Some conversions to use PacketBufferHandle #4011

kpschoedel · 2020-11-27T15:06:22Z

Problem

Code should use PacketBufferHandle rather than PacketBuffer *.

Summary of Changes

Converts most receive path code.
Converts most of //src/ble.
Introduces Handle versions of the AddToEnd/DetachTail pair.

Part of issue #2707 - Figure out a way to express PacketBuffer ownership in the type system

#### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system

kpschoedel · 2020-11-27T15:10:30Z

Roadmap: Next PR will cover the send path and substantially complete conversion, followed by cleanup of transitional code, and possible optimizations.

kpschoedel · 2020-11-27T15:30:04Z

src/ble/BLEEndPoint.cpp

-    //
-    // Beware this line should we ever use virtuals in this class or its
-    // super(s). See similar lines in chip::System::Layer end points.
-    memset(this, 0, sizeof(*this));


All members are in fact explicitly initialized below. (The change to PacketBufferHandle makes the compiler reject this, although it was already non-conforming anway.)

mspang · 2020-12-01T00:23:28Z

@kpschoedel conflict

bzbarsky-apple · 2020-12-01T05:22:16Z

src/ble/BLEEndPoint.cpp

-    buf = mBtpEngine.TxPacket();
-    mBtpEngine.ClearTxPacket();
-    PacketBuffer::Free(buf);
+    (void) mBtpEngine.TakeTxPacket();


I wonder whether we should just leave a ClearTxPacket API like we used to have. Seems more idiomatic, and probably less code, than the "take but ignore return value" pattern here...

Yes, it also touched mTxState. Not sure what I was thinking here.

bzbarsky-apple · 2020-12-01T05:22:28Z

src/ble/BLEEndPoint.cpp

-    buf = mBtpEngine.RxPacket();
-    mBtpEngine.ClearRxPacket();
-    PacketBuffer::Free(buf);
+    (void) mBtpEngine.TakeRxPacket();


Similar for ClearRxPacket

bzbarsky-apple · 2020-12-01T05:23:58Z

src/ble/BLEEndPoint.cpp

@@ -635,7 +629,7 @@ BLE_ERROR BLEEndPoint::SendCharacteristic(PacketBuffer * buf)
 *  kType_Data(0)       - data packet
 *  kType_Control(1)    - control packet
 */
-void BLEEndPoint::QueueTx(PacketBuffer * data, PacketType_t type)
+void BLEEndPoint::QueueTx(PacketBufferHandle data, PacketType_t type)


This one probably could use PacketBufferHandle && to reduce codesize, since it never takes early returns.... Not sure whether we want to do that.

I've been planning to check all PacketBufferHandle arguments as part of a final polish pass.

bzbarsky-apple · 2020-12-01T05:31:51Z

src/ble/BLEEndPoint.cpp

@@ -928,7 +905,7 @@ BLE_ERROR BLEEndPoint::HandleFragmentConfirmationReceived()
    // the stand-alone ack, and also the case where a window size < the immediate ack threshold was detected in
    // Receive(), but the stand-alone ack was deferred due to a pending outbound message fragment.
    if (mLocalReceiveWindowSize <= BLE_CONFIG_IMMEDIATE_ACK_WINDOW_THRESHOLD &&
-        !(mSendQueue != nullptr || mBtpEngine.TxState() == BtpEngine::kState_InProgress))
+        !(!mSendQueue.IsNull() || mBtpEngine.TxState() == BtpEngine::kState_InProgress))


I wonder whether it might be better to do:

Suggested change

!(!mSendQueue.IsNull() || mBtpEngine.TxState() == BtpEngine::kState_InProgress))

mSendQueue.IsNull() && mBtpEngine.TxState() != BtpEngine::kState_InProgress)

bzbarsky-apple · 2020-12-01T05:40:46Z

src/darwin/Framework/CHIP/CHIPDeviceStatusDelegateBridge.mm

 {
    NSLog(@"DeviceStatusDelegate received message from the device");

    size_t data_len = message->DataLength();
    // convert to NSData
    NSMutableData * dataBuffer = [[NSMutableData alloc] initWithBytes:message->Start() length:data_len];
-    message = message->Next();
+    (void) message.DetachTail();


This seems like it could just be message.FreeHead(), no? That's what the old code was doing, kind of, except for forgetting to free the buffer it was skipping...

