctpKcsHeader->ByteCount);
+ MctpKcsTrailer->Pec = HelperManageabilityGenerateCrc8 (MCTP_KCS_PACKET_ERROR_CODE_POLY, 0, ThisPackage, MctpKcsHeader->ByteCount);
*PacketBody = (UINT8 *)ThisPackage;
*PacketBodySize = MctpKcsHeader->ByteCount;
- *PacketTrailer = (MANAGEABILITY_TRANSPORT_TRAILER)Pec;
+ *PacketTrailer = (MANAGEABILITY_TRANSPORT_TRAILER)MctpKcsTrailer;
*PacketHeader = (MANAGEABILITY_TRANSPORT_HEADER)MctpKcsHeader;
*PacketHeaderSize = sizeof (MANAGEABILITY_MCTP_KCS_HEADER);
- *PacketTrailerSize = 1;
+ *PacketTrailerSize = sizeof (MANAGEABILITY_MCTP_KCS_TRAILER);
return EFI_SUCCESS;
} else {
DEBUG ((DEBUG_ERROR, "%a: No implementation of building up packet.", __func__));
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109909): https://edk2.groups.io/g/devel/message/109909
Mute This Topic: https://groups.io/mt/102134647/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* Re: [edk2-devel] [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure Konstantin Aladyshev
@ 2023-10-25 8:55 ` Attar, AbdulLateef (Abdul Lateef) via groups.io
0 siblings, 0 replies; 25+ messages in thread
From: Attar, AbdulLateef (Abdul Lateef) via groups.io @ 2023-10-25 8:55 UTC (permalink / raw)
To: Konstantin Aladyshev, devel@edk2.groups.io
Cc: Chang, Abner, nicklew@nvidia.com
[AMD Official Use Only - General]
Reviewed-by: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>
-----Original Message-----
From: Konstantin Aladyshev <aladyshev22@gmail.com>
Sent: Monday, October 23, 2023 6:35 PM
To: devel@edk2.groups.io
Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin Aladyshev <aladyshev22@gmail.com>
Subject: [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
Currently there is only a definition for the MCTP KCS HEADER structure.
Add definition for the MCTP KCS TRAILER structure as well.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Library/ManageabilityTransportMctpLib.h | 5 +++++
.../MctpProtocol/Common/MctpProtocolCommon.c | 14 +++++++-------
2 files changed, 12 insertions(+), 7 deletions(-)
diff --git a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
index 43bd142f4c..462e7436e6 100644
--- a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
+++ b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
@@ -39,6 +39,11 @@ typedef struct {
UINT8 DefiningBody; ///< Message type.
UINT8 ByteCount; ///< Byte count of payload.
} MANAGEABILITY_MCTP_KCS_HEADER;
+
+typedef struct {
+ UINT8 Pec; ///< MCTP over KCS Packet Error Code.
+} MANAGEABILITY_MCTP_KCS_TRAILER;
+
#define MCTP_KCS_NETFN_LUN 0xb0
#define DEFINING_BODY_DMTF_PRE_OS_WORKING_GROUP 0x01
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 1ad48efdc7..7576007f77 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -132,7 +132,7 @@ SetupMctpRequestTransportPacket (
MANAGEABILITY_MCTP_KCS_HEADER *MctpKcsHeader;
MCTP_TRANSPORT_HEADER *MctpTransportHeader;
MCTP_MESSAGE_HEADER *MctpMessageHeader;
- UINT8 *Pec;
+ MANAGEABILITY_MCTP_KCS_TRAILER *MctpKcsTrailer;
UINT8 *ThisPackage;
if ((PacketHeader == NULL) || (PacketHeaderSize == NULL) ||
@@ -151,8 +151,8 @@ SetupMctpRequestTransportPacket (
return EFI_OUT_OF_RESOURCES;
}
- Pec = (UINT8 *)AllocateZeroPool (sizeof (UINT8));
- if (Pec == NULL) {
+ MctpKcsTrailer = (MANAGEABILITY_MCTP_KCS_TRAILER *)AllocateZeroPool (sizeof (MANAGEABILITY_MCTP_KCS_TRAILER));
+ if (MctpKcsTrailer == NULL) {
DEBUG ((DEBUG_ERROR, "%a: Not enough resource for PEC.\n", __func__));
FreePool (MctpKcsHeader);
return EFI_OUT_OF_RESOURCES;
@@ -167,7 +167,7 @@ SetupMctpRequestTransportPacket (
if (ThisPackage == NULL) {
DEBUG ((DEBUG_ERROR, "%a: Not enough resource for package.\n", __func__));
FreePool (MctpKcsHeader);
- FreePool (Pec);
+ FreePool (MctpKcsTrailer);
return EFI_OUT_OF_RESOURCES;
}
@@ -193,14 +193,14 @@ SetupMctpRequestTransportPacket (
//
// Generate PEC follow SMBUS 2.0 specification.
- *Pec = HelperManageabilityGenerateCrc8 (MCTP_KCS_PACKET_ERROR_CODE_POLY, 0, ThisPackage, MctpKcsHeader->ByteCount);
+ MctpKcsTrailer->Pec = HelperManageabilityGenerateCrc8 (MCTP_KCS_PACKET_ERROR_CODE_POLY, 0, ThisPackage, MctpKcsHeader->ByteCount);
*PacketBody = (UINT8 *)ThisPackage;
*PacketBodySize = MctpKcsHeader->ByteCount;
- *PacketTrailer = (MANAGEABILITY_TRANSPORT_TRAILER)Pec;
+ *PacketTrailer = (MANAGEABILITY_TRANSPORT_TRAILER)MctpKcsTrailer;
*PacketHeader = (MANAGEABILITY_TRANSPORT_HEADER)MctpKcsHeader;
*PacketHeaderSize = sizeof (MANAGEABILITY_MCTP_KCS_HEADER);
- *PacketTrailerSize = 1;
+ *PacketTrailerSize = sizeof (MANAGEABILITY_MCTP_KCS_TRAILER);
return EFI_SUCCESS;
} else {
DEBUG ((DEBUG_ERROR, "%a: No implementation of building up packet.", __func__));
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110056): https://edk2.groups.io/g/devel/message/110056
Mute This Topic: https://groups.io/mt/102134647/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread
* [edk2-devel] [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure Konstantin Aladyshev
@ 2023-10-23 13:04 ` Konstantin Aladyshev
2023-10-25 8:57 ` Attar, AbdulLateef (Abdul Lateef) via groups.io
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 03/16] ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library Konstantin Aladyshev
` (14 subsequent siblings)
16 siblings, 1 reply; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:04 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
MTCP base specification marks EIDs 1-7 as reserved. Therefore return
EFI_INVALID_PARAMETER if such EIDs were provided to the
MctpSubmitMessage function.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Universal/MctpProtocol/Dxe/MctpProtocol.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
index 88bfd9b7e7..d0f49a1abb 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
@@ -78,6 +78,23 @@ MctpSubmitMessage (
return EFI_INVALID_PARAMETER;
}
+ //
+ // Check source EID and destination EID
+ //
+ if ((MctpSourceEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (MctpSourceEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ ) {
+ DEBUG ((DEBUG_ERROR, "%a: The value of MCTP source EID (%x) is reserved.\n", __func__, MctpSourceEndpointId));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ if ((MctpDestinationEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (MctpDestinationEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ ) {
+ DEBUG ((DEBUG_ERROR, "%a: The value of MCTP destination EID (%x) is reserved.\n", __func__, MctpDestinationEndpointId));
+ return EFI_INVALID_PARAMETER;
+ }
+
Status = CommonMctpSubmitMessage (
mTransportToken,
MctpType,
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109910): https://edk2.groups.io/g/devel/message/109910
Mute This Topic: https://groups.io/mt/102134648/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* Re: [edk2-devel] [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values Konstantin Aladyshev
@ 2023-10-25 8:57 ` Attar, AbdulLateef (Abdul Lateef) via groups.io
0 siblings, 0 replies; 25+ messages in thread
From: Attar, AbdulLateef (Abdul Lateef) via groups.io @ 2023-10-25 8:57 UTC (permalink / raw)
To: Konstantin Aladyshev, devel@edk2.groups.io
Cc: Chang, Abner, nicklew@nvidia.com
[AMD Official Use Only - General]
Reviewed-by: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>;
-----Original Message-----
From: Konstantin Aladyshev <aladyshev22@gmail.com>
Sent: Monday, October 23, 2023 6:35 PM
To: devel@edk2.groups.io
Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin Aladyshev <aladyshev22@gmail.com>
Subject: [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
MTCP base specification marks EIDs 1-7 as reserved. Therefore return EFI_INVALID_PARAMETER if such EIDs were provided to the MctpSubmitMessage function.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Universal/MctpProtocol/Dxe/MctpProtocol.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
index 88bfd9b7e7..d0f49a1abb 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.
+++ c
@@ -78,6 +78,23 @@ MctpSubmitMessage (
return EFI_INVALID_PARAMETER;
}
+ //
+ // Check source EID and destination EID
+ //
+ if ((MctpSourceEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (MctpSourceEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ ) {
+ DEBUG ((DEBUG_ERROR, "%a: The value of MCTP source EID (%x) is
+ reserved.\n", __func__, MctpSourceEndpointId));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ if ((MctpDestinationEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (MctpDestinationEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ ) {
+ DEBUG ((DEBUG_ERROR, "%a: The value of MCTP destination EID (%x) is
+ reserved.\n", __func__, MctpDestinationEndpointId));
+ return EFI_INVALID_PARAMETER;
+ }
+
Status = CommonMctpSubmitMessage (
mTransportToken,
MctpType,
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110057): https://edk2.groups.io/g/devel/message/110057
Mute This Topic: https://groups.io/mt/102134648/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply [flat|nested] 25+ messages in thread
* [edk2-devel] [PATCH edk2-platforms v3 03/16] ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 01/16] ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure Konstantin Aladyshev
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 02/16] ManageabilityPkg: Check MCTP EIDs for reserved values Konstantin Aladyshev
@ 2023-10-23 13:04 ` Konstantin Aladyshev
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 04/16] ManageabilityPkg: Check header fields in the MCTP response Konstantin Aladyshev
` (13 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:04 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
The Manageability KCS transport library needs to support requests both
from MCTP and IPMI transports. Currently the code only handles IPMI
case correctly.
In the MCTP case the communication should be based on the MCTP-over-KCS
specification (DSP0254). This specification defines a special KCS
binding header and trailer structures that need to be present in every
MCTP message.
The header structure contains a length field, therefore response packet
size is not needed to be known beforehand.
The trailer structure contains a PEC checksum that can be used to check
itegrity of the response message.
Modify Manageability KCS transport library code to check which message
is processed (IPMI or MCTP) and handle each case correctly based on its
own specification.
Tested:
- The IPMI KCS communication is tested by Abner Chang,
- The MCTP KCS communication is tested by Konstantin Aladyshev on the
AMD EthanolX CRB.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Common/KcsCommon.c | 284 +++++++++++++++---
.../MctpProtocol/Common/MctpProtocolCommon.c | 14 +-
2 files changed, 260 insertions(+), 38 deletions(-)
diff --git a/Features/ManageabilityPkg/Library/ManageabilityTransportKcsLib/Common/KcsCommon.c b/Features/ManageabilityPkg/Library/ManageabilityTransportKcsLib/Common/KcsCommon.c
index d5b54c04be..4f7e7d450f 100644
--- a/Features/ManageabilityPkg/Library/ManageabilityTransportKcsLib/Common/KcsCommon.c
+++ b/Features/ManageabilityPkg/Library/ManageabilityTransportKcsLib/Common/KcsCommon.c
@@ -8,16 +8,19 @@
**/
#include <Uefi.h>
#include <IndustryStandard/IpmiKcs.h>
+#include <IndustryStandard/Mctp.h>
#include <Library/BaseMemoryLib.h>
#include <Library/IoLib.h>
#include <Library/DebugLib.h>
#include <Library/ManageabilityTransportHelperLib.h>
+#include <Library/ManageabilityTransportMctpLib.h>
#include <Library/MemoryAllocationLib.h>
#include <Library/TimerLib.h>
#include "ManageabilityTransportKcs.h"
extern MANAGEABILITY_TRANSPORT_KCS_HARDWARE_INFO mKcsHardwareInfo;
+extern MANAGEABILITY_TRANSPORT_KCS *mSingleSessionToken;
/**
This function waits for parameter Flag to set.
@@ -379,6 +382,218 @@ KcsTransportRead (
return EFI_SUCCESS;
}
+/**
+ This funciton checks the KCS response data according to
+ manageability protocol.
+
+ @param[in] ResponseData Pointer to response data.
+ @param[in] ResponseDataSize Size of response data.
+ @param[out] AdditionalStatus Pointer to receive the additional status.
+
+ @retval EFI_SUCCESS KCS response header is checked without error
+ @retval EFI_DEVICE_ERROR KCS response header has problem.
+**/
+EFI_STATUS
+KcsCheckResponseData (
+ IN UINT8 *ResponseData,
+ IN UINT32 ResponseDataSize,
+ OUT MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS *AdditionalStatus
+ )
+{
+ EFI_STATUS Status;
+ MANAGEABILITY_MCTP_KCS_TRAILER MctpKcsPec;
+ UINT32 PecSize;
+ UINT8 CalculatedPec;
+ CHAR16 *CompletionCodeStr;
+
+ Status = EFI_SUCCESS;
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_NO_ERRORS;
+ if (CompareGuid (&gManageabilityProtocolMctpGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+ //
+ // For MCTP over KCS, check PEC
+ //
+ PecSize = sizeof (MANAGEABILITY_MCTP_KCS_TRAILER) + 1; // +1 to read last dummy byte that finishes KCS transfer
+ Status = KcsTransportRead (&MctpKcsPec.Pec, &PecSize);
+ if (EFI_ERROR (Status)) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Failed to read PEC with Status(%r)\n",
+ __func__,
+ Status
+ ));
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_ERROR;
+ return Status;
+ }
+
+ if (PecSize != sizeof (MctpKcsPec)) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Received PEC size is %d instead of %d\n",
+ __func__,
+ PecSize,
+ sizeof (MctpKcsPec)
+ ));
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_ERROR;
+ return EFI_DEVICE_ERROR;
+ }
+
+ HelperManageabilityDebugPrint ((VOID *)&MctpKcsPec.Pec, PecSize - 1, "MCTP over KCS Response PEC:\n");
+ CalculatedPec = HelperManageabilityGenerateCrc8 (MCTP_KCS_PACKET_ERROR_CODE_POLY, 0, ResponseData, ResponseDataSize);
+ if (CalculatedPec != MctpKcsPec.Pec) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Received PEC is 0x%02x instead of 0x%02x\n",
+ __func__,
+ MctpKcsPec.Pec,
+ CalculatedPec
+ ));
+ Status = EFI_DEVICE_ERROR;
+ }
+ } else if (CompareGuid (&gManageabilityProtocolIpmiGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+ //
+ // For IPMI over KCS
+ // Check and print Completion Code
+ //
+ Status = IpmiHelperCheckCompletionCode (*ResponseData, &CompletionCodeStr, AdditionalStatus);
+ if (!EFI_ERROR (Status)) {
+ DEBUG ((DEBUG_MANAGEABILITY_INFO, "Cc: %02x %s.\n", *((UINT8 *)ResponseData), CompletionCodeStr));
+ } else if (Status == EFI_NOT_FOUND) {
+ DEBUG ((DEBUG_ERROR, "Cc: %02x not defined in IpmiCompletionCodeMapping or invalid.\n", *((UINT8 *)ResponseData)));
+ }
+ }
+
+ return Status;
+}
+
+/**
+ This funciton reads the KCS response header according to
+ manageability protocol. Caller has to free the memory
+ allocated for response header.
+
+ @param[in] ResponseHeader Pointer to receive the response header.
+ @param[out] AdditionalStatus Pointer to receive the additional status.
+
+ @retval EFI_SUCCESS KCS response header is checked and returned
+ to caller.
+ @retval EFI_INVALID_PARAMETER One of the given parameter is incorrect.
+ @retval EFI_OUT_OF_RESOURCE Memory allocation is failed for ResponseHeader.
+ @retval EFI_DEVICE_ERROR Incorrect response header.
