CASE spec refresh #8137

jpk233 · 2021-07-06T15:47:20Z

Problem

CASE does not currently follow the current spec with message handling.
Fixes CASE handshake implementation looks nothing like the spec #5934 (along with the previously submitted CASE TLV changes)
Fixes CASE session fails when commissioning second device in a row #8207

Change overview

Added TrustedRootId parameter to CASESession’s EstablishSession method
Added new methods to Find and Generate Destination IDs (New CASE Spec updates)
Updated some CASESession method’s parameters that were using raw input buffers with ByteSpan and MutableByteSpan objects
Updated CASESession Nonce and Info vectors (new CASE Spec updates)
Updated SigmaR1 message construction method to generate DestinationID instead of sending all TrustedRootIDs available.
Minor comment updates to match CASE Spec’s steps
Added new method to OperationalCredentialSet to retrieve RootCertificate indexed by TrustedRootId
CASESession Tests: Updated NodeID parameter to match Node ID of the corresponding Test Vector Certificate

Testing

Tested using CASE Session Unit Tests

todo · 2021-07-06T15:47:22Z

read fabricId

connectedhomeip/src/protocols/secure_channel/CASESession.cpp

Lines 1151 to 1161 in f2f7df4

    
           // TODO: read fabricId 
        
           uint64_t fabricId              = 0; 
        
           uint16_t destinationMessageLen = kSigmaParamRandomNumberSize + kP256_PublicKey_Length + sizeof(fabricId) + sizeof(nodeId); 
        
           System::PacketBufferHandle destinationMessage = System::PacketBufferHandle::New(destinationMessageLen); 
        
           const ChipCertificateData * rootCertificate   = mOpCredSet->GetRootCertificate(*trustedRootId); 
        
           VerifyOrReturnError(!destinationMessage.IsNull(), CHIP_ERROR_NO_MEMORY); 
        
           VerifyOrReturnError(rootCertificate != nullptr, CHIP_ERROR_CERT_NOT_FOUND); 
        
           VerifyOrReturnError(!rootCertificate->mPublicKey.empty(), CHIP_ERROR_INTERNAL); 
        
           VerifyOrReturnError(rootCertificate->mPublicKey.size() == kP256_PublicKey_Length, CHIP_ERROR_INTERNAL); 
        
           Encoding::LittleEndian::BufferWriter bbuf(destinationMessage->Start(), destinationMessageLen);

This comment was generated by todo based on a `TODO` comment in `f2f7df4` in #8137. cc @jpk233.

src/channel/Channel.h

src/protocols/secure_channel/CASESession.cpp

tcarmelveilleux

Crypto algorithm used for destinationId in CASE does not match spec.

src/protocols/secure_channel/CASESession.cpp

todo · 2021-07-08T21:01:39Z

implement API for this in OperationalCredentialSet

connectedhomeip/src/protocols/secure_channel/CASESession.cpp

Lines 268 to 272 in 21f624c

    
           // TODO: implement API for this in OperationalCredentialSet 
        
           mTrustedRootId = operationalCredentialSet->GetOpCredSet()[opCredSetIndex].GetCertSet()[0].mAuthKeyId; 
        
           err = SendSigmaR1(); 
        
           SuccessOrExit(err);

This comment was generated by todo based on a `TODO` comment in `21f624c` in #8137. cc @jpk233.

todo · 2021-07-08T21:01:42Z

Parameterize fabricId instead of hardcoding it to 0

connectedhomeip/src/protocols/secure_channel/CASESession.cpp

Lines 331 to 341 in 21f624c

    
               // TODO: Parameterize fabricId instead of hardcoding it to 0 
        
               FabricId fabricId = 0; 
        
               // Generate an ephemeral keypair 
        
           #ifdef ENABLE_HSM_CASE_EPHERMAL_KEY 
        
               mEphemeralKey.SetKeyId(CASE_EPHEMERAL_KEY); 
        
           #endif 
        
               ReturnErrorOnFailure(mEphemeralKey.Initialize()); 
        
               // Fill in the random value 
        
               ReturnErrorOnFailure(DRBG_get_bytes(initiatorRandom, kSigmaParamRandomNumberSize));