bzbarsky-apple · 2020-12-01T06:01:13Z

src/platform/Darwin/BlePlatformDelegateImpl.mm

            return false;
        }

        bool BlePlatformDelegateImpl::SendWriteRequest(
-            BLE_CONNECTION_OBJECT connObj, const ChipBleUUID * svcId, const ChipBleUUID * charId, PacketBuffer * pBuf)
+            BLE_CONNECTION_OBJECT connObj, const ChipBleUUID * svcId, const ChipBleUUID * charId, PacketBufferHandle pBuf)


There are some NULL != pBuf bits in this function that need to become !pBuf.IsNull()

bzbarsky-apple · 2020-12-01T06:02:58Z

src/platform/ESP32/bluedroid/BLEManagerImpl.cpp

-            memset(&mCons[freeIndex], 0, sizeof(CHIPoBLEConState));
-            mCons[freeIndex].Allocated = 1;
-            mCons[freeIndex].ConId     = conId;
+            mCons[freeIndex].PendingIndBuf = nullptr;


I wonder whether CHIPoBLEConState should grow a Reset function... Followup perhaps?

Added to TODO list. Every Impl has its own version…

bzbarsky-apple · 2020-12-01T06:06:54Z

src/system/SystemPacketBuffer.cpp

@@ -707,5 +712,19 @@ PacketBuffer * PacketBuffer::BuildFreeList()

 #endif //  !CHIP_SYSTEM_CONFIG_USE_LWIP && CHIP_SYSTEM_CONFIG_PACKETBUFFER_MAXALLOC

+PacketBufferHandle PacketBufferHandle::DetachTail()


So this has quite different semantics from PacketBuffer::DetachTail: it returns the old head, not the old tail, and updates the identity of this rather than keeping that identity.

It seems to me that this basically does what FreeHead does, but instead of freeing the head it returns it. Maybe it should be called TakeHead or PopHead or some such? Calling it DetachTail really doesn't make sense, since the thing it's being called on becomes what used to be the tail.

Changed to PopHead().

bzbarsky-apple · 2020-12-01T06:07:51Z

src/system/SystemPacketBuffer.h

+     *  Detach and return the head of a buffer chain while updating this handle to point to the remaining buffers.
+     *  The current buffer must be the head of the chain.
+     *
+     *  @return the detached bufferly formerly at the head of the buffer chain.


Suggested change

* @return the detached bufferly formerly at the head of the buffer chain.

* @return the detached buffer formerly at the head of the buffer chain.

bzbarsky-apple · 2020-12-01T06:11:38Z

src/transport/raw/TCP.cpp

@@ -342,7 +342,7 @@ CHIP_ERROR TCPBase::ProcessReceivedBuffer(Inet::TCPEndPoint * endPoint, const Pe
        // when a buffer is empty, it can be released back to the app
        if (buffer->DataLength() == 0)
        {
-            buffer.Adopt(buffer->DetachTail());
+            (void) buffer.DetachTail();


Again, seems like this is doing a FreeHead?

* revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos

bzbarsky-apple · 2020-12-01T17:41:21Z

src/ble/BtpEngine.h

-    bool ClearRxPacket();
-    bool ClearTxPacket();
+    void ClearRxPacket();
+    PacketBufferHandle TakeRxPacket() { return std::move(mRxBuf); }


It seems like calling this requires a call to ClearRxPacket right afterward, right? And similar for TakeTxPacket? Maybe we should make those calls directly from inside this function instead of forcing consumers to do both....

Or perhaps what should happen is that TakeRxPacket should do all the state-maintenance and ClearRxPacket should just TakeRxPacket and ignore the return, or stick it into a stack var that then goes out of scope or something.

It would be simpler. The reason I left it is that there is one instance in BLEEndPoint::DriveSending() where mBtpEngineTest.DoTxTiming() uses the taken packet, and that CHIP_ENABLE_CHIPOBLE_TEST code currently isn't in tree so I can't tell whether it also depends on mTxState, and I don't know whether someone intends to import that test code.

Hmm. I don't know that this DoTxTiming thing is even going to exist in CHIP. That said, the Weave version doesn't depend on the engine's mTxState (note that it's being called on a different object: the test object, not the engine object). So I think it's safe to make this change. I'm OK with it being in this PR or in a followup, whichever works better for you.