+**/
+EFI_STATUS
+KcsReadResponseHeader (
+ IN UINT8 **ResponseHeader,
+ OUT MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS *AdditionalStatus
+ )
+{
+ EFI_STATUS Status;
+ UINT32 RspHeaderSize;
+ UINT32 ExpectedHeaderSize;
+ UINT8 *RspHeader;
+
+ if ((ResponseHeader == NULL) || (AdditionalStatus == NULL)) {
+ return EFI_INVALID_PARAMETER;
+ }
+
+ *ResponseHeader = NULL;
+ if (CompareGuid (&gManageabilityProtocolMctpGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+
+ // For MCTP over KCS
+ ExpectedHeaderSize = sizeof (MANAGEABILITY_MCTP_KCS_HEADER);
+ DEBUG ((
+ DEBUG_MANAGEABILITY_INFO,
+ "%a: Reading MCTP over KCS response header.\n",
+ __func__
+ ));
+ } else if (CompareGuid (&gManageabilityProtocolIpmiGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+ // For IPMI over KCS
+
+ ExpectedHeaderSize = sizeof (IPMI_KCS_RESPONSE_HEADER);
+ DEBUG ((
+ DEBUG_MANAGEABILITY_INFO,
+ "%a: Reading IPMI over KCS response header.\n",
+ __func__
+ ));
+ } else {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Unsupportted manageability protocol over KCS: %g.\n",
+ __func__,
+ mSingleSessionToken->Token.ManageabilityProtocolSpecification
+ ));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ RspHeader = (UINT8 *)AllocateZeroPool (ExpectedHeaderSize);
+ if (RspHeader == NULL) {
+ DEBUG ((DEBUG_ERROR, "Memory allocation failed for KCS response header!\n"));
+ return EFI_OUT_OF_RESOURCES;
+ }
+
+ RspHeaderSize = ExpectedHeaderSize;
+ Status = KcsTransportRead (RspHeader, &RspHeaderSize);
+ if (EFI_ERROR (Status)) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Failed to read KCS response header Status(%r)\n",
+ __func__,
+ Status
+ ));
+ FreePool (RspHeader);
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_ERROR;
+ return Status;
+ }
+
+ if (RspHeaderSize != 0) {
+ HelperManageabilityDebugPrint ((VOID *)RspHeader, RspHeaderSize, "KCS Response Header:\n");
+ }
+
+ if (ExpectedHeaderSize != RspHeaderSize) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "The size (%d bytes) of returned resposne header is not the same as expection (%d bytes)!\n",
+ RspHeaderSize,
+ ExpectedHeaderSize
+ ));
+ FreePool (RspHeader);
+ return EFI_DEVICE_ERROR;
+ }
+
+ if (CompareGuid (&gManageabilityProtocolMctpGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+ //
+ // MCTP over KCS
+ //
+ if (((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->NetFunc != MCTP_KCS_NETFN_LUN) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! MANAGEABILITY_MCTP_KCS_HEADER.NetFunc is equal 0x%02x instead of 0x%02x\n",
+ __func__,
+ ((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->NetFunc,
+ MCTP_KCS_NETFN_LUN
+ ));
+ FreePool (RspHeader);
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_ERROR;
+ return EFI_DEVICE_ERROR;
+ }
+
+ if (((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->DefiningBody != DEFINING_BODY_DMTF_PRE_OS_WORKING_GROUP) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! MANAGEABILITY_MCTP_KCS_HEADER.DefiningBody is equal 0x%02x instead of 0x%02x\n",
+ __func__,
+ ((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->DefiningBody,
+ DEFINING_BODY_DMTF_PRE_OS_WORKING_GROUP
+ ));
+ FreePool (RspHeader);
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_ERROR;
+ return EFI_DEVICE_ERROR;
+ }
+ }
+
+ *ResponseHeader = RspHeader;
+ *AdditionalStatus = MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS_NO_ERRORS;
+ return EFI_SUCCESS;
+}
+
/**
This service communicates with BMC using KCS protocol.
@@ -423,11 +638,9 @@ KcsTransportSendCommand (
OUT MANAGEABILITY_TRANSPORT_ADDITIONAL_STATUS *AdditionalStatus
)
{
- EFI_STATUS Status;
- UINT32 RspHeaderSize;
- IPMI_KCS_RESPONSE_HEADER RspHeader;
- UINT32 ExpectedResponseDataSize;
- CHAR16 *CompletionCodeStr;
+ EFI_STATUS Status;
+ UINT8 *RspHeader;
+ UINT32 ExpectedResponseDataSize;
if ((RequestData != NULL) && (RequestDataSize == 0)) {
DEBUG ((DEBUG_ERROR, "%a: Mismatched values of RequestData and RequestDataSize\n", __func__));
@@ -467,56 +680,59 @@ KcsTransportSendCommand (
RequestDataSize
);
if (EFI_ERROR (Status)) {
- DEBUG ((DEBUG_ERROR, "KCS Write Failed with Status(%r)", Status));
+ DEBUG ((DEBUG_ERROR, "KCS Write Failed with Status(%r)\n", Status));
return Status;
}
+ }
+ if ((ResponseData != NULL) && (ResponseDataSize != NULL) && (*ResponseDataSize != 0)) {
//
// Read the response header
- RspHeaderSize = sizeof (IPMI_KCS_RESPONSE_HEADER);
- Status = KcsTransportRead ((UINT8 *)&RspHeader, &RspHeaderSize);
+ //
+ Status = KcsReadResponseHeader (&RspHeader, AdditionalStatus);
if (EFI_ERROR (Status)) {
- DEBUG ((
- DEBUG_ERROR,
- "KCS read response header failed Status(%r), " \
- "RspNetFunctionLun = 0x%x, " \
- "Comamnd = 0x%x \n",
- Status,
- RspHeader.NetFunc,
- RspHeader.Command
- ));
return (Status);
}
//
- // Print out the response payloads.
- HelperManageabilityDebugPrint ((VOID *)&RspHeader, RspHeaderSize, "KCS Response Header:\n");
- }
+ // Override ResposeDataSize if the manageability protocol is MCTP.
+ //
+ if (CompareGuid (&gManageabilityProtocolMctpGuid, mSingleSessionToken->Token.ManageabilityProtocolSpecification)) {
+ if (*ResponseDataSize < ((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->ByteCount) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! MANAGEABILITY_MCTP_KCS_HEADER.ByteCount (0x%02x) is bigger than provided buffer (0x%02x)\n",
+ __func__,
+ ((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->ByteCount,
+ *ResponseDataSize
+ ));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ *ResponseDataSize = ((MANAGEABILITY_MCTP_KCS_HEADER *)RspHeader)->ByteCount;
+ }
+ FreePool (RspHeader);
- if ((ResponseData != NULL) && (ResponseDataSize != NULL) && (*ResponseDataSize != 0)) {
ExpectedResponseDataSize = *ResponseDataSize;
- Status = KcsTransportRead ((UINT8 *)ResponseData, ResponseDataSize);
+ Status = KcsTransportRead (ResponseData, ResponseDataSize);
if (EFI_ERROR (Status)) {
- DEBUG ((DEBUG_ERROR, "KCS response read Failed with Status(%r)", Status));
+ DEBUG ((DEBUG_ERROR, "KCS response read Failed with Status(%r)\n", Status));
}
- //
// Print out the response payloads.
if (*ResponseDataSize != 0) {
if (ExpectedResponseDataSize != *ResponseDataSize) {
- DEBUG ((DEBUG_ERROR, "Expected KCS response size : %d is not matched to returned size : %d.\n", ExpectedResponseDataSize, *ResponseDataSize));
- Status = EFI_DEVICE_ERROR;
+ DEBUG ((
+ DEBUG_ERROR,
+ "Expected KCS response size : %d is not matched to returned size : %d.\n",
+ ExpectedResponseDataSize,
+ *ResponseDataSize
+ ));
+ return EFI_DEVICE_ERROR;
}
HelperManageabilityDebugPrint ((VOID *)ResponseData, (UINT32)*ResponseDataSize, "KCS Response Data:\n");
-
- // Print Completion Code
- Status = IpmiHelperCheckCompletionCode (*((UINT8 *)ResponseData), &CompletionCodeStr, AdditionalStatus);
- if (!EFI_ERROR (Status)) {
- DEBUG ((DEBUG_MANAGEABILITY_INFO, "Cc: %02x %s.\n", *((UINT8 *)ResponseData), CompletionCodeStr));
- } else if (Status == EFI_NOT_FOUND) {
- DEBUG ((DEBUG_MANAGEABILITY_INFO, "Cc: %02x not defined in IpmiCompletionCodeMapping or invalid.\n", *((UINT8 *)ResponseData)));
- }
+ Status = KcsCheckResponseData (ResponseData, *ResponseDataSize, AdditionalStatus);
} else {
DEBUG ((DEBUG_ERROR, "No response, can't determine Completion Code.\n"));
}
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 7576007f77..e560c638d5 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -267,6 +267,9 @@ CommonMctpSubmitMessage (
MANAGEABILITY_TRANSPORT_TRAILER MctpTransportTrailer;
MANAGEABILITY_TRANSMISSION_MULTI_PACKAGES *MultiPackages;
MANAGEABILITY_TRANSMISSION_PACKAGE_ATTR *ThisPackage;
+ UINT8 *ResponseBuffer;
+ MCTP_TRANSPORT_HEADER *MctpTransportResponseHeader;
+ MCTP_MESSAGE_HEADER *MctpMessageResponseHeader;
if (TransportToken == NULL) {
DEBUG ((DEBUG_ERROR, "%a: No transport toke for MCTP\n", __func__));
@@ -435,11 +438,12 @@ CommonMctpSubmitMessage (
ThisPackage++;
}
+ ResponseBuffer = (UINT8 *)AllocatePool (*ResponseDataSize + sizeof (MCTP_TRANSPORT_HEADER) + sizeof (MCTP_MESSAGE_HEADER));
// Receive packet.
TransferToken.TransmitPackage.TransmitPayload = NULL;
TransferToken.TransmitPackage.TransmitSizeInByte = 0;
- TransferToken.ReceivePackage.ReceiveBuffer = ResponseData;
- TransferToken.ReceivePackage.ReceiveSizeInByte = *ResponseDataSize;
+ TransferToken.ReceivePackage.ReceiveBuffer = ResponseBuffer;
+ TransferToken.ReceivePackage.ReceiveSizeInByte = *ResponseDataSize + sizeof (MCTP_TRANSPORT_HEADER) + sizeof (MCTP_MESSAGE_HEADER);
TransferToken.TransmitHeader = NULL;
TransferToken.TransmitHeaderSize = 0;
TransferToken.TransmitTrailer = NULL;
@@ -461,8 +465,10 @@ CommonMctpSubmitMessage (
// Return transfer status.
//
*AdditionalTransferError = TransferToken.TransportAdditionalStatus;
- *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte;
- Status = TransferToken.TransferStatus;
+ *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte - sizeof (MCTP_TRANSPORT_HEADER) - sizeof (MCTP_MESSAGE_HEADER);
+ CopyMem (ResponseData, ResponseBuffer + sizeof (MCTP_TRANSPORT_HEADER) + sizeof (MCTP_MESSAGE_HEADER), *ResponseDataSize);
+ FreePool (ResponseBuffer);
+ Status = TransferToken.TransferStatus;
if (EFI_ERROR (Status)) {
DEBUG ((DEBUG_ERROR, "%a: Failed to send MCTP command over %s: %r\n", __func__, mTransportName, Status));
return Status;
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109911): https://edk2.groups.io/g/devel/message/109911
Mute This Topic: https://groups.io/mt/102134651/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 04/16] ManageabilityPkg: Check header fields in the MCTP response
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (2 preceding siblings ...)
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 03/16] ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library Konstantin Aladyshev
@ 2023-10-23 13:04 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 05/16] ManageabilityPkg: Correct typo in MCTP destination EID field Konstantin Aladyshev
` (12 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:04 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Add checks for the MCTP header fields in the MCTP response.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../MctpProtocol/Common/MctpProtocolCommon.c | 82 +++++++++++++++++++
1 file changed, 82 insertions(+)
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index e560c638d5..5844d54eb2 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -461,6 +461,88 @@ CommonMctpSubmitMessage (
&TransferToken
);
+ MctpTransportResponseHeader = (MCTP_TRANSPORT_HEADER *)ResponseBuffer;
+ if (MctpTransportResponseHeader->Bits.HeaderVersion != MCTP_KCS_HEADER_VERSION) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response HeaderVersion (0x%02x) doesn't match MCTP_KCS_HEADER_VERSION (0x%02x)\n",
+ __func__,
+ MctpTransportResponseHeader->Bits.HeaderVersion,
+ MCTP_KCS_HEADER_VERSION
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+ if (MctpTransportResponseHeader->Bits.MessageTag != MCTP_MESSAGE_TAG) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response MessageTag (0x%02x) doesn't match MCTP_MESSAGE_TAG (0x%02x)\n",
+ __func__,
+ MctpTransportResponseHeader->Bits.MessageTag,
+ MCTP_MESSAGE_TAG
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+ if (MctpTransportResponseHeader->Bits.TagOwner != MCTP_MESSAGE_TAG_OWNER_RESPONSE) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response TagOwner (0x%02x) doesn't match MCTP_MESSAGE_TAG_OWNER_RESPONSE (0x%02x)\n",
+ __func__,
+ MctpTransportResponseHeader->Bits.TagOwner,
+ MCTP_MESSAGE_TAG_OWNER_RESPONSE
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+ if (MctpTransportResponseHeader->Bits.SourceEndpointId != MctpDestinationEndpointId) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response SrcEID (0x%02x) doesn't match sent EID (0x%02x)\n",
+ __func__,
+ MctpTransportResponseHeader->Bits.SourceEndpointId,
+ MctpDestinationEndpointId
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+ if (MctpTransportResponseHeader->Bits.DestinationEndpointId != MctpSourceEndpointId) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response DestEID (0x%02x) doesn't match local EID (0x%02x)\n",
+ __func__,
+ MctpTransportResponseHeader->Bits.DestinationEndpointId,
+ MctpSourceEndpointId
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+
+ MctpMessageResponseHeader = (MCTP_MESSAGE_HEADER *)(MctpTransportResponseHeader + 1);
+ if (MctpMessageResponseHeader->Bits.MessageType != MctpType) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response MessageType (0x%02x) doesn't match sent MessageType (0x%02x)\n",
+ __func__,
+ MctpMessageResponseHeader->Bits.MessageType,
+ MctpType
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+
+ if (MctpMessageResponseHeader->Bits.IntegrityCheck != (UINT8)RequestDataIntegrityCheck) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Response IntegrityCheck (%d) doesn't match sent IntegrityCheck (%d)\n",
+ __func__,
+ MctpMessageResponseHeader->Bits.IntegrityCheck,
+ (UINT8)RequestDataIntegrityCheck
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_DEVICE_ERROR;
+ }
+
//
// Return transfer status.
//
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109912): https://edk2.groups.io/g/devel/message/109912
Mute This Topic: https://groups.io/mt/102134652/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 05/16] ManageabilityPkg: Correct typo in MCTP destination EID field
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (3 preceding siblings ...)
2023-10-23 13:04 ` [edk2-devel] [PATCH edk2-platforms v3 04/16] ManageabilityPkg: Check header fields in the MCTP response Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 06/16] ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD Konstantin Aladyshev
` (11 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Correct wrong structure member used for MCTP destination EID.
Signed-off-by: Abner Chang <abner.chang@amd.com>
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Universal/PldmProtocol/Common/PldmProtocolCommon.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index ce1e2cba95..1c4506d87f 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -119,7 +119,7 @@ SetupPldmRequestTransportPacket (
}
MctpHeader->SourceEndpointId = PcdGet8 (PcdMctpSourceEndpointId);
- MctpHeader->SourceEndpointId = PcdGet8 (PcdMctpDestinationEndpointId);
+ MctpHeader->DestinationEndpointId = PcdGet8 (PcdMctpDestinationEndpointId);
MctpHeader->MessageHeader.IntegrityCheck = FALSE;
MctpHeader->MessageHeader.MessageType = MCTP_MESSAGE_TYPE_PLDM;
*PacketHeader = (MANAGEABILITY_TRANSPORT_HEADER *)MctpHeader;
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109913): https://edk2.groups.io/g/devel/message/109913
Mute This Topic: https://groups.io/mt/102134653/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 06/16] ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (4 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 05/16] ManageabilityPkg: Correct typo in MCTP destination EID field Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 07/16] ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit Konstantin Aladyshev
` (10 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Although MtcpSubmit function receives source and destination MCTP EID
arguments these value are not used in any way currently. Instead the
code always uses EID values from the PCDs.
