Known Issues
- Cannot launch Target Indicator, getting error about missing VCRUNTIME140_1.dll / Cannot do any tracing with “Failed to load platform” error (Windows* host only)
Issue: This should not be an issue, but in case it appears - it means Microsoft* C++ Redistributables package is broken.
Workaround:
Go to “Add/Remove Programs”.
Uninstall all “Microsoft C++ Redistributables”.
Uninstall Intel® System Debugger.
Install Intel® System Debugger again.
- Trace and debug operations on Alder Lake S require changes in platforms OS workaround settings
Issue: Power state transition (S5, S4, WR) with probe mode detection method TapPolling causes debug tool hang. Meanwhile, TapPolling is required for run control on targets with OS running. Starting with U2122 release, Intel® System Debugger NDA enables you to manage probe mode detection logic based on use cases.
Workaround: Follow instructions at
<install_dir>/documentation/tca/usage/connecting_gui.html . Depending on your goals, complete step 7 as follows:For BIOS debugging or FW/SW tracing, select “Target runs no OS” (Default). Debugging and tracing on platform that does power transitions requires this option selected.
For working with Architectural Event Traces (AET) or WinDbg* over Intel® Direct Connect Interface (Intel® DCI), select “Target runs an OS”.
- Disconnecting and reconnecting Intel System Debugger to Alder Lake S target with survivability flag enabled causes NorthPeak (NPK) hardware hang
Issue: During NPK tool disconnect and reconnect, NPK IP is power gated and then power ungated and the default register value causes timeout and hang.
Workaround: Disable survivability mode if debugging across low power states is not required.
- Low power debug is not supported for Alder Lake P due to missing Low Voltage Margining (LVM) overrides
Issue: Alder Lake P Best Known Configuration (BKC) does not support low power.
Workaround: N/A
- Unicode error returned during target selection via console
Issue: When selecting a target with a command like trace.interactive_setup(), an error like UnicodeEncodeError: ‘latin-1’ codec can’t encode character ‘u2122’ in position 24: ordinal not in range(256) is displayed.
Workaround: Before starting the console session, execute the following command:
Windows:
set PYTHONIOENCODING=UTF8 Linux:
export PYTHONIOENCODING=UTF8
- Cannot connect Intel® System Debugger to Alder Lake P target with USB3 Debug Class (DbC)
Issue: Debug connection is not working with USB3 DbC connected to Alder Lake P. Reset break is also not functional.
Workaround: There are two ways to resolve connectivity issues with Intel System Debugger if USB3 DbC is detected:
In the final step of establishing target connection (“Connection Summary” dialog box), expand the “Platform Debug Configuration” tab and uncheck the “Configure High Bandwidth Transport” box.
Modify IFWI via Intel® Flash Image Tool (Intel® FIT) to enable USB2 port and avoid USB3 DbC detection.
- Cannot establish USB DbC connection to debug the 10th generation Intel® NUC with Alder Lake M or Alder Lake P
Issue: When attempting to connect to the target using USB DbC, both Target Indicator and Windows* Device Manager report errors on the USB device (USB DbC probe).
Workaround: N/A
- Target Indicator is not visible when being started on Linux* hosts with GNOME 3.26 and above (Linux* host only)
Issue: On Linux* hosts with GNOME 3.26 and above, the Target Indicator is not visible when being started because GNOME 3.26 removes the system tray that the Target Indicator heavily relies upon.
Workaround: You can start the Target Indicator executable with the
--window-only (or short-w ) flag, which makes the Target Indicator avoid using the system tray.- Cannot start the Target Indicator from a file explorer with older package versions (Linux* host only)
Issue: Starting the Target Indicator executable from a file explorer (for example, Nautilus) fails, since it is not recognized as an executable file.
Workaround: Upgrade the “file” package to a version higher than 5.36. Start the Target Indicator from the command line or via the provided application link (the package upgrade does not affect it). See the original bug report.
- Reset causes
libusb error and loss of run control on Linux* host Issue : Warm or cold reset of the target connected to Comet Lake CPU/Comet Lake PCH-H (CML-S) in user interface or command line (ipccli) causes
libusb error and loss of run control. Cold reset is necessary to connect to the target again.Workaround: Use the Intel® DCI USB DbC2 instead of USB DbC3 on the port used for the debug connection. You can configure this in the Intel® Flash Image Tool (Intel® FIT)
Launch the Intel® FIT (obtained from Intel® Validation Internet Portal) and change the following settings within your current configuration:
Debug > Direct Connect Interface Configuration > Direct Connect Interface (DCI) Enabled > Yes
Debug > Early USB DBC over Type-A Configuration > USB2 DbC port enable > USB2 Port <X>
where X identifies the post number. Depending on the target, set the following value:
1 - to enable USB2 DbC on the vertical USB type-A port on CML-U DDR4 boards
3 or 4 - to enable USB2 DbC on either of the two stacked USB type-A ports respectively on CML-U LP3 boards
8 - to enable USB2 DbC on the vertical USB type-A port on CML-S boards with CMP-H or CNP-H chipset.