OK. I'll add it to the next PR.

github-actions · 2020-12-01T18:47:36Z

Size increase report for "esp32-example-build" from 01031e7

File	Section	File	VM
chip-all-clusters-app.elf	.flash.text	348	348
chip-all-clusters-app.elf	.flash.rodata	4	4

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

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

sections,vmsize,filesize
.debug_info,0,7363
.debug_line,0,3301
.debug_loc,0,1173
.debug_ranges,0,1016
.debug_str,0,713
.debug_abbrev,0,364
.flash.text,348,348
.strtab,0,269
.debug_frame,0,120
.xt.prop._ZN4chip11DeviceLayer8Internal14BLEManagerImplD5Ev,0,52
.xt.lit._ZN4chip11DeviceLayer8Internal14BLEManagerImplD5Ev,0,48
.debug_aranges,0,40
.symtab,0,32
.flash.rodata,4,4
.xt.prop._ZN4chip8Encoding12LittleEndian7Write32ERPhj,0,2
.shstrtab,0,-1
[Unmapped],0,-4
.xt.prop._ZN4chip11DeviceLayer8Internal14BLEManagerImplD2Ev,0,-80

github-actions · 2020-12-01T18:48:05Z

Size increase report for "nrfconnect-example-build" from 01031e7

File	Section	File	VM
chip-lighting.elf	text	412	412
chip-lighting.elf	rodata	8	12
chip-lighting.elf	shell_root_cmds_sections	-12	-12
chip-shell.elf	text	348	348
chip-shell.elf	rodata	16	12
chip-shell.elf	log_const_sections	4	4
chip-lock.elf	text	412	412
chip-lock.elf	rodata	16	12
chip-lock.elf	shell_root_cmds_sections	-12	-12

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

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

sections,vmsize,filesize
.debug_info,0,6876
.debug_line,0,1755
.debug_loc,0,922
.debug_abbrev,0,805
.debug_ranges,0,680
.debug_str,0,438
text,412,412
.strtab,0,214
.debug_frame,0,148
.symtab,0,64
.debug_aranges,0,40
rodata,12,8
.shstrtab,0,-2
shell_root_cmds_sections,-12,-12

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

sections,vmsize,filesize
.debug_info,0,5046
.debug_line,0,1586
.debug_loc,0,963
.debug_abbrev,0,700
.debug_ranges,0,608
.debug_str,0,393
text,348,348
.strtab,0,164
.debug_frame,0,120
.symtab,0,64
.debug_aranges,0,32
rodata,12,16
log_const_sections,4,4

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

sections,vmsize,filesize
.debug_info,0,6974
.debug_line,0,1757
.debug_loc,0,926
.debug_abbrev,0,829
.debug_ranges,0,680
.debug_str,0,438
text,412,412
.strtab,0,214
.debug_frame,0,148
.symtab,0,64
.debug_aranges,0,40
rodata,12,16
.shstrtab,0,2
shell_root_cmds_sections,-12,-12

#### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts send paths and most other remaining code. - Adds some PacketBufferHandle convenience functions. - Removes AddToEnd_ForNow() and DetachTail_ForNow(). Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system

Encapsulate setting ESP32's BLEManagerImpl::CHIPoBLEConState. (Requested in review comments on PR project-chip#4011. Other existing `BLEManagerImpl`s don't use this pattern.)

Encapsulate setting ESP32's BLEManagerImpl::CHIPoBLEConState. (Requested in review comments on PR #4011. Other existing `BLEManagerImpl`s don't use this pattern.)

* Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io>

* RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (#3990) #### Problem Some conversions to use PacketBufferHandle (#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in #3909. * Add '-Wextra' to compiler flags (#3902) * Implement Level Control Cluster (#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (#4012) PR #3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (#4008) * Fix data loss or crash in TCPEndPoint with LwIP (#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes #4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (#4010) * Fix segmentation fault error in response echo message (#3984) * Add back Android default build coverage & fix the build (#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in #3340. * Update all-clusters-app gen/ folder with ZAP generated content (#3963) * Move src/inet/tests to auto-test-driver generation (#3997) * Rename TestUtils to UnitTestRegistration. (#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (#3991) * Cleanup zap chip-helper.js (#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (#3979) * Implement the missing part of Exchange Header in Transport layer (#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue #2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (#4032) * Add SSID and password to chip-tool pairing (#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (#4039) #### Problem PR #3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR #4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since #3704 and this PR should fix that. Also I have a PR (mostly) ready once #3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * fix compilation problem * fix minor diffs * cleanup rotating id in ble * fix dns * nits * fix compilation * Revert "fix minor diffs" This reverts commit 88fb69c. * nits * nit fixes * update allocation * update allocation * refactoring * revert settings.json * fix styling * Update README file * adding a build flag to bypass advertising the additional data field * fix styling * fixing README style * Fixing nits + added description about the additional data * fixed minor comment * remove ParseCommand * removing unused headers + nits * Fixing Additional data TLV Tags * Fix Styling * removed unnecessary logging tag * fixing writing rotating id * fixing minor styles * adding check for rotating id tag * restyling * update AdditionalDataPayloadParser to work on vector of bytes * restyling * Nit fix * Renaming QRCode command to SetupPayload command * update parse command for additional data * change compile flags * update some comments * fixing parsing/generating the additional data payload * changing the additional parser APIs * restyling * reverting pigweed repo changes * rename CHIP_ENABLE_ADDITIONAL_ADVERTISING * removing logging tag * adding extra validation in parser * adding docs to the additional data payload parser * Update ReadME.md Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com>

* RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (project-chip#3990) #### Problem Some conversions to use PacketBufferHandle (project-chip#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in project-chip#3909. * Add '-Wextra' to compiler flags (project-chip#3902) * Implement Level Control Cluster (project-chip#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (project-chip#4012) PR project-chip#3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (project-chip#4008) * Fix data loss or crash in TCPEndPoint with LwIP (project-chip#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes project-chip#4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (project-chip#4010) * Fix segmentation fault error in response echo message (project-chip#3984) * Add back Android default build coverage & fix the build (project-chip#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in project-chip#3340. * Update all-clusters-app gen/ folder with ZAP generated content (project-chip#3963) * Move src/inet/tests to auto-test-driver generation (project-chip#3997) * Rename TestUtils to UnitTestRegistration. (project-chip#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (project-chip#3991) * Cleanup zap chip-helper.js (project-chip#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (project-chip#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (project-chip#3979) * Implement the missing part of Exchange Header in Transport layer (project-chip#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (project-chip#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (project-chip#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (project-chip#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (project-chip#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (project-chip#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (project-chip#4032) * Add SSID and password to chip-tool pairing (project-chip#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (project-chip#4039) #### Problem PR project-chip#3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR project-chip#4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since project-chip#3704 and this PR should fix that. Also I have a PR (mostly) ready once project-chip#3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (project-chip#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (project-chip#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (project-chip#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * fix compilation problem * fix minor diffs * cleanup rotating id in ble * fix dns * nits * fix compilation * Revert "fix minor diffs" This reverts commit 88fb69c. * nits * nit fixes * update allocation * update allocation * refactoring * revert settings.json * fix styling * Update README file * adding a build flag to bypass advertising the additional data field * fix styling * fixing README style * Fixing nits + added description about the additional data * fixed minor comment * remove ParseCommand * removing unused headers + nits * Fixing Additional data TLV Tags * Fix Styling * removed unnecessary logging tag * fixing writing rotating id * fixing minor styles * adding check for rotating id tag * restyling * update AdditionalDataPayloadParser to work on vector of bytes * restyling * Nit fix * Renaming QRCode command to SetupPayload command * update parse command for additional data * change compile flags * update some comments * fixing parsing/generating the additional data payload * changing the additional parser APIs * restyling * reverting pigweed repo changes * rename CHIP_ENABLE_ADDITIONAL_ADVERTISING * removing logging tag * adding extra validation in parser * adding docs to the additional data payload parser * Update ReadME.md Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com>

…aligned with the spec (#9455) * RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (#3990) #### Problem Some conversions to use PacketBufferHandle (#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in #3909. * Add '-Wextra' to compiler flags (#3902) * Implement Level Control Cluster (#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (#4012) PR #3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (#4008) * Fix data loss or crash in TCPEndPoint with LwIP (#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes #4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (#4010) * Fix segmentation fault error in response echo message (#3984) * Add back Android default build coverage & fix the build (#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in #3340. * Update all-clusters-app gen/ folder with ZAP generated content (#3963) * Move src/inet/tests to auto-test-driver generation (#3997) * Rename TestUtils to UnitTestRegistration. (#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (#3991) * Cleanup zap chip-helper.js (#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (#3979) * Implement the missing part of Exchange Header in Transport layer (#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue #2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (#4032) * Add SSID and password to chip-tool pairing (#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (#4039) #### Problem PR #3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR #4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since #3704 and this PR should fix that. Also I have a PR (mostly) ready once #3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * Revert "Merge pull request #4 from hnnajh/rotating-id-test" This reverts commit 0235d05, reversing changes made to 3e1a4b9. * Storing RI in Octet String + Adding Binary format for BLE * Fixing rotating id parser + adding unittests * restyling * refactoring rotating id unit tests * Added more unit tests for Rotating Device Id * updated styling * refactor RI tests * Added RI Unittest + more validation * applying restyling * Fix CI * update styling * Fix CI * Update Styling * Fix CI * Restyling * Fixing nits * Using MutableByteSpan in RI generation * Fixing nits Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com>