To correct this issue modify function interface to receive source and
destination MCTP EIDs via pointers and use PCD values only if the
pointers are NULL.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
Signed-off-by: Abner Chang <abner.chang@amd.com>
---
.../Include/Protocol/MctpProtocol.h | 12 ++++--
.../Dxe/ManageabilityTransportMctp.c | 4 +-
.../MctpProtocol/Common/MctpProtocolCommon.c | 4 +-
.../Universal/MctpProtocol/Dxe/MctpProtocol.c | 42 ++++++++++++++-----
4 files changed, 44 insertions(+), 18 deletions(-)
diff --git a/Features/ManageabilityPkg/Include/Protocol/MctpProtocol.h b/Features/ManageabilityPkg/Include/Protocol/MctpProtocol.h
index 85e42f157d..c96b986c44 100644
--- a/Features/ManageabilityPkg/Include/Protocol/MctpProtocol.h
+++ b/Features/ManageabilityPkg/Include/Protocol/MctpProtocol.h
@@ -28,8 +28,12 @@ typedef struct _EDKII_MCTP_PROTOCOL EDKII_MCTP_PROTOCOL;
@param[in] This EDKII_MCTP_PROTOCOL instance.
@param[in] MctpType MCTP message type.
- @param[in] MctpSourceEndpointId MCTP source endpoint ID.
- @param[in] MctpDestinationEndpointId MCTP source endpoint ID.
+ @param[in] MctpSourceEndpointId Pointer of MCTP source endpoint ID.
+ Set to NULL means use platform PCD value
+ (PcdMctpSourceEndpointId).
+ @param[in] MctpDestinationEndpointId Pointer of MCTP destination endpoint ID.
+ Set to NULL means use platform PCD value
+ (PcdMctpDestinationEndpointId).
@param[in] RequestDataIntegrityCheck Indicates whether MCTP message has
integrity check byte.
@param[in] RequestData Message Data.
@@ -58,8 +62,8 @@ EFI_STATUS
(EFIAPI *MCTP_SUBMIT_COMMAND)(
IN EDKII_MCTP_PROTOCOL *This,
IN UINT8 MctpType,
- IN UINT8 MctpSourceEndpointId,
- IN UINT8 MctpDestinationEndpointId,
+ IN UINT8 *MctpSourceEndpointId,
+ IN UINT8 *MctpDestinationEndpointId,
IN BOOLEAN RequestDataIntegrityCheck,
IN UINT8 *RequestData,
IN UINT32 RequestDataSize,
diff --git a/Features/ManageabilityPkg/Library/ManageabilityTransportMctpLib/Dxe/ManageabilityTransportMctp.c b/Features/ManageabilityPkg/Library/ManageabilityTransportMctpLib/Dxe/ManageabilityTransportMctp.c
index c520e2302d..249104c873 100644
--- a/Features/ManageabilityPkg/Library/ManageabilityTransportMctpLib/Dxe/ManageabilityTransportMctp.c
+++ b/Features/ManageabilityPkg/Library/ManageabilityTransportMctpLib/Dxe/ManageabilityTransportMctp.c
@@ -205,8 +205,8 @@ MctpTransportTransmitReceive (
Status = mMctpProtocol->Functions.Version1_0->MctpSubmitCommand (
mMctpProtocol,
TransmitHeader->MessageHeader.MessageType,
- TransmitHeader->SourceEndpointId,
- TransmitHeader->DestinationEndpointId,
+ &TransmitHeader->SourceEndpointId,
+ &TransmitHeader->DestinationEndpointId,
(BOOLEAN)TransmitHeader->MessageHeader.IntegrityCheck,
TransferToken->TransmitPackage.TransmitPayload,
TransferToken->TransmitPackage.TransmitSizeInByte,
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 5844d54eb2..3128aadd15 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -175,8 +175,8 @@ SetupMctpRequestTransportPacket (
MctpTransportHeader = (MCTP_TRANSPORT_HEADER *)ThisPackage;
MctpTransportHeader->Bits.Reserved = 0;
MctpTransportHeader->Bits.HeaderVersion = MCTP_KCS_HEADER_VERSION;
- MctpTransportHeader->Bits.DestinationEndpointId = PcdGet8 (PcdMctpDestinationEndpointId);
- MctpTransportHeader->Bits.SourceEndpointIdId = PcdGet8 (PcdMctpSourceEndpointId);
+ MctpTransportHeader->Bits.DestinationEndpointId = MctpDestinationEndpointId;
+ MctpTransportHeader->Bits.SourceEndpointId = MctpSourceEndpointId;
MctpTransportHeader->Bits.MessageTag = MCTP_MESSAGE_TAG;
MctpTransportHeader->Bits.TagOwner = MCTP_MESSAGE_TAG_OWNER_REQUEST;
MctpTransportHeader->Bits.PacketSequence = mMctpPacketSequence & MCTP_PACKET_SEQUENCE_MASK;
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
index d0f49a1abb..73445bf816 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Dxe/MctpProtocol.c
@@ -29,8 +29,12 @@ UINT32 mTransportMaximumPayload;
@param[in] This EDKII_MCTP_PROTOCOL instance.
@param[in] MctpType MCTP message type.
- @param[in] MctpSourceEndpointId MCTP source endpoint ID.
- @param[in] MctpDestinationEndpointId MCTP source endpoint ID.
+ @param[in] MctpSourceEndpointId Pointer of MCTP source endpoint ID.
+ Set to NULL means use platform PCD value
+ (PcdMctpSourceEndpointId).
+ @param[in] MctpDestinationEndpointId Pointer of MCTP destination endpoint ID.
+ Set to NULL means use platform PCD value
+ (PcdMctpDestinationEndpointId).
@param[in] RequestDataIntegrityCheck Indicates whether MCTP message has
integrity check byte.
@param[in] RequestData Message Data.
@@ -59,8 +63,8 @@ EFIAPI
MctpSubmitMessage (
IN EDKII_MCTP_PROTOCOL *This,
IN UINT8 MctpType,
- IN UINT8 MctpSourceEndpointId,
- IN UINT8 MctpDestinationEndpointId,
+ IN UINT8 *MctpSourceEndpointId,
+ IN UINT8 *MctpDestinationEndpointId,
IN BOOLEAN RequestDataIntegrityCheck,
IN UINT8 *RequestData,
IN UINT32 RequestDataSize,
@@ -72,24 +76,42 @@ MctpSubmitMessage (
)
{
EFI_STATUS Status;
+ UINT8 SourceEid;
+ UINT8 DestinationEid;
if ((RequestData == NULL) && (ResponseData == NULL)) {
DEBUG ((DEBUG_ERROR, "%a: Both RequestData and ResponseData are NULL\n", __func__));
return EFI_INVALID_PARAMETER;
}
+ if (MctpSourceEndpointId == NULL) {
+ SourceEid = PcdGet8 (PcdMctpSourceEndpointId);
+ DEBUG ((DEBUG_MANAGEABILITY, "%a: Use PcdMctpSourceEndpointId for MCTP source EID: %x\n", __func__, SourceEid));
+ } else {
+ SourceEid = *MctpSourceEndpointId;
+ DEBUG ((DEBUG_MANAGEABILITY, "%a: MCTP source EID: %x\n", __func__, SourceEid));
+ }
+
+ if (MctpDestinationEndpointId == NULL) {
+ DestinationEid = PcdGet8 (PcdMctpDestinationEndpointId);
+ DEBUG ((DEBUG_MANAGEABILITY, "%a: Use PcdMctpDestinationEndpointId for MCTP destination EID: %x\n", __func__, DestinationEid));
+ } else {
+ DestinationEid = *MctpDestinationEndpointId;
+ DEBUG ((DEBUG_MANAGEABILITY, "%a: MCTP destination EID: %x\n", __func__, DestinationEid));
+ }
+
//
// Check source EID and destination EID
//
- if ((MctpSourceEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
- (MctpSourceEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ if ((SourceEid >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (SourceEid <= MCTP_RESERVED_ENDPOINT_END_ID)
) {
DEBUG ((DEBUG_ERROR, "%a: The value of MCTP source EID (%x) is reserved.\n", __func__, MctpSourceEndpointId));
return EFI_INVALID_PARAMETER;
}
- if ((MctpDestinationEndpointId >= MCTP_RESERVED_ENDPOINT_START_ID) &&
- (MctpDestinationEndpointId <= MCTP_RESERVED_ENDPOINT_END_ID)
+ if ((DestinationEid >= MCTP_RESERVED_ENDPOINT_START_ID) &&
+ (DestinationEid <= MCTP_RESERVED_ENDPOINT_END_ID)
) {
DEBUG ((DEBUG_ERROR, "%a: The value of MCTP destination EID (%x) is reserved.\n", __func__, MctpDestinationEndpointId));
return EFI_INVALID_PARAMETER;
@@ -98,8 +120,8 @@ MctpSubmitMessage (
Status = CommonMctpSubmitMessage (
mTransportToken,
MctpType,
- MctpSourceEndpointId,
- MctpDestinationEndpointId,
+ SourceEid,
+ DestinationEid,
RequestDataIntegrityCheck,
RequestData,
RequestDataSize,
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109914): https://edk2.groups.io/g/devel/message/109914
Mute This Topic: https://groups.io/mt/102134654/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 07/16] ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (5 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 06/16] ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 08/16] ManageabilityPkg: Don't check MCTP header fields if transfer has failed Konstantin Aladyshev
` (9 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Currently the MCTP TAG_OWNER bit is checked against 1 both in MTCP
request and response.
According to the MTCP Base specification in case of the MCTP response
the TAG_OWNER bit should be equal to 0.
Correct MCTP_MESSAGE_TAG_OWNER_RESPONSE flag value to fix the issue.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Include/Library/ManageabilityTransportMctpLib.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
index 462e7436e6..a8dc8a8519 100644
--- a/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
+++ b/Features/ManageabilityPkg/Include/Library/ManageabilityTransportMctpLib.h
@@ -51,8 +51,8 @@ typedef struct {
// is not defined by the specification.
#define MCTP_MESSAGE_TAG 0x1
-#define MCTP_MESSAGE_TAG_OWNER_REQUEST 0x01
-#define MCTP_MESSAGE_TAG_OWNER_RESPONSE 0x01
+#define MCTP_MESSAGE_TAG_OWNER_REQUEST 1
+#define MCTP_MESSAGE_TAG_OWNER_RESPONSE 0
#define MCTP_PACKET_SEQUENCE_MASK 0x3
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109915): https://edk2.groups.io/g/devel/message/109915
Mute This Topic: https://groups.io/mt/102134655/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 08/16] ManageabilityPkg: Don't check MCTP header fields if transfer has failed
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (6 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 07/16] ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 09/16] ManageabilityPkg: Use correct constants for PLDM header checks Konstantin Aladyshev
` (8 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
If MCTP KCS communication has failed we need to abort MCTP transfer
function before checking any MCTP header data.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../MctpProtocol/Common/MctpProtocolCommon.c | 16 +++++++---------
1 file changed, 7 insertions(+), 9 deletions(-)
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 3128aadd15..4aae4fcba9 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -461,6 +461,13 @@ CommonMctpSubmitMessage (
&TransferToken
);
+ *AdditionalTransferError = TransferToken.TransportAdditionalStatus;
+ Status = TransferToken.TransferStatus;
+ if (EFI_ERROR (Status)) {
+ DEBUG ((DEBUG_ERROR, "%a: Failed to send MCTP command over %s: %r\n", __func__, mTransportName, Status));
+ return Status;
+ }
+
MctpTransportResponseHeader = (MCTP_TRANSPORT_HEADER *)ResponseBuffer;
if (MctpTransportResponseHeader->Bits.HeaderVersion != MCTP_KCS_HEADER_VERSION) {
DEBUG ((
@@ -543,18 +550,9 @@ CommonMctpSubmitMessage (
return EFI_DEVICE_ERROR;
}
- //
- // Return transfer status.
- //
- *AdditionalTransferError = TransferToken.TransportAdditionalStatus;
*ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte - sizeof (MCTP_TRANSPORT_HEADER) - sizeof (MCTP_MESSAGE_HEADER);
CopyMem (ResponseData, ResponseBuffer + sizeof (MCTP_TRANSPORT_HEADER) + sizeof (MCTP_MESSAGE_HEADER), *ResponseDataSize);
FreePool (ResponseBuffer);
- Status = TransferToken.TransferStatus;
- if (EFI_ERROR (Status)) {
- DEBUG ((DEBUG_ERROR, "%a: Failed to send MCTP command over %s: %r\n", __func__, mTransportName, Status));
- return Status;
- }
return Status;
}
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109916): https://edk2.groups.io/g/devel/message/109916
Mute This Topic: https://groups.io/mt/102134656/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 09/16] ManageabilityPkg: Use correct constants for PLDM header checks
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (7 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 08/16] ManageabilityPkg: Don't check MCTP header fields if transfer has failed Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 10/16] ManageabilityPkg: Return error on multiple-packet MCTP responses Konstantin Aladyshev
` (7 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Currently PldmProtocol code uses magic numbers in the PLDM header
checks. Since PLDM headers have all the necessary definitions replace
magic numbers with the appropriate defines.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Universal/PldmProtocol/Common/PldmProtocolCommon.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index 1c4506d87f..4edfe05955 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -321,15 +321,15 @@ CommonPldmSubmitCommand (
//
// Check the integrity of response. data.
ResponseHeader = (PLDM_RESPONSE_HEADER *)FullPacketResponseData;
- if ((ResponseHeader->PldmHeader.DatagramBit != 0) ||
- (ResponseHeader->PldmHeader.RequestBit != 0) ||
+ if ((ResponseHeader->PldmHeader.DatagramBit != (!PLDM_MESSAGE_HEADER_IS_DATAGRAM)) ||
+ (ResponseHeader->PldmHeader.RequestBit != PLDM_MESSAGE_HEADER_IS_RESPONSE) ||
(ResponseHeader->PldmHeader.InstanceId != mPldmRequestInstanceId) ||
(ResponseHeader->PldmHeader.PldmType != PldmType) ||
(ResponseHeader->PldmHeader.PldmTypeCommandCode != PldmCommand))
{
DEBUG ((DEBUG_ERROR, "PLDM integrity check of response data is failed.\n"));
- DEBUG ((DEBUG_ERROR, " Request bit = %d (Expected value: 0)\n"));
- DEBUG ((DEBUG_ERROR, " Datagram = %d (Expected value: 0)\n"));
+ DEBUG ((DEBUG_ERROR, " Datagram = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.DatagramBit, (!PLDM_MESSAGE_HEADER_IS_DATAGRAM)));
+ DEBUG ((DEBUG_ERROR, " Request bit = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.RequestBit, PLDM_MESSAGE_HEADER_IS_RESPONSE));
DEBUG ((DEBUG_ERROR, " Instance ID = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.InstanceId, mPldmRequestInstanceId));
DEBUG ((DEBUG_ERROR, " Pldm Type = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.PldmType, PldmType));
DEBUG ((DEBUG_ERROR, " Pldm Command = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.PldmTypeCommandCode, PldmCommand));
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109917): https://edk2.groups.io/g/devel/message/109917
Mute This Topic: https://groups.io/mt/102134657/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 10/16] ManageabilityPkg: Return error on multiple-packet MCTP responses
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (8 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 09/16] ManageabilityPkg: Use correct constants for PLDM header checks Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 11/16] ManageabilityPkg: Add PLDM terminus PCDs Konstantin Aladyshev
` (6 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Since the current driver doesn't yet support handling of
multiple-packet MCTP responses, return EFI_UNSUPPORTED error in such
cases.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../MctpProtocol/Common/MctpProtocolCommon.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
index 4aae4fcba9..3709ab16eb 100644
--- a/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/MctpProtocol/Common/MctpProtocolCommon.c
@@ -524,6 +524,17 @@ CommonMctpSubmitMessage (
FreePool (ResponseBuffer);
return EFI_DEVICE_ERROR;
}
+ if ((MctpTransportResponseHeader->Bits.StartOfMessage != 1) ||
+ (MctpTransportResponseHeader->Bits.EndOfMessage != 1) ||
+ (MctpTransportResponseHeader->Bits.PacketSequence != 0)) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: Error! Multiple-packet MCTP responses are not supported by the current driver\n",
+ __func__
+ ));
+ FreePool (ResponseBuffer);
+ return EFI_UNSUPPORTED;
+ }
MctpMessageResponseHeader = (MCTP_MESSAGE_HEADER *)(MctpTransportResponseHeader + 1);
if (MctpMessageResponseHeader->Bits.MessageType != MctpType) {
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109918): https://edk2.groups.io/g/devel/message/109918
Mute This Topic: https://groups.io/mt/102134658/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 11/16] ManageabilityPkg: Add PLDM terminus PCDs
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (9 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 10/16] ManageabilityPkg: Return error on multiple-packet MCTP responses Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 12/16] PldmProtocolDxe: Correct TID argument usage Konstantin Aladyshev
` (5 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
From: Abner Chang <abner.chang@amd.com>
Add PLDM source and destination terminus IDs for transmiting PLDM
message.