This comment was generated by todo based on a `TODO` comment in `21f624c` in #8137. cc @jpk233.

todo · 2021-07-08T21:01:45Z

Parameterize fabricId instead of hardcoding it to 0

connectedhomeip/src/protocols/secure_channel/CASESession.cpp

Lines 1119 to 1129 in 21f624c

    
           // TODO: Parameterize fabricId instead of hardcoding it to 0 
        
           FabricId fabricId = 0; 
        
           trustedRootId = mOpCredSet->GetTrustedRootId(i); 
        
           ReturnErrorOnFailure(DecodeChipCert(mOpCredSet->GetDevOpCred(trustedRootId), mOpCredSet->GetDevOpCredLen(trustedRootId), 
        
                                               nodeOperationalCertificate)); 
        
           ReturnErrorOnFailure(nodeOperationalCertificate.mSubjectDN.GetCertChipId(nodeId)); 
        
           // TODO: add method to retrieve FabricId. Call it here

This comment was generated by todo based on a `TODO` comment in `21f624c` in #8137. cc @jpk233.

todo · 2021-07-08T21:01:48Z

add method to retrieve FabricId. Call it here

connectedhomeip/src/protocols/secure_channel/CASESession.cpp

Lines 1128 to 1138 in 21f624c

    
           // TODO: add method to retrieve FabricId. Call it here 
        
           { 
        
               MutableByteSpan candidateSpan(candidate, sizeof(candidate)); 
        
               ReturnErrorOnFailure(GenerateDestinationID(initiatorRandom, &trustedRootId, nodeId, fabricId, candidateSpan)); 
        
           } 
        
           if (memcmp(destinationId.data(), candidate, destinationId.size()) == 0) 
        
           { 
        
               mTrustedRootId = trustedRootId; 
        
               break;

This comment was generated by todo based on a `TODO` comment in `21f624c` in #8137. cc @jpk233.

src/protocols/secure_channel/CASESession.cpp

src/controller/CHIPDevice.h

src/protocols/secure_channel/CASESession.h

src/transport/AdminPairingTable.cpp

src/protocols/secure_channel/tests/TestCASESession.cpp

src/credentials/CHIPCert.h

src/protocols/secure_channel/CASESession.cpp

src/controller/CHIPDevice.h

src/transport/FabricTable.cpp

src/transport/FabricTable.h

src/credentials/CHIPCert.cpp

Update GetCertFabricId method to use UINT64_MAX macro as an invalid reference, and added a final sanity check to see if there were actually any fabricIds present in the certificate. Added doxygen to GetCredentials method.

github-actions · 2021-07-28T02:01:21Z

Size increase report for "esp32-example-build" from 5d91959

File	Section	File	VM
chip-ipv6only-app.elf	.flash.text	172	172
chip-all-clusters-app.elf	.flash.text	632	632
chip-all-clusters-app.elf	.dram0.data	8	16
chip-all-clusters-app.elf	.dram0.heap_start	0	-8
chip-all-clusters-app.elf	.flash.rodata	-8	-8
chip-all-clusters-app.elf	.dram0.bss	0	-56
chip-shell.elf	.flash.text	24	24
chip-lock-app.elf	.flash.text	400	400
chip-lock-app.elf	[2 Others]	-8	-8
chip-lock-app.elf	.dram0.bss	0	-56

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

Comparing ./master_artifact/chip-ipv6only-app.elf and ./pull_artifact/chip-ipv6only-app.elf:

sections,vmsize,filesize
.flash.text,172,172
[Unmapped],0,-172

Comparing ./master_artifact/chip-all-clusters-app.elf and ./pull_artifact/chip-all-clusters-app.elf:

sections,vmsize,filesize
.debug_info,0,12064
.debug_loc,0,2027
.debug_str,0,854
.debug_ranges,0,832
.strtab,0,790
.debug_line,0,665
.flash.text,632,632
.symtab,0,192
.debug_frame,0,156
.debug_aranges,0,40
.debug_abbrev,0,21
.dram0.data,16,8
.riscv.attributes,0,1
.shstrtab,0,-2
.dram0.heap_start,-8,0
.flash.rodata,-8,-8
[ELF Headers],0,-40
.dram0.bss,-56,0
[Unmapped],0,-632