…aligned with the spec (project-chip#9455) * RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (project-chip#3990) #### Problem Some conversions to use PacketBufferHandle (project-chip#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in project-chip#3909. * Add '-Wextra' to compiler flags (project-chip#3902) * Implement Level Control Cluster (project-chip#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (project-chip#4012) PR project-chip#3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (project-chip#4008) * Fix data loss or crash in TCPEndPoint with LwIP (project-chip#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes project-chip#4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (project-chip#4010) * Fix segmentation fault error in response echo message (project-chip#3984) * Add back Android default build coverage & fix the build (project-chip#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in project-chip#3340. * Update all-clusters-app gen/ folder with ZAP generated content (project-chip#3963) * Move src/inet/tests to auto-test-driver generation (project-chip#3997) * Rename TestUtils to UnitTestRegistration. (project-chip#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (project-chip#3991) * Cleanup zap chip-helper.js (project-chip#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (project-chip#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (project-chip#3979) * Implement the missing part of Exchange Header in Transport layer (project-chip#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (project-chip#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (project-chip#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (project-chip#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (project-chip#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (project-chip#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (project-chip#4032) * Add SSID and password to chip-tool pairing (project-chip#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (project-chip#4039) #### Problem PR project-chip#3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR project-chip#4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since project-chip#3704 and this PR should fix that. Also I have a PR (mostly) ready once project-chip#3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (project-chip#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (project-chip#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (project-chip#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * Revert "Merge pull request #4 from hnnajh/rotating-id-test" This reverts commit 0235d05, reversing changes made to 3e1a4b9. * Storing RI in Octet String + Adding Binary format for BLE * Fixing rotating id parser + adding unittests * restyling * refactoring rotating id unit tests * Added more unit tests for Rotating Device Id * updated styling * refactor RI tests * Added RI Unittest + more validation * applying restyling * Fix CI * update styling * Fix CI * Update Styling * Fix CI * Restyling * Fixing nits * Using MutableByteSpan in RI generation * Fixing nits Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com>

* RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (#3990) #### Problem Some conversions to use PacketBufferHandle (#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in #3909. * Add '-Wextra' to compiler flags (#3902) * Implement Level Control Cluster (#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (#4012) PR #3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (#4008) * Fix data loss or crash in TCPEndPoint with LwIP (#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes #4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (#4010) * Fix segmentation fault error in response echo message (#3984) * Add back Android default build coverage & fix the build (#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in #3340. * Update all-clusters-app gen/ folder with ZAP generated content (#3963) * Move src/inet/tests to auto-test-driver generation (#3997) * Rename TestUtils to UnitTestRegistration. (#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (#3991) * Cleanup zap chip-helper.js (#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (#3979) * Implement the missing part of Exchange Header in Transport layer (#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue #2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (#4032) * Add SSID and password to chip-tool pairing (#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (#4039) #### Problem PR #3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR #4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since #3704 and this PR should fix that. Also I have a PR (mostly) ready once #3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * Revert "Merge pull request #4 from hnnajh/rotating-id-test" This reverts commit 0235d05, reversing changes made to 3e1a4b9. * increment lifetime counter * Incrementing lifetime counter at each BLE/MDNS advertisement * address comments * Update src/include/platform/ConfigurationManager.h Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> * removing unnecessary lines * nit * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * restyling * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * fix comments * incrementing lifetimecounter once * remove unnecessary include Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com>