Signed-off-by: Abner Chang <abner.chang@amd.com>
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
Features/ManageabilityPkg/ManageabilityPkg.dec | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/Features/ManageabilityPkg/ManageabilityPkg.dec b/Features/ManageabilityPkg/ManageabilityPkg.dec
index 14fe0fd2e0..eb0ee67cba 100644
--- a/Features/ManageabilityPkg/ManageabilityPkg.dec
+++ b/Features/ManageabilityPkg/ManageabilityPkg.dec
@@ -72,6 +72,12 @@
# @Prompt MCTP KCS (Memory mapped) I/O base address
gManageabilityPkgTokenSpaceGuid.PcdMctpKcsBaseAddress|0xca2|UINT32|0x00000004
+ ## This value is the PLDM source and destination terminus ID for transmiting PLDM message.
+ # @Prompt PLDM source terminus ID
+ gManageabilityPkgTokenSpaceGuid.PcdPldmSourceTerminusId|0|UINT8|0x00000040
+ # @Prompt PLDM destination terminus ID
+ gManageabilityPkgTokenSpaceGuid.PcdPldmDestinationEndpointId|0|UINT8|0x00000041
+
## This is the value of SOL channels supported on platform.
# @Prompt SOL channel number
gManageabilityPkgTokenSpaceGuid.PcdMaxSolChannels|3|UINT8|0x00000100
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109919): https://edk2.groups.io/g/devel/message/109919
Mute This Topic: https://groups.io/mt/102134661/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 12/16] PldmProtocolDxe: Correct TID argument usage
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (10 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 11/16] ManageabilityPkg: Add PLDM terminus PCDs Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 13/16] ManageabilityPkg/PldmProtocol: Remove PLDM command table Konstantin Aladyshev
` (4 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
From: Abner Chang <abner.chang@amd.com>
Currently the PLDM source/destination TID arguments for the PldmSubmit
function are not actually used in any way in the underlying MCTP
communication. The code just uses MCTP source/destination EID PCDs. So
we have to restructure code to actually use provided PLDM TIDs.
On the other case the PldmSubmitCommand function from the
PldmProtocolLib doesn't even accept the source/destination TID
arguments.
To address both these facts correct TID argument usage in the following
way:
- by default the TID values are taken from the built-time PCDs,
- user have an ability to provide custom TIDs either via PldmSubmit
function arguments or by calling PldmSetTerminus API.
Signed-off-by: Abner Chang <abner.chang@amd.com>
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Include/Library/BasePldmProtocolLib.h | 16 ++++++
.../Include/Protocol/PldmProtocol.h | 18 +++---
.../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 +++++++++++++++-
.../Dxe/PldmProtocolLib.inf | 6 +-
.../PldmProtocol/Common/PldmProtocolCommon.c | 28 +++++++---
.../PldmProtocol/Common/PldmProtocolCommon.h | 22 +++++---
.../Universal/PldmProtocol/Dxe/PldmProtocol.c | 56 ++++++++++++++++---
.../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 --
8 files changed, 162 insertions(+), 37 deletions(-)
diff --git a/Features/ManageabilityPkg/Include/Library/BasePldmProtocolLib.h b/Features/ManageabilityPkg/Include/Library/BasePldmProtocolLib.h
index 5523ac3a4d..a698197263 100644
--- a/Features/ManageabilityPkg/Include/Library/BasePldmProtocolLib.h
+++ b/Features/ManageabilityPkg/Include/Library/BasePldmProtocolLib.h
@@ -9,6 +9,22 @@
#ifndef EDKII_PLDM_PROTOCOL_LIB_H_
#define EDKII_PLDM_PROTOCOL_LIB_H_
+/**
+ This function sets the PLDM source termius and destination terminus
+ ID for SMBIOS PLDM transfer.
+
+ @param[in] SourceId PLDM source teminus ID.
+ @param[in] DestinationId PLDM destination teminus ID.
+
+ @retval EFI_SUCCESS The terminus is set successfully.
+ @retval EFI_INVALID_PARAMETER The terminus is set unsuccessfully.
+**/
+EFI_STATUS
+PldmSetTerminus (
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId
+);
+
/**
This service enables submitting commands via EDKII PLDM protocol.
diff --git a/Features/ManageabilityPkg/Include/Protocol/PldmProtocol.h b/Features/ManageabilityPkg/Include/Protocol/PldmProtocol.h
index 651997e1ad..02efb3015a 100644
--- a/Features/ManageabilityPkg/Include/Protocol/PldmProtocol.h
+++ b/Features/ManageabilityPkg/Include/Protocol/PldmProtocol.h
@@ -26,13 +26,15 @@ typedef struct _EDKII_PLDM_PROTOCOL EDKII_PLDM_PROTOCOL;
/**
This service enables submitting commands via EDKII PLDM protocol.
- @param[in] This EDKII_PLDM_PROTOCOL instance.
- @param[in] PldmType PLDM message type.
- @param[in] Command PLDM Command of PLDM message type.
- @param[in] RequestData Command Request Data.
- @param[in] RequestDataSize Size of Command Request Data.
- @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
- @param[in, out] ResponseDataSize Size of Command Response Data.
+ @param[in] This EDKII_PLDM_PROTOCOL instance.
+ @param[in] PldmType PLDM message type.
+ @param[in] Command PLDM Command of PLDM message type.
+ @param[in] PldmTerminusSourceId PLDM source teminus ID.
+ @param[in] PldmTerminusDestinationId PLDM destination teminus ID.
+ @param[in] RequestData Command Request Data.
+ @param[in] RequestDataSize Size of Command Request Data.
+ @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
+ @param[in, out] ResponseDataSize Size of Command Response Data.
@retval EFI_SUCCESS The command byte stream was successfully submit to the device and a response was successfully received.
@retval EFI_NOT_FOUND The command was not successfully sent to the device or a response was not successfully received from the device.
@@ -49,6 +51,8 @@ EFI_STATUS
IN EDKII_PLDM_PROTOCOL *This,
IN UINT8 PldmType,
IN UINT8 Command,
+ IN UINT8 PldmTerminusSourceId,
+ IN UINT8 PldmTerminusDestinationId,
IN UINT8 *RequestData,
IN UINT32 RequestDataSize,
OUT UINT8 *ResponseData,
diff --git a/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.c b/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.c
index 267bd8fbc1..37231b0756 100644
--- a/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.c
+++ b/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.c
@@ -9,10 +9,34 @@
#include <PiDxe.h>
#include <Protocol/PldmProtocol.h>
#include <Library/DebugLib.h>
+#include <Library/PcdLib.h>
#include <Library/ManageabilityTransportHelperLib.h>
#include <Library/UefiBootServicesTableLib.h>
-EDKII_PLDM_PROTOCOL *mEdkiiPldmProtocol = NULL;
+EDKII_PLDM_PROTOCOL *mEdkiiPldmProtocol = NULL;
+UINT8 mSourcePldmTerminusId = 0;
+UINT8 mDestinationPldmTerminusId = 0;
+
+/**
+ This function sets the PLDM source termius and destination terminus
+ ID for SMBIOS PLDM transfer.
+
+ @param[in] SourceId PLDM source teminus ID.
+ @param[in] DestinationId PLDM destination teminus ID.
+
+ @retval EFI_SUCCESS The terminus is set successfully.
+ @retval EFI_INVALID_PARAMETER The terminus is set unsuccessfully.
+**/
+EFI_STATUS
+PldmSetTerminus (
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId
+)
+{
+ mSourcePldmTerminusId = SourceId;
+ mDestinationPldmTerminusId = DestinationId;
+ return EFI_SUCCESS;
+}
/**
This service enables submitting commands via EDKII PLDM protocol.
@@ -69,6 +93,8 @@ PldmSubmitCommand (
mEdkiiPldmProtocol,
PldmType,
Command,
+ mSourcePldmTerminusId,
+ mDestinationPldmTerminusId,
RequestData,
RequestDataSize,
ResponseData,
@@ -85,3 +111,24 @@ PldmSubmitCommand (
return Status;
}
+/**
+
+ Initialize mSourcePldmTerminusId and mDestinationPldmTerminusId.
+
+ @param ImageHandle The image handle.
+ @param SystemTable The system table.
+
+ @retval EFI_SUCCESS Protocol listener is registered successfully.
+
+**/
+EFI_STATUS
+EFIAPI
+PldmProtocolLibConstructor (
+ IN EFI_HANDLE ImageHandle,
+ IN EFI_SYSTEM_TABLE *SystemTable
+ )
+{
+
+ PldmSetTerminus (PcdGet8(PcdPldmSourceTerminusId), PcdGet8(PcdPldmDestinationEndpointId));
+ return EFI_SUCCESS;
+}
diff --git a/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.inf b/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.inf
index 1233d76726..19c84840b6 100644
--- a/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.inf
+++ b/Features/ManageabilityPkg/Library/PldmProtocolLibrary/Dxe/PldmProtocolLib.inf
@@ -16,7 +16,7 @@
MODULE_TYPE = DXE_DRIVER
VERSION_STRING = 1.0
LIBRARY_CLASS = PldmProtocolLib|DXE_RUNTIME_DRIVER DXE_DRIVER DXE_CORE UEFI_DRIVER UEFI_APPLICATION
-
+ CONSTRUCTOR = PldmProtocolLibConstructor
#
# VALID_ARCHITECTURES = IA32 X64
#
@@ -40,3 +40,7 @@
[Protocols]
gEdkiiPldmProtocolGuid ## ALWAYS_CONSUMES
+[FixedPcd]
+ gManageabilityPkgTokenSpaceGuid.PcdPldmSourceTerminusId
+ gManageabilityPkgTokenSpaceGuid.PcdPldmDestinationEndpointId
+
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index 4edfe05955..ea3d4a22b2 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -64,6 +64,8 @@ GetFullPacketResponseSize (
@param[in] TransportToken The transport interface.
@param[in] PldmType PLDM message type.
@param[in] PldmCommand PLDM command of this PLDM type.
+ @param[in] SourceId PLDM source teminus ID.
+ @param[in] DestinationId PLDM destination teminus ID.
@param[out] PacketHeader The pointer to receive header of request.
@param[out] PacketHeaderSize Packet header size in bytes.
@param[in, out] PacketBody The request body.
@@ -88,6 +90,8 @@ SetupPldmRequestTransportPacket (
IN MANAGEABILITY_TRANSPORT_TOKEN *TransportToken,
IN UINT8 PldmType,
IN UINT8 PldmCommand,
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId,
OUT MANAGEABILITY_TRANSPORT_HEADER *PacketHeader,
OUT UINT16 *PacketHeaderSize,
IN OUT UINT8 **PacketBody,
@@ -118,8 +122,8 @@ SetupPldmRequestTransportPacket (
return EFI_OUT_OF_RESOURCES;
}
- MctpHeader->SourceEndpointId = PcdGet8 (PcdMctpSourceEndpointId);
- MctpHeader->DestinationEndpointId = PcdGet8 (PcdMctpDestinationEndpointId);
+ MctpHeader->SourceEndpointId = SourceId;
+ MctpHeader->DestinationEndpointId = DestinationId;
MctpHeader->MessageHeader.IntegrityCheck = FALSE;
MctpHeader->MessageHeader.MessageType = MCTP_MESSAGE_TYPE_PLDM;
*PacketHeader = (MANAGEABILITY_TRANSPORT_HEADER *)MctpHeader;
@@ -161,13 +165,15 @@ SetupPldmRequestTransportPacket (
/**
Common code to submit PLDM commands
- @param[in] TransportToken Transport token.
- @param[in] PldmType PLDM message type.
- @param[in] PldmCommand PLDM command of this PLDM type.
- @param[in] RequestData Command Request Data.
- @param[in] RequestDataSize Size of Command Request Data.
- @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
- @param[in, out] ResponseDataSize Size of Command Response Data.
+ @param[in] TransportToken Transport token.
+ @param[in] PldmType PLDM message type.
+ @param[in] PldmCommand PLDM command of this PLDM type.
+ @param[in] PldmTerminusSourceId PLDM source teminus ID.
+ @param[in] PldmTerminusDestinationId PLDM destination teminus ID.
+ @param[in] RequestData Command Request Data.
+ @param[in] RequestDataSize Size of Command Request Data.
+ @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
+ @param[in, out] ResponseDataSize Size of Command Response Data.
@retval EFI_SUCCESS The command byte stream was successfully submit to the device and a response was successfully received.
@retval EFI_NOT_FOUND The command was not successfully sent to the device or a response was not successfully received from the device.
@@ -182,6 +188,8 @@ CommonPldmSubmitCommand (
IN MANAGEABILITY_TRANSPORT_TOKEN *TransportToken,
IN UINT8 PldmType,
IN UINT8 PldmCommand,
+ IN UINT8 PldmTerminusSourceId,
+ IN UINT8 PldmTerminusDestinationId,
IN UINT8 *RequestData OPTIONAL,
IN UINT32 RequestDataSize,
OUT UINT8 *ResponseData OPTIONAL,
@@ -225,6 +233,8 @@ CommonPldmSubmitCommand (
TransportToken,
PldmType,
PldmCommand,
+ PldmTerminusSourceId,
+ PldmTerminusDestinationId,
&PldmTransportHeader,
&HeaderSize,
&ThisRequestData,
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
index 231d6e802e..79431dd3b1 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
@@ -44,6 +44,8 @@ SetupPldmTransportHardwareInformation (
@param[in] TransportToken The transport interface.
@param[in] PldmType PLDM message type.
@param[in] PldmCommand PLDM command of this PLDM type.
+ @param[in] SourceId PLDM source teminus ID.
+ @param[in] DestinationId PLDM destination teminus ID.
@param[out] PacketHeader The pointer to receive header of request.
@param[out] PacketHeaderSize Packet header size in bytes.
@param[in, out] PacketBody The request body.
@@ -68,6 +70,8 @@ SetupPldmRequestTransportPacket (
IN MANAGEABILITY_TRANSPORT_TOKEN *TransportToken,
IN UINT8 PldmType,
IN UINT8 PldmCommand,
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId,
OUT MANAGEABILITY_TRANSPORT_HEADER *PacketHeader,
OUT UINT16 *PacketHeaderSize,
IN OUT UINT8 **PacketBody,
@@ -79,13 +83,15 @@ SetupPldmRequestTransportPacket (
/**
Common code to submit PLDM commands
- @param[in] TransportToken Transport token.
- @param[in] PldmType PLDM message type.
- @param[in] PldmCommand PLDM command of this PLDM type.
- @param[in] RequestData Command Request Data.
- @param[in] RequestDataSize Size of Command Request Data.
- @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
- @param[in, out] ResponseDataSize Size of Command Response Data.
+ @param[in] TransportToken Transport token.
+ @param[in] PldmType PLDM message type.
+ @param[in] PldmCommand PLDM command of this PLDM type.
+ @param[in] PldmTerminusSourceId PLDM source teminus ID.
+ @param[in] PldmTerminusDestinationId PLDM destination teminus ID.
+ @param[in] RequestData Command Request Data.
+ @param[in] RequestDataSize Size of Command Request Data.
+ @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
+ @param[in, out] ResponseDataSize Size of Command Response Data.
@retval EFI_SUCCESS The command byte stream was successfully submit to the device and a response was successfully received.