Comparing ./master_artifact/chip-shell.elf and ./pull_artifact/chip-shell.elf:

sections,vmsize,filesize
.debug_info,0,1434
.debug_str,0,274
.debug_line,0,158
.debug_loc,0,143
.debug_ranges,0,72
.debug_abbrev,0,39
.flash.text,24,24
.shstrtab,0,-2
.strtab,0,-2
[Unmapped],0,-24

Comparing ./master_artifact/chip-persistent-storage.elf and ./pull_artifact/chip-persistent-storage.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-temperature-measurement-app.elf and ./pull_artifact/chip-temperature-measurement-app.elf:

BLOAT EXECUTION FAILED WITH CODE 1:
bloaty: integer overflow

Comparing ./master_artifact/chip-lock-app.elf and ./pull_artifact/chip-lock-app.elf:

sections,vmsize,filesize
.debug_info,0,11100
.debug_loc,0,5867
.debug_line,0,1566
.debug_ranges,0,976
.debug_str,0,854
.strtab,0,790
.flash.text,400,400
.shstrtab,0,322
.symtab,0,304
.xt.prop._ZNK4chip4SpanIhE7SubSpanEjj,0,140
.debug_frame,0,120
.xt.prop._ZNK4chip11CASESession10GetIPKListEv,0,88
.debug_abbrev,0,81
.xt.prop._ZN4chip10CASEServer10GetSessionEv,0,76
.xt.prop._ZNK4chip11CASESession17GetIPKListEntriesEv,0,76
.xt.prop._ZZNK4chip11CASESession10GetIPKListEvE11ipkListSpan,0,52
.xt.lit._ZNK4chip11CASESession10GetIPKListEv,0,48
.debug_aranges,0,32
[2 Others],-8,-8
.dram0.bss,-56,0
[Unmapped],0,-392

Comparing ./master_artifact/chip-pigweed-app.elf and ./pull_artifact/chip-pigweed-app.elf:

sections,vmsize,filesize

github-actions · 2021-07-28T02:01:40Z

Size increase report for "nrfconnect-example-build" from 5d91959

File	Section	File	VM
chip-shell.elf	text	56	56
chip-shell.elf	[LOAD #3 [RW]]	0	20
chip-shell.elf	device_handles	8	8
chip-shell.elf	bss	0	-20
chip-lock.elf	text	436	436
chip-lock.elf	datas	8	8
chip-lock.elf	device_handles	-4	-4
chip-lock.elf	rodata	-8	-8
chip-lock.elf	[LOAD #3 [RW]]	0	-20
chip-lock.elf	bss	0	-44

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

Comparing ./master_artifact/chip-shell.elf and ./pull_artifact/chip-shell.elf:

sections,vmsize,filesize
.debug_info,0,1508
.debug_loc,0,471
.debug_str,0,261
.debug_ranges,0,72
.debug_line,0,64
text,56,56
[LOAD #3 [RW]],20,0
.debug_abbrev,0,12
device_handles,8,8
.debug_frame,0,4
.shstrtab,0,-1
bss,-20,0
.symtab,0,-32
.strtab,0,-35

Comparing ./master_artifact/chip-lock.elf and ./pull_artifact/chip-lock.elf:

sections,vmsize,filesize
.debug_info,0,10156
.debug_loc,0,5702
.strtab,0,865
.debug_line,0,809
.debug_str,0,786
text,436,436
.debug_ranges,0,408
.symtab,0,336
.debug_frame,0,148
.debug_aranges,0,48
datas,8,8
.debug_abbrev,0,-1
.shstrtab,0,-1
device_handles,-4,-4
rodata,-8,-8
[LOAD #3 [RW]],-20,0
bss,-44,0

github-actions · 2021-07-28T02:02:15Z

Size increase report for "gn_qpg-example-build" from 5d91959

File	Section	File	VM
chip-qpg6100-lighting-example.out	.text	428	428
chip-qpg6100-lighting-example.out	.heap	0	48
chip-qpg6100-lighting-example.out	.data	8	8
chip-qpg6100-lighting-example.out	.bss	0	-56