* RotatingId: version0 * RotatingId: version1 * RotatingId: version0 * RotatingId: version1 * Fix Darwin host build (project-chip#3990) #### Problem Some conversions to use PacketBufferHandle (project-chip#3909) broke Darwin builds, which aren't currently run in CI. #### Summary of Changes Fix src/platform/Darwin/BleConnectionDelegateImpl.mm to match the API change in project-chip#3909. * Add '-Wextra' to compiler flags (project-chip#3902) * Implement Level Control Cluster (project-chip#3806) * New seekbar in Android CHIPTool * Sample usage in lighting-app/nrfconnect Signed-off-by: Markus Becker <markus.becker@tridonic.com> * Fix Rendezvous over BLE after recent changes (project-chip#4012) PR project-chip#3704 introduced a change that the BLE transport in RendezvousSession is only initialized when PeerAddress in RendezvousParams is of type BLE. However, PeerAddress isn't initialized anywhere. As a result Rendezvous over BLE stopped working between Android CHIPTool and accessories. Btw, remove an assert related to the storage delegate as it seems an optional member of the device controller. * [thread] fix invalid configuration of active dataset (project-chip#4008) * Fix data loss or crash in TCPEndPoint with LwIP (project-chip#4022) * Fix data loss or crash in TCPEndPoint with LwIP #### Problem Under the configuration CHIP_SYSTEM_CONFIG_USE_LWIP, in some cases where the data size exceeds the TCP window size, TCPEndPoint can either die or lose data when accounting of un-acked data falls out of sync. #### Summary of Changes Imported fix from Weave: * This change removes separate accounting of the unsent data position and replaces it with simple counting of sent-but-not-acked data and a skip-loop at the start of DriveSending(). Fixes project-chip#4013 - Data loss or crash in TCPEndPoint with LwIP * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Update lighting-app gen/ folder with ZAP generated content (project-chip#4010) * Fix segmentation fault error in response echo message (project-chip#3984) * Add back Android default build coverage & fix the build (project-chip#3966) mDNS doesn't build with no device layer. Remove it from the build and add back the coverage that would catch this that was lost in project-chip#3340. * Update all-clusters-app gen/ folder with ZAP generated content (project-chip#3963) * Move src/inet/tests to auto-test-driver generation (project-chip#3997) * Rename TestUtils to UnitTestRegistration. (project-chip#4021) Looking to remove usage of 'Utils' unless we have really no choice. 'UnitTestRegistration' seems clearer in what it does compared to 'TestUtils'. * Update src/lib/core/tests to auto-test-driver generation (project-chip#3991) * Cleanup zap chip-helper.js (project-chip#3973) * Cleanup zap chip-helper.js * Restyled by clang-format Co-authored-by: Restyled.io <commits@restyled.io> * Move src/transport/tests to auto-test-driver generation (project-chip#3999) * Move src/transport/tests to auto-test-driver generation * Add relevant libraries (and more test-capable libs) to nrf. * Refactor inet test helpers (to not include actual inet tests), try to make qemu allow better linkage but still failed for transport tests so disabled for now * Added more tests on esp32 qemu * Restyle fixes * Fix cast errors in InetCommon * Disable raw tests from zephyr: somehow they fail running out of endpoints * Disable DNS test on zephyr * Remove inet endpoint test from zephyr * Remove inet endpoint test from zephyr - fix again * Modify gitignore * Restyle fixes * Use CHIPDeviceController instead of CHIPDeviceController_deprecated (project-chip#3979) * Implement the missing part of Exchange Header in Transport layer (project-chip#4017) * Implement the missing part of Exchange Header in Transport layer * Revert comment 'if' back to 'iff'("if and only if") * Remove duplicated send flag defines and put ExchangeMgr/ExchangeConte… (project-chip#3994) * Remove duplicated send flag defines and put ExchangeMgr/ExchangeContext under the same namespace as CRMP * Rename kSendFlag_Default to kSendFlag_None * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation (project-chip#3998) * Move src/lib/asn1/tests and src/ble/tests to auto-test-driver generation * Remove one more unused file * Attempt to enable asn1 and ble tests in esp32 - see if they pass or not * Fix merge error * Update include header for ASN1 test * Include ASN1 in libCHIP * Some conversions to use PacketBufferHandle (project-chip#4011) * Some conversions to use PacketBufferHandle #### Problem Code should use `PacketBufferHandle` rather than `PacketBuffer *`. #### Summary of Changes - Converts remaining receive path in //src/inet and //src/transport. - Converts most of //src/ble. - Introduces Handle versions of the `AddToEnd`/`DetachTail` pair. Part of issue project-chip#2707 - Figure out a way to express PacketBuffer ownership in the type system * Restyled by clang-format * review * revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos Co-authored-by: Restyled.io <commits@restyled.io> * Move src/system/tests to auto-test-driver generation (project-chip#4000) * Move src/system/tests to auto-test-driver generation * Remove a TCP/IP init call that was killing qemu * Remove explicit "all" target from root build file (project-chip#3967) The "all" target exists implicitly and contains everything. We don't need to specify one, and it's misleading to do so specifying deps has no effect. * Make src/setup_payload compile with -Werror=conversion (project-chip#4032) * Add SSID and password to chip-tool pairing (project-chip#4054) * Get temperature-measurement and all-clusters-app to use examples/common/chip-app-server (project-chip#4039) #### Problem PR project-chip#3704 introduced a change where a `PeerAddress` is now required in order to start `RendezvousSession`. Sadly the multiple code paths bootstrapping `RendezvousSession` has not been updated. PR project-chip#4012 add a fix for some of the `examples/` but not for the `all-clusters-app` nor the `temperature-measurement-app`. To avoid such situation, this PR merge `examples/common/chip-app-server` and the custom code from `all-clusters-app` and `temperature-measurement-app`. One of the more discutable change of this PR (imo) is the code that moves the custom `Echo` mechanism from the `all-clusters-app` to `chip-app-server`. I was hoping to get rid of it before doing this change but the `all-clusters-app` and the `temperature-measurement-app` are broken since project-chip#3704 and this PR should fix that. Also I have a PR (mostly) ready once project-chip#3979 lands to get rid of this `Echo` specific code and replace it by a manufacturer specific `ping` command. #### Summary of Changes * Remove `EchoServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `all-clusters-app` * Remove `ResponseServer.cpp`, `RendezvousDeviceDelegate.cpp` and `include/RendezvousDeviceDelegate.h` from `temperature-measurement-app` * Introduce `chip-app-server/include/AppDelegate.h` in order to keep the behavior the `all-clusters-app` that turns on/off leds on different events. Maybe it should be converted to some types of `ChipDeviceEvent` or `CHIPCallback` at some point. * Fix `chip-app-server` to accomodate for the specifics of `all-clusters-app` * Add all Thread ULA addresses to the lwip interface (project-chip#4053) ULA prefixes will used for CHIP network so we need to add all these addresses to the interface. * Remove src/lib/message. (project-chip#4055) * Remove src/lib/message. Updated messaging implementation lives in src/messaging and the src/lib/message is not currently compiled or linked in. This saves us from looking at this code when some refactoring is needed (e.g. the SystemPacketBuffer changes). * Remove one more unused file * [ChipTool] Add Payload Parse Command (project-chip#3696) * [ChipTool] Add Payload Parse Command * [ChipTool] Add Payload Parse Command * [ChipTool] Restyle issues resolved * Restyled by whitespace * Resolve build errors caused by Command.h schema change Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: Restyled.io <commits@restyled.io> * rename ParseCommand to QRCodeParseCommand * adding AdditionalDataParseCommand version 0 * Adding parsing logic + logging * adding another parsing method * Basic Parsing is DONE * fixing memory issue * removing some logs * removing more logs * minor update * Add RotatingDeviceId to DNS-SD * Revert "Merge pull request #4 from hnnajh/rotating-id-test" This reverts commit 0235d05, reversing changes made to 3e1a4b9. * increment lifetime counter * Incrementing lifetime counter at each BLE/MDNS advertisement * address comments * Update src/include/platform/ConfigurationManager.h Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> * removing unnecessary lines * nit * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * restyling * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * Update src/app/server/Dnssd.cpp Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com> * fix comments * incrementing lifetimecounter once * remove unnecessary include Co-authored-by: Kevin Schoedel <67607049+kpschoedel@users.noreply.github.com> Co-authored-by: Vivien Nicolas <vnicolas@apple.com> Co-authored-by: Markus Becker <Markus.Becker@tridonic.com> Co-authored-by: Damian Królik <66667989+Damian-Nordic@users.noreply.github.com> Co-authored-by: Łukasz Duda <lukasz.duda@nordicsemi.no> Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Yufeng Wang <44623591+yufengwangca@users.noreply.github.com> Co-authored-by: Michael Spang <spang@google.com> Co-authored-by: Andrei Litvin <andrei@andy314.com> Co-authored-by: jepenven-silabs <67962328+jepenven-silabs@users.noreply.github.com> Co-authored-by: Boris Zbarsky <bzbarsky@apple.com> Co-authored-by: Jiacheng Guo <gjc@google.com> Co-authored-by: Liju Jayakumar <26148162+lijujayakumar@users.noreply.github.com> Co-authored-by: lijayaku <lijayaku@amazon.com> Co-authored-by: chrisdecenzo <61757564+chrisdecenzo@users.noreply.github.com>