@retval EFI_NOT_FOUND The command was not successfully sent to the device or a response was not successfully received from the device.
@@ -100,6 +106,8 @@ CommonPldmSubmitCommand (
IN MANAGEABILITY_TRANSPORT_TOKEN *TransportToken,
IN UINT8 PldmType,
IN UINT8 PldmCommand,
+ IN UINT8 PldmTerminusSourceId,
+ IN UINT8 PldmTerminusDestinationId,
IN UINT8 *RequestData OPTIONAL,
IN UINT32 RequestDataSize,
OUT UINT8 *ResponseData OPTIONAL,
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
index b2ca69b05f..726747416c 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
@@ -25,13 +25,15 @@ UINT32 TransportMaximumPayload;
/**
This service enables submitting commands via EDKII PLDM protocol.
- @param[in] This EDKII_PLDM_PROTOCOL instance.
- @param[in] PldmType PLDM message type.
- @param[in] Command PLDM Command of PLDM message type.
- @param[in] RequestData Command Request Data.
- @param[in] RequestDataSize Size of Command Request Data.
- @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
- @param[in, out] ResponseDataSize Size of Command Response Data.
+ @param[in] This EDKII_PLDM_PROTOCOL instance.
+ @param[in] PldmType PLDM message type.
+ @param[in] Command PLDM Command of PLDM message type.
+ @param[in] PldmTerminusSourceId PLDM source teminus ID.
+ @param[in] PldmTerminusDestinationId PLDM destination teminus ID.
+ @param[in] RequestData Command Request Data.
+ @param[in] RequestDataSize Size of Command Request Data.
+ @param[out] ResponseData Command Response Data. The completion code is the first byte of response data.
+ @param[in, out] ResponseDataSize Size of Command Response Data.
@retval EFI_SUCCESS The command byte stream was successfully submit to the device and a response was successfully received.
@retval EFI_NOT_FOUND The command was not successfully sent to the device or a response was not successfully received from the device.
@@ -39,7 +41,7 @@ UINT32 TransportMaximumPayload;
@retval EFI_DEVICE_ERROR PLDM transport interface Device hardware error.
@retval EFI_TIMEOUT The command time out.
@retval EFI_UNSUPPORTED The command was not successfully sent to the device.
- @retval EFI_OUT_OF_RESOURCES The resource allocation is out of resource or data size error.
+ @retval EFI_OUT_OF_RESOURCES The resource allcation is out of resource or data size error.
@retval EFI_INVALID_PARAMETER Both RequestData and ResponseData are NULL
**/
EFI_STATUS
@@ -48,6 +50,8 @@ PldmSubmitCommand (
IN EDKII_PLDM_PROTOCOL *This,
IN UINT8 PldmType,
IN UINT8 Command,
+ IN UINT8 PldmTerminusSourceId,
+ IN UINT8 PldmTerminusDestinationId,
IN UINT8 *RequestData,
IN UINT32 RequestDataSize,
OUT UINT8 *ResponseData,
@@ -61,10 +65,46 @@ PldmSubmitCommand (
return EFI_INVALID_PARAMETER;
}
+ if (RequestData != NULL && RequestDataSize == 0) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: RequestDataSize == 0, however RequestData is not NULL for PLDM type: 0x%x, Command: 0x%x.\n",
+ __func__,
+ PldmType,
+ Command
+ ));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ if (ResponseData == NULL && *ResponseDataSize != 0) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: *ResponseDataSize != 0, however ResponseData is NULL for PLDM type: 0x%x, Command: 0x%x.\n",
+ __func__,
+ PldmType,
+ Command
+ ));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ if (ResponseData != NULL && *ResponseDataSize == 0) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: *ResponseDataSize == 0, however ResponseData is not NULL for PLDM type: 0x%x, Command: 0x%x.\n",
+ __func__,
+ PldmType,
+ Command
+ ));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ DEBUG ((DEBUG_MANAGEABILITY, "%a: Source terminus ID: 0x%x, Destination terminus ID: 0x%x.\n"));
Status = CommonPldmSubmitCommand (
mTransportToken,
PldmType,
Command,
+ PldmTerminusSourceId,
+ PldmTerminusDestinationId,
RequestData,
RequestDataSize,
ResponseData,
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocolDxe.inf b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocolDxe.inf
index 006f77b09a..aef25f6438 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocolDxe.inf
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocolDxe.inf
@@ -42,9 +42,5 @@
[Protocols]
gEdkiiPldmProtocolGuid
-[FixedPcd]
- gManageabilityPkgTokenSpaceGuid.PcdMctpSourceEndpointId
- gManageabilityPkgTokenSpaceGuid.PcdMctpDestinationEndpointId
-
[Depex]
TRUE
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109920): https://edk2.groups.io/g/devel/message/109920
Mute This Topic: https://groups.io/mt/102134662/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 13/16] ManageabilityPkg/PldmProtocol: Remove PLDM command table
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (11 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 12/16] PldmProtocolDxe: Correct TID argument usage Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 14/16] ManageabilityPkg: Return error on PLDM header check fails Konstantin Aladyshev
` (3 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
From: Abner Chang <abner.chang@amd.com>
In case of the PLDM/MCTP communication response size doesn't have to be
known beforehand, the caller just need to provide the buffer big enough
to accomodate the response.
Remove PLDM command table for retrieving the response payload size and
correct the code to fix the response buffer size handling.
Also update the message for error conditions.
Signed-off-by: Abner Chang <abner.chang@amd.com>
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../PldmProtocol/Common/PldmProtocolCommon.c | 100 +++---------------
.../PldmProtocol/Common/PldmProtocolCommon.h | 3 +
.../Universal/PldmProtocol/Dxe/PldmProtocol.c | 13 ++-
3 files changed, 31 insertions(+), 85 deletions(-)
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index ea3d4a22b2..bc72ce07b3 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -21,42 +21,6 @@
extern CHAR16 *mTransportName;
extern UINT8 mPldmRequestInstanceId;
-PLDM_MESSAGE_PACKET_MAPPING PldmMessagePacketMappingTable[] = {
- { PLDM_TYPE_SMBIOS, PLDM_GET_SMBIOS_STRUCTURE_TABLE_METADATA_COMMAND_CODE, sizeof (PLDM_GET_SMBIOS_STRUCTURE_TABLE_METADATA_RESPONSE_FORMAT) },
- { PLDM_TYPE_SMBIOS, PLDM_SET_SMBIOS_STRUCTURE_TABLE_METADATA_COMMAND_CODE, sizeof (PLDM_SET_SMBIOS_STRUCTURE_TABLE_METADATA_RESPONSE_FORMAT) },
- { PLDM_TYPE_SMBIOS, PLDM_SET_SMBIOS_STRUCTURE_TABLE_COMMAND_CODE, sizeof (PLDM_SET_SMBIOS_STRUCTURE_TABLE_REQUEST_FORMAT) }
-};
-
-/**
- This function returns the expected full size of PLDM response message.
-
- @param[in] PldmType PLDM message type.
- @param[in] PldmCommand PLDM command of this PLDM type.
-
- @retval Zero No matched entry for this PldmType/PldmCommand.
- @retval None-zero Size of full packet is returned.
-**/
-UINT32
-GetFullPacketResponseSize (
- IN UINT8 PldmType,
- IN UINT8 PldmCommand
- )
-{
- INT16 Index;
- PLDM_MESSAGE_PACKET_MAPPING *ThisEntry;
-
- ThisEntry = PldmMessagePacketMappingTable;
- for (Index = 0; Index < (sizeof (PldmMessagePacketMappingTable)/ sizeof (PLDM_MESSAGE_PACKET_MAPPING)); Index++) {
- if ((PldmType == ThisEntry->PldmType) && (PldmCommand == ThisEntry->PldmCommand)) {
- return ThisEntry->ResponseSize;
- }
-
- ThisEntry++;
- }
-
- return 0;
-}
-
/**
This functions setup the final header/body/trailer packets for
the acquired transport interface.
@@ -267,10 +231,10 @@ CommonPldmSubmitCommand (
TransferToken.TransmitPackage.TransmitTimeoutInMillisecond = MANAGEABILITY_TRANSPORT_NO_TIMEOUT;
// Set receive packet.
- FullPacketResponseDataSize = GetFullPacketResponseSize (PldmType, PldmCommand);
- if (FullPacketResponseDataSize == 0) {
- DEBUG ((DEBUG_ERROR, " No mapping entry in PldmMessagePacketMappingTable for PLDM Type:%d Command %d\n", PldmType, PldmCommand));
- ASSERT (FALSE);
+ if (ResponseData == NULL && *ResponseDataSize == 0) {
+ FullPacketResponseDataSize = sizeof (PLDM_RESPONSE_HEADER);
+ } else {
+ FullPacketResponseDataSize = *ResponseDataSize + sizeof (PLDM_RESPONSE_HEADER);
}
FullPacketResponseData = (UINT8 *)AllocateZeroPool (FullPacketResponseDataSize);
@@ -306,6 +270,7 @@ CommonPldmSubmitCommand (
);
//
// Check the response size.
+ //
if (TransferToken.ReceivePackage.ReceiveSizeInByte < sizeof (PLDM_RESPONSE_HEADER)) {
DEBUG ((
DEBUG_MANAGEABILITY_INFO,
@@ -315,21 +280,13 @@ CommonPldmSubmitCommand (
TransferToken.ReceivePackage.ReceiveSizeInByte,
FullPacketResponseDataSize
));
- if (ResponseDataSize != NULL) {
- if (*ResponseDataSize > TransferToken.ReceivePackage.ReceiveSizeInByte) {
- *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte;
- }
- }
-
- if (ResponseData != NULL) {
- CopyMem ((VOID *)ResponseData, (VOID *)FullPacketResponseData, *ResponseDataSize);
- }
-
+ HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
goto ErrorExit;
}
//
// Check the integrity of response. data.
+ //
ResponseHeader = (PLDM_RESPONSE_HEADER *)FullPacketResponseData;
if ((ResponseHeader->PldmHeader.DatagramBit != (!PLDM_MESSAGE_HEADER_IS_DATAGRAM)) ||
(ResponseHeader->PldmHeader.RequestBit != PLDM_MESSAGE_HEADER_IS_RESPONSE) ||
@@ -343,22 +300,16 @@ CommonPldmSubmitCommand (
DEBUG ((DEBUG_ERROR, " Instance ID = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.InstanceId, mPldmRequestInstanceId));
DEBUG ((DEBUG_ERROR, " Pldm Type = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.PldmType, PldmType));
DEBUG ((DEBUG_ERROR, " Pldm Command = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.PldmTypeCommandCode, PldmCommand));
- if (ResponseDataSize != NULL) {
- if (*ResponseDataSize > TransferToken.ReceivePackage.ReceiveSizeInByte) {
- *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte;
- }
- }
-
- if (ResponseData != NULL) {
- CopyMem ((VOID *)ResponseData, (VOID *)FullPacketResponseData, *ResponseDataSize);
- }
+ DEBUG ((DEBUG_ERROR, " Pldm Completion Code = 0x%x\n", ResponseHeader->PldmCompletionCode));
+ HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
goto ErrorExit;
}
//
// Check the response size
- if (TransferToken.ReceivePackage.ReceiveSizeInByte != FullPacketResponseDataSize) {
+ //
+ if (TransferToken.ReceivePackage.ReceiveSizeInByte > FullPacketResponseDataSize) {
DEBUG ((
DEBUG_ERROR,
"The response size is incorrect: Response size %d (Expected %d), Completion code %d.\n",
@@ -366,38 +317,21 @@ CommonPldmSubmitCommand (
FullPacketResponseDataSize,
ResponseHeader->PldmCompletionCode
));
- if (ResponseDataSize != NULL) {
- if (*ResponseDataSize > TransferToken.ReceivePackage.ReceiveSizeInByte) {
- *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte;
- }
- }
-
- if (ResponseData != NULL) {
- CopyMem ((VOID *)ResponseData, (VOID *)FullPacketResponseData, *ResponseDataSize);
- }
+ HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
goto ErrorExit;
}
- if (*ResponseDataSize != (TransferToken.ReceivePackage.ReceiveSizeInByte - sizeof (PLDM_RESPONSE_HEADER))) {
+ if (*ResponseDataSize < GET_PLDM_MESSAGE_PAYLOAD_SIZE(TransferToken.ReceivePackage.ReceiveSizeInByte)) {
DEBUG ((DEBUG_ERROR, " The size of response is not matched to RequestDataSize assigned by caller.\n"));
DEBUG ((
DEBUG_ERROR,
"Caller expects %d, the response size minus PLDM_RESPONSE_HEADER size is %d, Completion Code %d.\n",
*ResponseDataSize,
- TransferToken.ReceivePackage.ReceiveSizeInByte - sizeof (PLDM_RESPONSE_HEADER),
+ GET_PLDM_MESSAGE_PAYLOAD_SIZE(TransferToken.ReceivePackage.ReceiveSizeInByte),
ResponseHeader->PldmCompletionCode
));
- if (ResponseDataSize != NULL) {
- if (*ResponseDataSize > TransferToken.ReceivePackage.ReceiveSizeInByte) {
- *ResponseDataSize = TransferToken.ReceivePackage.ReceiveSizeInByte;
- }
- }
-
- if (ResponseData != NULL) {
- CopyMem ((VOID *)ResponseData, (VOID *)FullPacketResponseData, *ResponseDataSize);
- }
-
+ HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, GET_PLDM_MESSAGE_PAYLOAD_SIZE(TransferToken.ReceivePackage.ReceiveSizeInByte), "Failed response payload\n");
goto ErrorExit;
}
@@ -406,10 +340,10 @@ CommonPldmSubmitCommand (
// Copy response data (without header) to caller's buffer.
if ((ResponseData != NULL) && (*ResponseDataSize != 0)) {
- *ResponseDataSize = FullPacketResponseDataSize - sizeof (PLDM_RESPONSE_HEADER);
+ *ResponseDataSize = GET_PLDM_MESSAGE_PAYLOAD_SIZE(TransferToken.ReceivePackage.ReceiveSizeInByte);
CopyMem (
(VOID *)ResponseData,
- (VOID *)(FullPacketResponseData + sizeof (PLDM_RESPONSE_HEADER)),
+ GET_PLDM_MESSAGE_PAYLOAD_PTR(FullPacketResponseData),
*ResponseDataSize
);
}
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
index 79431dd3b1..eb273c4f46 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.h
@@ -12,6 +12,9 @@
#include <IndustryStandard/Pldm.h>
#include <Library/ManageabilityTransportLib.h>
+#define GET_PLDM_MESSAGE_PAYLOAD_SIZE(PayloadSize) (PayloadSize - sizeof (PLDM_RESPONSE_HEADER))
+#define GET_PLDM_MESSAGE_PAYLOAD_PTR(PayloadPtr) ((UINT8 *)PayloadPtr + sizeof (PLDM_RESPONSE_HEADER))
+
typedef struct {
UINT8 PldmType;
UINT8 PldmCommand;
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
index 726747416c..058f98e677 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Dxe/PldmProtocol.c
@@ -60,8 +60,17 @@ PldmSubmitCommand (
{
EFI_STATUS Status;
- if ((RequestData == NULL) && (ResponseData == NULL)) {
- DEBUG ((DEBUG_ERROR, "%a: Both RequestData and ResponseData are NULL\n", __func__));
+ //
+ // Check the given input parameters.
+ //
+ if (RequestData == NULL && RequestDataSize != 0) {
+ DEBUG ((
+ DEBUG_ERROR,
+ "%a: RequestDataSize != 0, however RequestData is NULL for PLDM type: 0x%x, Command: 0x%x.\n",
+ __func__,
+ PldmType,
+ Command
+ ));
return EFI_INVALID_PARAMETER;
}
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109921): https://edk2.groups.io/g/devel/message/109921
Mute This Topic: https://groups.io/mt/102134663/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 14/16] ManageabilityPkg: Return error on PLDM header check fails
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (12 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 13/16] ManageabilityPkg/PldmProtocol: Remove PLDM command table Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 15/16] PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API Konstantin Aladyshev
` (2 subsequent siblings)
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Currently PldmSubmit command returns EFI_SUCCESS even if the response
header checks have failed.