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

Comparing ./master_artifact/chip-qpg6100-lighting-example.out and ./pull_artifact/chip-qpg6100-lighting-example.out:

sections,vmsize,filesize
.debug_info,0,10180
.debug_loc,0,5065
.strtab,0,855
.debug_str,0,796
.debug_line,0,742
.text,428,428
.symtab,0,336
.debug_frame,0,144
.debug_ranges,0,136
.debug_abbrev,0,49
.debug_aranges,0,48
.heap,48,0
.data,8,8
.shstrtab,0,1
.bss,-56,0
[Unmapped],0,-428

Comparing ./master_artifact/chip-qpg6100-lighting-example.out.map and ./pull_artifact/chip-qpg6100-lighting-example.out.map:

BLOAT EXECUTION FAILED WITH CODE 1:
bloaty: unknown file type for file './pull_artifact/chip-qpg6100-lighting-example.out.map'

* CASE spec refresh * Remove TrustedRootID from Sigma messages. Replce HKDF algorithm with HMAC-SHA256 Algorithm in GenerateDestinationID method. * Replace TrustedRootId parameter with an index value for the OperationalCredentialSet. Added CredentialsIndex parameter to AdminPairingTable's GetCredentials method. Removed FabricSecret. Removed kIPKInfo. Removed Deprecated ComputeIPK method - Replaced it with dummy RetrieveIPK method. * cast 1 to uint8_t to avoid compiler errors on other platforms * Fix uint8_t conversion * Replace CASETLVTags with TLV::ContextTag ID, switch unsafe statement to an actual OperationalCredentialSet method to retrieve a TrustedRootId, renamed kMAX_Hash_Length refs to kSHA256_Hash_Length to match the Spec. * Add GenerateDestinationID Test * Added ReleaseLastCert method to OperationalCredentialSet Class. This method will release the last certificate data in the set. Added call to ReleaseLastCert method during CASESession after a successful validation of the NOC certificate. Updated CASE Unit Tests to force 3 certificates maximum. This will guarantee that for the tests to work fine, CASESession must release the NOC certificate every time. CASESession: change LoadCerts to LoadCert - Only NOC is transferred during CASE Protocol. No need to handle ICA * remove fabricId parameters/methods from CASESession. Retrieve it from the NOC instead. * Updated ReleaseLastCert method from CHIPCertificateSet class: not using a const ChipCertificateData type anymore to avoid confusion. Removed redundant comment from ReleaseLastCert method. Wrote some TODO items to update OperationalCredentialSet class in order to work with size_t variable counters: useful to index more than 255 Credentials. Update DestinationIDGeneration Test to use Spec's test vectors. Added static assert to check TBEData2 and TBEData3 Nonce Lengths : they must match. Added new method to Estimate TLV Struct overhead. Updated GenerateDestinationID to be stateless: now directly accesses the inputs as raw memory buffers. Updated HandleSigma methods to handle TLV tags sequentially. Removed redundant GetLength and GetType calls during TLV Reads. Renamed encryptionKeyId to initiatorSessionId and responderSessionId. Fixed typo in ENABLE_HSM_CASE_EPHEMERAL_KEY macro. * Update casesession with latest comments * Trigger Build * added IPK to CASESession Serializable data: IPK is needed to DeriveSecureSession, so it needs to be stored. Updated FindDestinationIDCandidate loop variables to size_t. Also updated names to reflect what they are indexing. * Added CASESession protected API (Virtual) to get the IPK List Entries. They can be overridden by the Unit Tests in order to feed in the test IPK vectors. * Update for fabric class rename * Restyling * Replace mCredentialsIndex pointer to an actual uint8_t variable. Update GetCertFabricId method to use UINT64_MAX macro as an invalid reference, and added a final sanity check to see if there were actually any fabricIds present in the certificate. Added doxygen to GetCredentials method. * Restyle fix

boring-cyborg bot added the controller label Jul 6, 2021

pullapprove bot requested review from chrisdecenzo, jelderton, jmartinez-silabs, LuDuda, mspang, saurabhst, vivien-apple and woody-apple July 6, 2021 15:47

pullapprove bot added the review - pending label Jul 6, 2021