boring-cyborg bot added app ble controller darwin examples inet lib labels Nov 27, 2020

kpschoedel commented Nov 27, 2020

View reviewed changes

kpschoedel marked this pull request as ready for review November 30, 2020 14:10

kpschoedel requested a review from bzbarsky-apple November 30, 2020 14:11

pullapprove bot requested review from andy31415, BroderickCarlin, chrisdecenzo, hawk248, jelderton, mspang, rwalker-apple, saurabhst and shana-apple November 30, 2020 14:11

andy31415 approved these changes Nov 30, 2020

View reviewed changes

rwalker-apple approved these changes Nov 30, 2020

View reviewed changes

saurabhst approved these changes Nov 30, 2020

View reviewed changes

Merge remote-tracking branch 'origin/master' into x2707-03

b8ac5e9

restyled-io bot mentioned this pull request Nov 30, 2020

Restyle Some conversions to use PacketBufferHandle #4031

Closed

BroderickCarlin approved these changes Nov 30, 2020

View reviewed changes

Restyled by clang-format

f904a80

mspang approved these changes Dec 1, 2020

View reviewed changes

bzbarsky-apple requested changes Dec 1, 2020

View reviewed changes

kpschoedel added 2 commits December 1, 2020 09:35

Merge remote-tracking branch 'origin/master' into x2707-03

1d08d8f

review

acb10c0

* revive BtpEngine::Clear[TR]xPacket() * simplify conditional * (void) message.DetachTail() → message.FreeHead() * remove ExchangeContext::kSendFlag_RetainBuffer * missed pBuf.IsNull() * DetachHead() → PopTail() * typos

bzbarsky-apple approved these changes Dec 1, 2020

View reviewed changes

Merge remote-tracking branch 'origin/master' into x2707-03

f8d146d

rwalker-apple merged commit 316a6f5 into project-chip:master Dec 1, 2020

kpschoedel deleted the x2707-03 branch December 1, 2020 19:53

kpschoedel mentioned this pull request Dec 4, 2020

Set/Reset on CHIPoBLEConState #4095

Merged

andy31415 pushed a commit that referenced this pull request Dec 7, 2020

Set/Reset on CHIPoBLEConState (#4095)

cf1cd59

Encapsulate setting ESP32's BLEManagerImpl::CHIPoBLEConState. (Requested in review comments on PR #4011. Other existing `BLEManagerImpl`s don't use this pattern.)

kpschoedel mentioned this pull request Dec 10, 2020

Received packet buffer is not getting freed after processing incoming message in IOS app #4150

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Some conversions to use PacketBufferHandle #4011

Some conversions to use PacketBufferHandle #4011

kpschoedel commented Nov 27, 2020

kpschoedel commented Nov 27, 2020

kpschoedel Nov 27, 2020

mspang commented Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

bzbarsky-apple Dec 1, 2020

kpschoedel Dec 1, 2020

github-actions bot commented Dec 1, 2020

github-actions bot commented Dec 1, 2020

	!(!mSendQueue.IsNull() \|\| mBtpEngine.TxState() == BtpEngine::kState_InProgress))
	mSendQueue.IsNull() && mBtpEngine.TxState() != BtpEngine::kState_InProgress)

		@@ -707,5 +712,19 @@ PacketBuffer * PacketBuffer::BuildFreeList()

		#endif // !CHIP_SYSTEM_CONFIG_USE_LWIP && CHIP_SYSTEM_CONFIG_PACKETBUFFER_MAXALLOC

		PacketBufferHandle PacketBufferHandle::DetachTail()

	* @return the detached bufferly formerly at the head of the buffer chain.
	* @return the detached buffer formerly at the head of the buffer chain.

Some conversions to use PacketBufferHandle #4011

Some conversions to use PacketBufferHandle #4011

Conversation

kpschoedel commented Nov 27, 2020

Problem

Summary of Changes

kpschoedel commented Nov 27, 2020

Choose a reason for hiding this comment

mspang commented Dec 1, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Dec 1, 2020

github-actions bot commented Dec 1, 2020