Correct the code to return errors in such cases.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Universal/PldmProtocol/Common/PldmProtocolCommon.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index bc72ce07b3..04f250e57c 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -241,7 +241,7 @@ CommonPldmSubmitCommand (
if (FullPacketResponseData == NULL) {
DEBUG ((DEBUG_ERROR, " Not enough memory for FullPacketResponseDataSize.\n"));
Status = EFI_OUT_OF_RESOURCES;
- goto ErrorExit2;
+ goto ErrorExit;
}
// Print out PLDM packet.
@@ -281,6 +281,7 @@ CommonPldmSubmitCommand (
FullPacketResponseDataSize
));
HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
+ Status = EFI_DEVICE_ERROR;
goto ErrorExit;
}
@@ -303,6 +304,7 @@ CommonPldmSubmitCommand (
DEBUG ((DEBUG_ERROR, " Pldm Completion Code = 0x%x\n", ResponseHeader->PldmCompletionCode));
HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
+ Status = EFI_DEVICE_ERROR;
goto ErrorExit;
}
@@ -319,6 +321,7 @@ CommonPldmSubmitCommand (
));
HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, TransferToken.ReceivePackage.ReceiveSizeInByte, "Failed response payload\n");
+ Status = EFI_DEVICE_ERROR;
goto ErrorExit;
}
@@ -332,6 +335,7 @@ CommonPldmSubmitCommand (
ResponseHeader->PldmCompletionCode
));
HelperManageabilityDebugPrint ((VOID *)FullPacketResponseData, GET_PLDM_MESSAGE_PAYLOAD_SIZE(TransferToken.ReceivePackage.ReceiveSizeInByte), "Failed response payload\n");
+ Status = EFI_DEVICE_ERROR;
goto ErrorExit;
}
@@ -350,13 +354,12 @@ CommonPldmSubmitCommand (
// Return transfer status.
//
-ErrorExit:
Status = TransferToken.TransferStatus;
if (EFI_ERROR (Status)) {
DEBUG ((DEBUG_ERROR, "%a: Failed to send PLDM command over %s\n", __func__, mTransportName));
}
-ErrorExit2:
+ErrorExit:
if (PldmTransportHeader != NULL) {
FreePool ((VOID *)PldmTransportHeader);
}
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109922): https://edk2.groups.io/g/devel/message/109922
Mute This Topic: https://groups.io/mt/102134664/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 15/16] PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (13 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 14/16] ManageabilityPkg: Return error on PLDM header check fails Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 16/16] ManageabilityPkg: Check PLDM completion code Konstantin Aladyshev
2023-10-24 2:44 ` [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Chang, Abner via groups.io
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
From: Abner Chang <abner.chang@amd.com>
Currently all PLDM functions inside the PLDM_SMBIOS_TRANSFER_PROTOCOL
use PLDM terminus PCDs for the MCTP addressing.
Add additional function to the protocol API to provide user a way to
use custom TIDs.
Signed-off-by: Abner Chang <abner.chang@amd.com>
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Protocol/PldmSmbiosTransferProtocol.h | 26 +++++++++++++++++
.../PldmSmbiosTransferDxe.c | 28 +++++++++++++++++++
2 files changed, 54 insertions(+)
diff --git a/Features/ManageabilityPkg/Include/Protocol/PldmSmbiosTransferProtocol.h b/Features/ManageabilityPkg/Include/Protocol/PldmSmbiosTransferProtocol.h
index 7903e12726..8b23d39682 100644
--- a/Features/ManageabilityPkg/Include/Protocol/PldmSmbiosTransferProtocol.h
+++ b/Features/ManageabilityPkg/Include/Protocol/PldmSmbiosTransferProtocol.h
@@ -23,6 +23,31 @@ typedef struct _EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL EDKII_PLDM_SMBIOS_TRANSFER_
#define EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL_VERSION ((EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL_VERSION_MAJOR << 8) |\
EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL_VERSION_MINOR)
+/**
+ This function sets PLDM SMBIOS transfer source and destination
+ PLDM terminus ID.
+
+ @param [in] This EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL instance.
+ @param [in] SourceId PLDM source teminus ID.
+ Set to PLDM_TERMINUS_ID_UNASSIGNED means use
+ platform default PLDM terminus ID.
+ (gManageabilityPkgTokenSpaceGuid.PcdPldmSourceTerminusId)
+ @param [in] DestinationId PLDM destination teminus ID.
+ Set to PLDM_TERMINUS_ID_UNASSIGNED means use
+ platform default PLDM terminus ID.
+ (gManageabilityPkgTokenSpaceGuid.PcdPldmDestinationEndpointId)
+
+ @retval EFI_SUCCESS Get SMBIOS table metadata Successfully.
+ @retval EFI_INVALID_PARAMETER Invalid value of source or destination
+ PLDM terminus ID.
+**/
+typedef
+EFI_STATUS
+(EFIAPI *PLDM_GET_SMBIOS_TRANSFER_TERMINUS_ID)(
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId
+ );
+
/**
This function gets SMBIOS table metadata.
@@ -151,6 +176,7 @@ EFI_STATUS
// EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL
//
typedef struct {
+ PLDM_GET_SMBIOS_TRANSFER_TERMINUS_ID SetPldmSmbiosTransferTerminusId;
PLDM_GET_SMBIOS_STRUCTURE_TABLE_METADATA GetSmbiosStructureTableMetaData;
PLDM_SET_SMBIOS_STRUCTURE_TABLE_METADATA SetSmbiosStructureTableMetaData;
PLDM_GET_SMBIOS_STRUCTURE_TABLE GetSmbiosStructureTable;
diff --git a/Features/ManageabilityPkg/Universal/PldmSmbiosTransferDxe/PldmSmbiosTransferDxe.c b/Features/ManageabilityPkg/Universal/PldmSmbiosTransferDxe/PldmSmbiosTransferDxe.c
index fdf033f0b1..357a7d49e4 100644
--- a/Features/ManageabilityPkg/Universal/PldmSmbiosTransferDxe/PldmSmbiosTransferDxe.c
+++ b/Features/ManageabilityPkg/Universal/PldmSmbiosTransferDxe/PldmSmbiosTransferDxe.c
@@ -25,6 +25,33 @@
UINT32 SetSmbiosStructureTableHandle;
+/**
+ This function sets PLDM SMBIOS transfer source and destination
+ PLDM terminus ID.
+
+ @param [in] This EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL instance.
+ @param [in] SourceId PLDM source teminus ID.
+ Set to PLDM_TERMINUS_ID_UNASSIGNED means use
+ platform default PLDM terminus ID.
+ (gManageabilityPkgTokenSpaceGuid.PcdPldmSourceTerminusId)
+ @param [in] DestinationId PLDM destination teminus ID.
+ Set to PLDM_TERMINUS_ID_UNASSIGNED means use
+ platform default PLDM terminus ID.
+ (gManageabilityPkgTokenSpaceGuid.PcdPldmDestinationEndpointId)
+
+ @retval EFI_SUCCESS Get SMBIOS table metadata Successfully.
+ @retval EFI_INVALID_PARAMETER Invalid value of source or destination
+ PLDM terminus ID.
+**/
+EFI_STATUS
+SetPldmSmbiosTransferTerminusId (
+ IN UINT8 SourceId,
+ IN UINT8 DestinationId
+ )
+{
+ return PldmSetTerminus(SourceId, DestinationId);
+}
+
/**
Get the full size of SMBIOS structure including optional strings that follow the formatted structure.
@@ -457,6 +484,7 @@ GetSmbiosStructureByHandle (
}
EDKII_PLDM_SMBIOS_TRANSFER_PROTOCOL_V1_0 mPldmSmbiosTransferProtocolV10 = {
+ SetPldmSmbiosTransferTerminusId,
GetSmbiosStructureTableMetaData,
SetSmbiosStructureTableMetaData,
GetSmbiosStructureTable,
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109923): https://edk2.groups.io/g/devel/message/109923
Mute This Topic: https://groups.io/mt/102134665/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* [edk2-devel] [PATCH edk2-platforms v3 16/16] ManageabilityPkg: Check PLDM completion code
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (14 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 15/16] PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API Konstantin Aladyshev
@ 2023-10-23 13:05 ` Konstantin Aladyshev
2023-10-24 2:44 ` [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Chang, Abner via groups.io
16 siblings, 0 replies; 25+ messages in thread
From: Konstantin Aladyshev @ 2023-10-23 13:05 UTC (permalink / raw)
To: devel; +Cc: abner.chang, AbdulLateef.Attar, nicklew, Konstantin Aladyshev
Every PLDM response contains 1 byte Completion code which is needed
to be checked against SUCCESS value which means that the PLDM command
was accepted and completed normally.
If it is not equal to that value it is necessary to return error status
so the caller would know that Response buffer is invalid.
Signed-off-by: Konstantin Aladyshev <aladyshev22@gmail.com>
---
.../Universal/PldmProtocol/Common/PldmProtocolCommon.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
index 04f250e57c..2f2f76603e 100644
--- a/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
+++ b/Features/ManageabilityPkg/Universal/PldmProtocol/Common/PldmProtocolCommon.c
@@ -293,7 +293,8 @@ CommonPldmSubmitCommand (
(ResponseHeader->PldmHeader.RequestBit != PLDM_MESSAGE_HEADER_IS_RESPONSE) ||
(ResponseHeader->PldmHeader.InstanceId != mPldmRequestInstanceId) ||
(ResponseHeader->PldmHeader.PldmType != PldmType) ||
- (ResponseHeader->PldmHeader.PldmTypeCommandCode != PldmCommand))
+ (ResponseHeader->PldmHeader.PldmTypeCommandCode != PldmCommand) ||
+ (ResponseHeader->PldmCompletionCode != PLDM_COMPLETION_CODE_SUCCESS))
{
DEBUG ((DEBUG_ERROR, "PLDM integrity check of response data is failed.\n"));
DEBUG ((DEBUG_ERROR, " Datagram = %d (Expected value: %d)\n", ResponseHeader->PldmHeader.DatagramBit, (!PLDM_MESSAGE_HEADER_IS_DATAGRAM)));
--
2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109924): https://edk2.groups.io/g/devel/message/109924
Mute This Topic: https://groups.io/mt/102134666/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply related [flat|nested] 25+ messages in thread* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-10-23 13:04 [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Konstantin Aladyshev
` (15 preceding siblings ...)
2023-10-23 13:05 ` [edk2-devel] [PATCH edk2-platforms v3 16/16] ManageabilityPkg: Check PLDM completion code Konstantin Aladyshev
@ 2023-10-24 2:44 ` Chang, Abner via groups.io
2023-10-26 5:03 ` Chang, Abner via groups.io
16 siblings, 1 reply; 25+ messages in thread
From: Chang, Abner via groups.io @ 2023-10-24 2:44 UTC (permalink / raw)
To: Konstantin Aladyshev, devel@edk2.groups.io
Cc: Attar, AbdulLateef (Abdul Lateef), nicklew@nvidia.com
[AMD Official Use Only - General]
For entire V3 and the additional patch 16/16,
Reviewed-by: Abner Chang <abner.chang@amd.com>
I will do the Uncrustify check and merge this patch set once the corresponding edk2 changes are merged.
Thanks
Abner
> -----Original Message-----
> From: Konstantin Aladyshev <aladyshev22@gmail.com>
> Sent: Monday, October 23, 2023 9:05 PM
> To: devel@edk2.groups.io
> Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul
> Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin
> Aladyshev <aladyshev22@gmail.com>
> Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>
> Caution: This message originated from an External Source. Use proper caution
> when opening attachments, clicking links, or responding.
>
>
> The Manageability KCS transport library needs to support requests both
> from MCTP and IPMI transports. Currently the code only handles IPMI
> case correctly.
> In the MCTP case the communication should be based on the MCTP-over-KCS
> specification (DSP0254). This specification defines a special KCS
> binding header and trailer structures that need to be present in every
> MCTP message.
> The header structure contains a length field, therefore response packet
> size is not needed to be known beforehand.
> The trailer structure contains a PEC checksum that can be used to check
> itegrity of the response message.
> Modify Manageability KCS transport library code to check which message
> is processed (IPMI or MCTP) and handle each case correctly based on its
> own specification.
> This patch is a result of a joint effort from the Konstantin Aladyshev
> <aladyshev22@gmail.com> and Abner Chang <abner.chang@amd.com>.
>
> Tested:
> PLDM communication between the HOST and BMC was tested with both
> components implemented via open-source software:
> - The HOST (UEFI firmware) part was based one the edk2 [1] and
> edk2-platforms [2] code,
> - The BMC part was based on the openbmc [3] distribution.
>
> The testing process and all the necessary utilities are described in
> the [4] repository.
>
> The provided changes keep IPMI over KCS stack working as reported by
> Abner Chang.
>
> [1]: https://github.com/tianocore/edk2
> [2]: https://github.com/tianocore/edk2-platforms
> [3]: https://github.com/openbmc/openbmc
> [4]: https://github.com/Kostr/PLDM
>
> Changes v2 -> v3:
> - Add new patch that adds PLDM completion code check
>
> Changes v1 -> v2:
> - Add new patches with corrections for the PLDM protocol. The
> resulting communication via EDKII_PLDM_PROTOCOL was successfully
> tested.
>
> Abner Chang (4):
> ManageabilityPkg: Add PLDM terminus PCDs
> PldmProtocolDxe: Correct TID argument usage
> ManageabilityPkg/PldmProtocol: Remove PLDM command table
> PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
>
> Konstantin Aladyshev (12):
> ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
> ManageabilityPkg: Check MCTP EIDs for reserved values
> ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
> ManageabilityPkg: Check header fields in the MCTP response
> ManageabilityPkg: Correct typo in MCTP destination EID field
> ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
> ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
> ManageabilityPkg: Don't check MCTP header fields if transfer has
> failed
> ManageabilityPkg: Use correct constants for PLDM header checks
> ManageabilityPkg: Return error on multiple-packet MCTP responses
> ManageabilityPkg: Return error on PLDM header check fails
> ManageabilityPkg: Check PLDM completion code
>
> .../Include/Library/BasePldmProtocolLib.h | 16 +
> .../Library/ManageabilityTransportMctpLib.h | 9 +-
> .../Include/Protocol/MctpProtocol.h | 12 +-
> .../Include/Protocol/PldmProtocol.h | 18 +-
> .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
> .../Common/KcsCommon.c | 284 +++++++++++++++---
> .../Dxe/ManageabilityTransportMctp.c | 4 +-
> .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
> .../Dxe/PldmProtocolLib.inf | 6 +-
> .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
> .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
> .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
> .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
> .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
> .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
> .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
> .../PldmSmbiosTransferDxe.c | 28 ++
> 17 files changed, 690 insertions(+), 194 deletions(-)
>
> --
> 2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#109980): https://edk2.groups.io/g/devel/message/109980
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-10-24 2:44 ` [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support Chang, Abner via groups.io
@ 2023-10-26 5:03 ` Chang, Abner via groups.io
2023-10-31 6:37 ` Yoshinoya
0 siblings, 1 reply; 25+ messages in thread
From: Chang, Abner via groups.io @ 2023-10-26 5:03 UTC (permalink / raw)
To: Konstantin Aladyshev, devel@edk2.groups.io
Cc: Attar, AbdulLateef (Abdul Lateef), nicklew@nvidia.com
[AMD Official Use Only - General]
Merged. I will send another patch for Uncrustify updates if any.
Abner
> -----Original Message-----
> From: Chang, Abner
> Sent: Tuesday, October 24, 2023 10:44 AM
> To: Konstantin Aladyshev <aladyshev22@gmail.com>; devel@edk2.groups.io
> Cc: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>;
> nicklew@nvidia.com
> Subject: RE: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>
> For entire V3 and the additional patch 16/16,
> Reviewed-by: Abner Chang <abner.chang@amd.com>
>
> I will do the Uncrustify check and merge this patch set once the corresponding
> edk2 changes are merged.
> Thanks
> Abner
>
> > -----Original Message-----
> > From: Konstantin Aladyshev <aladyshev22@gmail.com>
> > Sent: Monday, October 23, 2023 9:05 PM
> > To: devel@edk2.groups.io
> > Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul
> > Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin
> > Aladyshev <aladyshev22@gmail.com>
> > Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
> >
> > Caution: This message originated from an External Source. Use proper
> caution
> > when opening attachments, clicking links, or responding.
> >
> >
> > The Manageability KCS transport library needs to support requests both
> > from MCTP and IPMI transports. Currently the code only handles IPMI
> > case correctly.
> > In the MCTP case the communication should be based on the MCTP-over-
> KCS
> > specification (DSP0254). This specification defines a special KCS
> > binding header and trailer structures that need to be present in every
> > MCTP message.
> > The header structure contains a length field, therefore response packet
> > size is not needed to be known beforehand.
> > The trailer structure contains a PEC checksum that can be used to check
> > itegrity of the response message.
> > Modify Manageability KCS transport library code to check which message
> > is processed (IPMI or MCTP) and handle each case correctly based on its
> > own specification.
> > This patch is a result of a joint effort from the Konstantin Aladyshev
> > <aladyshev22@gmail.com> and Abner Chang <abner.chang@amd.com>.
> >
> > Tested:
> > PLDM communication between the HOST and BMC was tested with both
> > components implemented via open-source software:
> > - The HOST (UEFI firmware) part was based one the edk2 [1] and
> > edk2-platforms [2] code,
> > - The BMC part was based on the openbmc [3] distribution.
> >
> > The testing process and all the necessary utilities are described in
> > the [4] repository.
> >
> > The provided changes keep IPMI over KCS stack working as reported by
> > Abner Chang.
> >
> > [1]: https://github.com/tianocore/edk2
> > [2]: https://github.com/tianocore/edk2-platforms
> > [3]: https://github.com/openbmc/openbmc
> > [4]: https://github.com/Kostr/PLDM
> >
> > Changes v2 -> v3:
> > - Add new patch that adds PLDM completion code check
> >
> > Changes v1 -> v2:
> > - Add new patches with corrections for the PLDM protocol. The
> > resulting communication via EDKII_PLDM_PROTOCOL was successfully
> > tested.
> >
> > Abner Chang (4):
> > ManageabilityPkg: Add PLDM terminus PCDs
> > PldmProtocolDxe: Correct TID argument usage
> > ManageabilityPkg/PldmProtocol: Remove PLDM command table
> > PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
> >
> > Konstantin Aladyshev (12):
> > ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
> > ManageabilityPkg: Check MCTP EIDs for reserved values
> > ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
> > ManageabilityPkg: Check header fields in the MCTP response
> > ManageabilityPkg: Correct typo in MCTP destination EID field
> > ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
> > ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
> > ManageabilityPkg: Don't check MCTP header fields if transfer has
> > failed
> > ManageabilityPkg: Use correct constants for PLDM header checks
> > ManageabilityPkg: Return error on multiple-packet MCTP responses
> > ManageabilityPkg: Return error on PLDM header check fails
> > ManageabilityPkg: Check PLDM completion code
> >
> > .../Include/Library/BasePldmProtocolLib.h | 16 +
> > .../Library/ManageabilityTransportMctpLib.h | 9 +-
> > .../Include/Protocol/MctpProtocol.h | 12 +-
> > .../Include/Protocol/PldmProtocol.h | 18 +-
> > .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
> > .../Common/KcsCommon.c | 284 +++++++++++++++---
> > .../Dxe/ManageabilityTransportMctp.c | 4 +-
> > .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
> > .../Dxe/PldmProtocolLib.inf | 6 +-
> > .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
> > .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
> > .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
> > .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
> > .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
> > .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
> > .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
> > .../PldmSmbiosTransferDxe.c | 28 ++
> > 17 files changed, 690 insertions(+), 194 deletions(-)
> >
> > --
> > 2.34.1
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110070): https://edk2.groups.io/g/devel/message/110070
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-10-26 5:03 ` Chang, Abner via groups.io
@ 2023-10-31 6:37 ` Yoshinoya
2023-10-31 7:00 ` Chang, Abner via groups.io
0 siblings, 1 reply; 25+ messages in thread
From: Yoshinoya @ 2023-10-31 6:37 UTC (permalink / raw)
To: devel, abner.chang
Cc: Konstantin Aladyshev, Attar, AbdulLateef (Abdul Lateef),
nicklew@nvidia.com
[-- Attachment #1: Type: text/plain, Size: 6298 bytes --]
Hi, Abner:
I ask you a favor.
KCS actually layer on LPC interface.
KCS is a usual BMC-to-CPU communication channel.
Does MCTP-over-KCS feature require some special LPC interface hardware changes?
Or, MCTP is just a software stack and it uses current LPC interface, not any special hardward design change requirement.
Thanks
At 2023-10-26 13:03:02, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io> wrote:
>[AMD Official Use Only - General]
>
>Merged. I will send another patch for Uncrustify updates if any.
>
>Abner
>
>> -----Original Message-----
>> From: Chang, Abner
>> Sent: Tuesday, October 24, 2023 10:44 AM
>> To: Konstantin Aladyshev <aladyshev22@gmail.com>; devel@edk2.groups.io
>> Cc: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>;
>> nicklew@nvidia.com
>> Subject: RE: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>>
>> For entire V3 and the additional patch 16/16,
>> Reviewed-by: Abner Chang <abner.chang@amd.com>
>>
>> I will do the Uncrustify check and merge this patch set once the corresponding
>> edk2 changes are merged.
>> Thanks
>> Abner
>>
>> > -----Original Message-----
>> > From: Konstantin Aladyshev <aladyshev22@gmail.com>
>> > Sent: Monday, October 23, 2023 9:05 PM
>> > To: devel@edk2.groups.io
>> > Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul
>> > Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin
>> > Aladyshev <aladyshev22@gmail.com>
>> > Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>> >
>> > Caution: This message originated from an External Source. Use proper
>> caution
>> > when opening attachments, clicking links, or responding.
>> >
>> >
>> > The Manageability KCS transport library needs to support requests both
>> > from MCTP and IPMI transports. Currently the code only handles IPMI
>> > case correctly.
>> > In the MCTP case the communication should be based on the MCTP-over-
>> KCS
>> > specification (DSP0254). This specification defines a special KCS
>> > binding header and trailer structures that need to be present in every
>> > MCTP message.
>> > The header structure contains a length field, therefore response packet
>> > size is not needed to be known beforehand.
>> > The trailer structure contains a PEC checksum that can be used to check
>> > itegrity of the response message.
>> > Modify Manageability KCS transport library code to check which message
>> > is processed (IPMI or MCTP) and handle each case correctly based on its
>> > own specification.
>> > This patch is a result of a joint effort from the Konstantin Aladyshev
>> > <aladyshev22@gmail.com> and Abner Chang <abner.chang@amd.com>.
>> >
>> > Tested:
>> > PLDM communication between the HOST and BMC was tested with both
>> > components implemented via open-source software:
>> > - The HOST (UEFI firmware) part was based one the edk2 [1] and
>> > edk2-platforms [2] code,
>> > - The BMC part was based on the openbmc [3] distribution.
>> >
>> > The testing process and all the necessary utilities are described in
>> > the [4] repository.
>> >
>> > The provided changes keep IPMI over KCS stack working as reported by
>> > Abner Chang.
>> >
>> > [1]: https://github.com/tianocore/edk2
>> > [2]: https://github.com/tianocore/edk2-platforms
>> > [3]: https://github.com/openbmc/openbmc
>> > [4]: https://github.com/Kostr/PLDM
>> >
>> > Changes v2 -> v3:
>> > - Add new patch that adds PLDM completion code check
>> >
>> > Changes v1 -> v2:
>> > - Add new patches with corrections for the PLDM protocol. The
>> > resulting communication via EDKII_PLDM_PROTOCOL was successfully
>> > tested.
>> >
>> > Abner Chang (4):
>> > ManageabilityPkg: Add PLDM terminus PCDs
>> > PldmProtocolDxe: Correct TID argument usage
>> > ManageabilityPkg/PldmProtocol: Remove PLDM command table
>> > PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
>> >
>> > Konstantin Aladyshev (12):
>> > ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
>> > ManageabilityPkg: Check MCTP EIDs for reserved values
>> > ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
>> > ManageabilityPkg: Check header fields in the MCTP response
>> > ManageabilityPkg: Correct typo in MCTP destination EID field
>> > ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
>> > ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
>> > ManageabilityPkg: Don't check MCTP header fields if transfer has
>> > failed
>> > ManageabilityPkg: Use correct constants for PLDM header checks
>> > ManageabilityPkg: Return error on multiple-packet MCTP responses
>> > ManageabilityPkg: Return error on PLDM header check fails
>> > ManageabilityPkg: Check PLDM completion code
>> >
>> > .../Include/Library/BasePldmProtocolLib.h | 16 +
>> > .../Library/ManageabilityTransportMctpLib.h | 9 +-
>> > .../Include/Protocol/MctpProtocol.h | 12 +-
>> > .../Include/Protocol/PldmProtocol.h | 18 +-
>> > .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
>> > .../Common/KcsCommon.c | 284 +++++++++++++++---
>> > .../Dxe/ManageabilityTransportMctp.c | 4 +-
>> > .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
>> > .../Dxe/PldmProtocolLib.inf | 6 +-
>> > .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
>> > .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
>> > .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
>> > .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
>> > .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
>> > .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
>> > .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
>> > .../PldmSmbiosTransferDxe.c | 28 ++
>> > 17 files changed, 690 insertions(+), 194 deletions(-)
>> >
>> > --
>> > 2.34.1
>
>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110374): https://edk2.groups.io/g/devel/message/110374
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
[-- Attachment #2: Type: text/html, Size: 8313 bytes --]
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-10-31 6:37 ` Yoshinoya
@ 2023-10-31 7:00 ` Chang, Abner via groups.io
2023-11-01 5:19 ` Yoshinoya
0 siblings, 1 reply; 25+ messages in thread
From: Chang, Abner via groups.io @ 2023-10-31 7:00 UTC (permalink / raw)
To: Yoshinoya, devel@edk2.groups.io
Cc: Konstantin Aladyshev, Attar, AbdulLateef (Abdul Lateef),
nicklew@nvidia.com
[-- Attachment #1: Type: text/plain, Size: 7995 bytes --]
MCTP over KCS defines two types of KCS-like access, one is compatible with IPMI KCS, another is memory map I/O access. It doesn't require any special HW, just require the I/O cycle or memory cycle can be delivered to the management controller.
Yes, we can consider MCTP is a software layer protocol and the that protocol is HW (transport interface) agnostic.
Abner
From: Yoshinoya <yoshinoyatoko@163.com>
Sent: Tuesday, October 31, 2023 2:38 PM
To: devel@edk2.groups.io; Chang, Abner <Abner.Chang@amd.com>
Cc: Konstantin Aladyshev <aladyshev22@gmail.com>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com
Subject: Re:Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
Hi, Abner:
I ask you a favor.
KCS actually layer on LPC interface.
KCS is a usual BMC-to-CPU communication channel.
Does MCTP-over-KCS feature require some special LPC interface hardware changes?
Or, MCTP is just a software stack and it uses current LPC interface, not any special hardward design change requirement.
Thanks
At 2023-10-26 13:03:02, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io<mailto:abner.chang=amd.com@groups.io>> wrote:
>[AMD Official Use Only - General]
>
>Merged. I will send another patch for Uncrustify updates if any.
>
>Abner
>
>> -----Original Message-----
>> From: Chang, Abner
>> Sent: Tuesday, October 24, 2023 10:44 AM
>> To: Konstantin Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>; devel@edk2.groups.io<mailto:devel@edk2.groups.io>
>> Cc: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com<mailto:AbdulLateef.Attar@amd.com>>;
>> nicklew@nvidia.com<mailto:nicklew@nvidia.com>
>> Subject: RE: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>>
>> For entire V3 and the additional patch 16/16,
>> Reviewed-by: Abner Chang <abner.chang@amd.com<mailto:abner.chang@amd.com>>
>>
>> I will do the Uncrustify check and merge this patch set once the corresponding
>> edk2 changes are merged.
>> Thanks
>> Abner
>>
>> > -----Original Message-----
>> > From: Konstantin Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>
>> > Sent: Monday, October 23, 2023 9:05 PM
>> > To: devel@edk2.groups.io<mailto:devel@edk2.groups.io>
>> > Cc: Chang, Abner <Abner.Chang@amd.com<mailto:Abner.Chang@amd.com>>; Attar, AbdulLateef (Abdul
>> > Lateef) <AbdulLateef.Attar@amd.com<mailto:AbdulLateef.Attar@amd.com>>; nicklew@nvidia.com<mailto:nicklew@nvidia.com>; Konstantin
>> > Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>
>> > Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>> >
>> > Caution: This message originated from an External Source. Use proper
>> caution
>> > when opening attachments, clicking links, or responding.
>> >
>> >
>> > The Manageability KCS transport library needs to support requests both
>> > from MCTP and IPMI transports. Currently the code only handles IPMI
>> > case correctly.
>> > In the MCTP case the communication should be based on the MCTP-over-
>> KCS
>> > specification (DSP0254). This specification defines a special KCS
>> > binding header and trailer structures that need to be present in every
>> > MCTP message.
>> > The header structure contains a length field, therefore response packet
>> > size is not needed to be known beforehand.
>> > The trailer structure contains a PEC checksum that can be used to check
>> > itegrity of the response message.
>> > Modify Manageability KCS transport library code to check which message
>> > is processed (IPMI or MCTP) and handle each case correctly based on its
>> > own specification.
>> > This patch is a result of a joint effort from the Konstantin Aladyshev
>> > <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>> and Abner Chang <abner.chang@amd.com<mailto:abner.chang@amd.com>>.
>> >
>> > Tested:
>> > PLDM communication between the HOST and BMC was tested with both
>> > components implemented via open-source software:
>> > - The HOST (UEFI firmware) part was based one the edk2 [1] and
>> > edk2-platforms [2] code,
>> > - The BMC part was based on the openbmc [3] distribution.
>> >
>> > The testing process and all the necessary utilities are described in
>> > the [4] repository.
>> >
>> > The provided changes keep IPMI over KCS stack working as reported by
>> > Abner Chang.
>> >
>> > [1]: https://github.com/tianocore/edk2
>> > [2]: https://github.com/tianocore/edk2-platforms
>> > [3]: https://github.com/openbmc/openbmc
>> > [4]: https://github.com/Kostr/PLDM
>> >
>> > Changes v2 -> v3:
>> > - Add new patch that adds PLDM completion code check
>> >
>> > Changes v1 -> v2:
>> > - Add new patches with corrections for the PLDM protocol. The
>> > resulting communication via EDKII_PLDM_PROTOCOL was successfully
>> > tested.
>> >
>> > Abner Chang (4):
>> > ManageabilityPkg: Add PLDM terminus PCDs
>> > PldmProtocolDxe: Correct TID argument usage
>> > ManageabilityPkg/PldmProtocol: Remove PLDM command table
>> > PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
>> >
>> > Konstantin Aladyshev (12):
>> > ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
>> > ManageabilityPkg: Check MCTP EIDs for reserved values
>> > ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
>> > ManageabilityPkg: Check header fields in the MCTP response
>> > ManageabilityPkg: Correct typo in MCTP destination EID field
>> > ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
>> > ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
>> > ManageabilityPkg: Don't check MCTP header fields if transfer has
>> > failed
>> > ManageabilityPkg: Use correct constants for PLDM header checks
>> > ManageabilityPkg: Return error on multiple-packet MCTP responses
>> > ManageabilityPkg: Return error on PLDM header check fails
>> > ManageabilityPkg: Check PLDM completion code
>> >
>> > .../Include/Library/BasePldmProtocolLib.h | 16 +
>> > .../Library/ManageabilityTransportMctpLib.h | 9 +-
>> > .../Include/Protocol/MctpProtocol.h | 12 +-
>> > .../Include/Protocol/PldmProtocol.h | 18 +-
>> > .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
>> > .../Common/KcsCommon.c | 284 +++++++++++++++---
>> > .../Dxe/ManageabilityTransportMctp.c | 4 +-
>> > .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
>> > .../Dxe/PldmProtocolLib.inf | 6 +-
>> > .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
>> > .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
>> > .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
>> > .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
>> > .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
>> > .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
>> > .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
>> > .../PldmSmbiosTransferDxe.c | 28 ++
>> > 17 files changed, 690 insertions(+), 194 deletions(-)
>> >
>> > --
>> > 2.34.1
>
>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110377): https://edk2.groups.io/g/devel/message/110377
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
[-- Attachment #2: Type: text/html, Size: 23696 bytes --]
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-10-31 7:00 ` Chang, Abner via groups.io
@ 2023-11-01 5:19 ` Yoshinoya
2023-11-01 5:53 ` Chang, Abner via groups.io
0 siblings, 1 reply; 25+ messages in thread
From: Yoshinoya @ 2023-11-01 5:19 UTC (permalink / raw)
To: devel, abner.chang
Cc: Konstantin Aladyshev, Attar, AbdulLateef (Abdul Lateef),
nicklew@nvidia.com
[-- Attachment #1: Type: text/plain, Size: 7406 bytes --]
Hi, Abner:
Got it, Thanks.
Is there any plan to implement MTCP-over-smbus or MTCP-over-PCIE ?
Thanks
At 2023-10-31 15:00:03, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io> wrote:
MCTP over KCS defines two types of KCS-like access, one is compatible with IPMI KCS, another is memory map I/O access. It doesn’t require any special HW, just require the I/O cycle or memory cycle can be delivered to the management controller.
Yes, we can consider MCTP is a software layer protocol and the that protocol is HW (transport interface) agnostic.
Abner
From: Yoshinoya <yoshinoyatoko@163.com>
Sent: Tuesday, October 31, 2023 2:38 PM
To: devel@edk2.groups.io; Chang, Abner <Abner.Chang@amd.com>
Cc: Konstantin Aladyshev <aladyshev22@gmail.com>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com
Subject: Re:Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
| |
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
|
Hi, Abner:
I ask you a favor.
KCS actually layer on LPC interface.
KCS is a usual BMC-to-CPU communication channel.
Does MCTP-over-KCS feature require some special LPC interface hardware changes?
Or, MCTP is just a software stack and it uses current LPC interface, not any special hardward design change requirement.
Thanks
At 2023-10-26 13:03:02, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io> wrote:
>[AMD Official Use Only - General]
>
>Merged. I will send another patch for Uncrustify updates if any.
>
>Abner
>
>> -----Original Message-----
>> From: Chang, Abner
>> Sent: Tuesday, October 24, 2023 10:44 AM
>> To: Konstantin Aladyshev <aladyshev22@gmail.com>; devel@edk2.groups.io
>> Cc: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>;
>> nicklew@nvidia.com
>> Subject: RE: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>>
>> For entire V3 and the additional patch 16/16,
>> Reviewed-by: Abner Chang <abner.chang@amd.com>
>>
>> I will do the Uncrustify check and merge this patch set once the corresponding
>> edk2 changes are merged.
>> Thanks
>> Abner
>>
>> > -----Original Message-----
>> > From: Konstantin Aladyshev <aladyshev22@gmail.com>
>> > Sent: Monday, October 23, 2023 9:05 PM
>> > To: devel@edk2.groups.io
>> > Cc: Chang, Abner <Abner.Chang@amd.com>; Attar, AbdulLateef (Abdul
>> > Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com; Konstantin
>> > Aladyshev <aladyshev22@gmail.com>
>> > Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>> >
>> > Caution: This message originated from an External Source. Use proper
>> caution
>> > when opening attachments, clicking links, or responding.
>> >
>> >
>> > The Manageability KCS transport library needs to support requests both
>> > from MCTP and IPMI transports. Currently the code only handles IPMI
>> > case correctly.
>> > In the MCTP case the communication should be based on the MCTP-over-
>> KCS
>> > specification (DSP0254). This specification defines a special KCS
>> > binding header and trailer structures that need to be present in every
>> > MCTP message.
>> > The header structure contains a length field, therefore response packet
>> > size is not needed to be known beforehand.
>> > The trailer structure contains a PEC checksum that can be used to check
>> > itegrity of the response message.
>> > Modify Manageability KCS transport library code to check which message
>> > is processed (IPMI or MCTP) and handle each case correctly based on its
>> > own specification.
>> > This patch is a result of a joint effort from the Konstantin Aladyshev
>> > <aladyshev22@gmail.com> and Abner Chang <abner.chang@amd.com>.
>> >
>> > Tested:
>> > PLDM communication between the HOST and BMC was tested with both
>> > components implemented via open-source software:
>> > - The HOST (UEFI firmware) part was based one the edk2 [1] and
>> > edk2-platforms [2] code,
>> > - The BMC part was based on the openbmc [3] distribution.
>> >
>> > The testing process and all the necessary utilities are described in
>> > the [4] repository.
>> >
>> > The provided changes keep IPMI over KCS stack working as reported by
>> > Abner Chang.
>> >
>> > [1]: https://github.com/tianocore/edk2
>> > [2]: https://github.com/tianocore/edk2-platforms
>> > [3]: https://github.com/openbmc/openbmc
>> > [4]: https://github.com/Kostr/PLDM
>> >
>> > Changes v2 -> v3:
>> > - Add new patch that adds PLDM completion code check
>> >
>> > Changes v1 -> v2:
>> > - Add new patches with corrections for the PLDM protocol. The
>> > resulting communication via EDKII_PLDM_PROTOCOL was successfully
>> > tested.
>> >
>> > Abner Chang (4):
>> > ManageabilityPkg: Add PLDM terminus PCDs
>> > PldmProtocolDxe: Correct TID argument usage
>> > ManageabilityPkg/PldmProtocol: Remove PLDM command table
>> > PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
>> >
>> > Konstantin Aladyshev (12):
>> > ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
>> > ManageabilityPkg: Check MCTP EIDs for reserved values
>> > ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
>> > ManageabilityPkg: Check header fields in the MCTP response
>> > ManageabilityPkg: Correct typo in MCTP destination EID field
>> > ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
>> > ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
>> > ManageabilityPkg: Don't check MCTP header fields if transfer has
>> > failed
>> > ManageabilityPkg: Use correct constants for PLDM header checks
>> > ManageabilityPkg: Return error on multiple-packet MCTP responses
>> > ManageabilityPkg: Return error on PLDM header check fails
>> > ManageabilityPkg: Check PLDM completion code
>> >
>> > .../Include/Library/BasePldmProtocolLib.h | 16 +
>> > .../Library/ManageabilityTransportMctpLib.h | 9 +-
>> > .../Include/Protocol/MctpProtocol.h | 12 +-
>> > .../Include/Protocol/PldmProtocol.h | 18 +-
>> > .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
>> > .../Common/KcsCommon.c | 284 +++++++++++++++---
>> > .../Dxe/ManageabilityTransportMctp.c | 4 +-
>> > .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
>> > .../Dxe/PldmProtocolLib.inf | 6 +-
>> > .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
>> > .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
>> > .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
>> > .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
>> > .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
>> > .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
>> > .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
>> > .../PldmSmbiosTransferDxe.c | 28 ++
>> > 17 files changed, 690 insertions(+), 194 deletions(-)
>> >
>> > --
>> > 2.34.1
>
>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110466): https://edk2.groups.io/g/devel/message/110466
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
[-- Attachment #2: Type: text/html, Size: 22328 bytes --]
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
2023-11-01 5:19 ` Yoshinoya
@ 2023-11-01 5:53 ` Chang, Abner via groups.io
0 siblings, 0 replies; 25+ messages in thread
From: Chang, Abner via groups.io @ 2023-11-01 5:53 UTC (permalink / raw)
To: Yoshinoya, devel@edk2.groups.io
Cc: Konstantin Aladyshev, Attar, AbdulLateef (Abdul Lateef),
nicklew@nvidia.com
[-- Attachment #1: Type: text/plain, Size: 9127 bytes --]
[AMD Official Use Only - General]
We need the use case on our platforms, it says we don’t have plan at the moment. However we are welcome anyone from community to implement it. 😊
Abner
From: Yoshinoya <yoshinoyatoko@163.com>
Sent: Wednesday, November 1, 2023 1:20 PM
To: devel@edk2.groups.io; Chang, Abner <Abner.Chang@amd.com>
Cc: Konstantin Aladyshev <aladyshev22@gmail.com>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com>; nicklew@nvidia.com
Subject: Re:Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
Hi, Abner:
Got it, Thanks.
Is there any plan to implement MTCP-over-smbus or MTCP-over-PCIE ?
Thanks
At 2023-10-31 15:00:03, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io<mailto:abner.chang=amd.com@groups.io>> wrote:
MCTP over KCS defines two types of KCS-like access, one is compatible with IPMI KCS, another is memory map I/O access. It doesn’t require any special HW, just require the I/O cycle or memory cycle can be delivered to the management controller.
Yes, we can consider MCTP is a software layer protocol and the that protocol is HW (transport interface) agnostic.
Abner
From: Yoshinoya <yoshinoyatoko@163.com<mailto:yoshinoyatoko@163.com>>
Sent: Tuesday, October 31, 2023 2:38 PM
To: devel@edk2.groups.io<mailto:devel@edk2.groups.io>; Chang, Abner <Abner.Chang@amd.com<mailto:Abner.Chang@amd.com>>
Cc: Konstantin Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>; Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com<mailto:AbdulLateef.Attar@amd.com>>; nicklew@nvidia.com<mailto:nicklew@nvidia.com>
Subject: Re:Re: [edk2-devel] [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
Hi, Abner:
I ask you a favor.
KCS actually layer on LPC interface.
KCS is a usual BMC-to-CPU communication channel.
Does MCTP-over-KCS feature require some special LPC interface hardware changes?
Or, MCTP is just a software stack and it uses current LPC interface, not any special hardward design change requirement.
Thanks
At 2023-10-26 13:03:02, "Chang, Abner via groups.io" <abner.chang=amd.com@groups.io<mailto:abner.chang=amd.com@groups.io>> wrote:
>[AMD Official Use Only - General]
>
>Merged. I will send another patch for Uncrustify updates if any.
>
>Abner
>
>> -----Original Message-----
>> From: Chang, Abner
>> Sent: Tuesday, October 24, 2023 10:44 AM
>> To: Konstantin Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>; devel@edk2.groups.io<mailto:devel@edk2.groups.io>
>> Cc: Attar, AbdulLateef (Abdul Lateef) <AbdulLateef.Attar@amd.com<mailto:AbdulLateef.Attar@amd.com>>;
>> nicklew@nvidia.com<mailto:nicklew@nvidia.com>
>> Subject: RE: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>>
>> For entire V3 and the additional patch 16/16,
>> Reviewed-by: Abner Chang <abner.chang@amd.com<mailto:abner.chang@amd.com>>
>>
>> I will do the Uncrustify check and merge this patch set once the corresponding
>> edk2 changes are merged.
>> Thanks
>> Abner
>>
>> > -----Original Message-----
>> > From: Konstantin Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>
>> > Sent: Monday, October 23, 2023 9:05 PM
>> > To: devel@edk2.groups.io<mailto:devel@edk2.groups.io>
>> > Cc: Chang, Abner <Abner.Chang@amd.com<mailto:Abner.Chang@amd.com>>; Attar, AbdulLateef (Abdul
>> > Lateef) <AbdulLateef.Attar@amd.com<mailto:AbdulLateef.Attar@amd.com>>; nicklew@nvidia.com<mailto:nicklew@nvidia.com>; Konstantin
>> > Aladyshev <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>>
>> > Subject: [PATCH edk2-platforms v3 00/16] MTCP-over-KCS support
>> >
>> > Caution: This message originated from an External Source. Use proper
>> caution
>> > when opening attachments, clicking links, or responding.
>> >
>> >
>> > The Manageability KCS transport library needs to support requests both
>> > from MCTP and IPMI transports. Currently the code only handles IPMI
>> > case correctly.
>> > In the MCTP case the communication should be based on the MCTP-over-
>> KCS
>> > specification (DSP0254). This specification defines a special KCS
>> > binding header and trailer structures that need to be present in every
>> > MCTP message.
>> > The header structure contains a length field, therefore response packet
>> > size is not needed to be known beforehand.
>> > The trailer structure contains a PEC checksum that can be used to check
>> > itegrity of the response message.
>> > Modify Manageability KCS transport library code to check which message
>> > is processed (IPMI or MCTP) and handle each case correctly based on its
>> > own specification.
>> > This patch is a result of a joint effort from the Konstantin Aladyshev
>> > <aladyshev22@gmail.com<mailto:aladyshev22@gmail.com>> and Abner Chang <abner.chang@amd.com<mailto:abner.chang@amd.com>>.
>> >
>> > Tested:
>> > PLDM communication between the HOST and BMC was tested with both
>> > components implemented via open-source software:
>> > - The HOST (UEFI firmware) part was based one the edk2 [1] and
>> > edk2-platforms [2] code,
>> > - The BMC part was based on the openbmc [3] distribution.
>> >
>> > The testing process and all the necessary utilities are described in
>> > the [4] repository.
>> >
>> > The provided changes keep IPMI over KCS stack working as reported by
>> > Abner Chang.
>> >
>> > [1]: https://github.com/tianocore/edk2
>> > [2]: https://github.com/tianocore/edk2-platforms
>> > [3]: https://github.com/openbmc/openbmc
>> > [4]: https://github.com/Kostr/PLDM
>> >
>> > Changes v2 -> v3:
>> > - Add new patch that adds PLDM completion code check
>> >
>> > Changes v1 -> v2:
>> > - Add new patches with corrections for the PLDM protocol. The
>> > resulting communication via EDKII_PLDM_PROTOCOL was successfully
>> > tested.
>> >
>> > Abner Chang (4):
>> > ManageabilityPkg: Add PLDM terminus PCDs
>> > PldmProtocolDxe: Correct TID argument usage
>> > ManageabilityPkg/PldmProtocol: Remove PLDM command table
>> > PldmSmbiosTransferDxe: Implement Set PLDM terminus ID API
>> >
>> > Konstantin Aladyshev (12):
>> > ManageabilityPkg: Add definition for the MCTP KCS TRAILER structure
>> > ManageabilityPkg: Check MCTP EIDs for reserved values
>> > ManageabilityPkg: Support both MCTP and IPMI in KCS tranport library
>> > ManageabilityPkg: Check header fields in the MCTP response
>> > ManageabilityPkg: Correct typo in MCTP destination EID field
>> > ManageabilityPkg: Update the algorithm of using MCTP endpoint ID PCD
>> > ManageabilityPkg: Correct value for the MCTP TAG_OWNER response bit
>> > ManageabilityPkg: Don't check MCTP header fields if transfer has
>> > failed
>> > ManageabilityPkg: Use correct constants for PLDM header checks
>> > ManageabilityPkg: Return error on multiple-packet MCTP responses
>> > ManageabilityPkg: Return error on PLDM header check fails
>> > ManageabilityPkg: Check PLDM completion code
>> >
>> > .../Include/Library/BasePldmProtocolLib.h | 16 +
>> > .../Library/ManageabilityTransportMctpLib.h | 9 +-
>> > .../Include/Protocol/MctpProtocol.h | 12 +-
>> > .../Include/Protocol/PldmProtocol.h | 18 +-
>> > .../Protocol/PldmSmbiosTransferProtocol.h | 26 ++
>> > .../Common/KcsCommon.c | 284 +++++++++++++++---
>> > .../Dxe/ManageabilityTransportMctp.c | 4 +-
>> > .../PldmProtocolLibrary/Dxe/PldmProtocolLib.c | 49 ++-
>> > .../Dxe/PldmProtocolLib.inf | 6 +-
>> > .../ManageabilityPkg/ManageabilityPkg.dec | 6 +
>> > .../MctpProtocol/Common/MctpProtocolCommon.c | 129 +++++++-
>> > .../Universal/MctpProtocol/Dxe/MctpProtocol.c | 51 +++-
>> > .../PldmProtocol/Common/PldmProtocolCommon.c | 148 +++------
>> > .../PldmProtocol/Common/PldmProtocolCommon.h | 25 +-
>> > .../Universal/PldmProtocol/Dxe/PldmProtocol.c | 69 ++++-
>> > .../PldmProtocol/Dxe/PldmProtocolDxe.inf | 4 -
>> > .../PldmSmbiosTransferDxe.c | 28 ++
>> > 17 files changed, 690 insertions(+), 194 deletions(-)
>> >
>> > --
>> > 2.34.1
>
>
>
>
>
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#110467): https://edk2.groups.io/g/devel/message/110467
Mute This Topic: https://groups.io/mt/102134646/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-
[-- Attachment #2: Type: text/html, Size: 30469 bytes --]
^ permalink raw reply [flat|nested] 25+ messages in thread