Build the image and flash the target with the new image.
- Disconnecting a probe crashes the target (Windows* host only)
Issue: Intel® Silicon View Technology (Intel® SVT) Closed Chassis Adapter (CCA) and Intel® In-Target Probe (Intel® ITP) - XDP3 probes are supporting hot plug/unplug from a target, whereas for Intel® DCI DbC the connection is bidirectional. If a probe connection is lost, the debugger reports a Power Loss event.
Workaround: If the target was running, reconnect the probe (the disconnection has no effect on the running target). If the target was halted, reboot the target and restart the debugging session (the debug context is lost).
- Platform security policy might inhibit debugger operation
Issue: In some platforms, the security policy might disable JTAG access to the CPU. This is intended to prevent reverse-engineering. In this case, the Intel® System Debugger will successfully connect to the target, but will not be able to discover any CPUs on the JTAG bus and will report that no CPU is available.
Workaround: Ensure that that platform firmware has enabled access to the CPUs via JTAG. You can do it by flashing a special “debug” firmware into the target. In some cases the CPU or the CPU module might physically disable JTAG access, especially in production or near-production versions. In this case, please work with the platform business unit to obtain a JTAG-enabled hardware.
- Installation fails if Microsoft .NET Framework* is not installed (Windows* host only)
Issue: Microsoft .NET Framework 3.5 and 4.0 are critical prerequisites successful installation of Intel® DFx Abstraction Layer (Intel® DAL).
Workaround:
If you want to use Intel® DAL, install Microsoft .NET Framework 3.5 and Microsoft .NET Framework 4.0 and retrigger installation.
If you do not need Intel® DAL and wish to proceed the installation without .Net framework, select “Custom installation” in the installer window, de-select Intel® DFx Abstraction Layer, and continue installation.
- Intel® DFx Abstraction Layer (Intel® DAL) logs cannot be archived (Windows* host only)
Issue: When you attempting to archive Intel® DAL logs using the commands
itp.daldebugloggerlevel("default") followed byitp.loggerarchive() , a “Permission Denied” error is printed.Workaround: Before every execution of the commands above, navigate to
<install-dir>\tools\DAL_<version> and modify thePublicLogging.log4net file to replace “DALBase.log” with a location where you have write permissions (for example,C:\Temp\DALBase.log ).- Cannot connect Intel® System Debugger to Raptor Lake CPU / Alder Lake PCH-S target with Intel® In-Target Probe (Intel® ITP) - XDP3
Issue: Debug connection is not working with Intel® ITP XPD3 connected to Raptor Lake CPU / Alder Lake PCH-S.
Workaround: Use a different debug probe (see Supported Platforms and Probes for available connection options).
Target Connection Assistant
- Unlock State in the Connection view is frozen
Issue: Unlock State displayed in the Connection view always shows “Locked” even for functional targets.
Workaround: Ignore the status.
- Sapphire Rapids 4-socket modular system is not detected automatically
Issue: Sapphire Rapids 4-socket system is not automatically detected by the Target Connection Assistant.
Workaround:
Open the XML file located at
<install-dir>\tools\OpenIPC_<version>\Config\OpenIpcConfig.xml and edit the Name attribute as follows:<DefaultIpcConfig Name="SPR_EBG_OpenDCI_Modular" />
Launch Intel® System Debugger, start establishing target connection, and choose “Manually select target” (for instructions, see the Intel System Debugger documentation).
In the final step of the connection wizard, specify the new OpenIPC provider name as follows:
- Raptor Lake PCH is not detected correctly
Issue: Target Connection Assistant erroneously detects Raptor Lake PCH as Alder Lake PCH (ADP).
Workaround: Select detected Alder Lake PCH as target configuration. Raptor Lake PCH will work as expected.
Intel® System Debugger - System Debug
- Intel® Processor Trace (Intel® PT) decoding timeouts
Issue: An error message “ITrace: Processor Trace decoding timed out” is displayed in the console.
Workaround: Reduce the scope of the executing trace capture, preferably by setting breakpoints at the point of interest. For instructions and alternative solutions, refer to the Troubleshooting chapter of the documentation.
- Source view points to a wrong function
Issue: Disassembly or Debug views show different function name as compared to the Source view. This issue only occurs if function have same addresses.
Workaround: Consider the information in the Disassembly and Debug views correct.
- Limited call stack depth
Issue: Call stack only shows 2 to 3 frames.
Workaround: Use stepping to return from the last function in the visible stack frames in order to get deeper call flow.
- Pressing Step-In (F5) Multiple Times with Simics® simulator crashes the Intel® System Debugger
Issue: When pressing Step-In multiple times really fast might cause the Intel System Debugger to crash.
Workaround: Only reproducible when clicked in short bursts without a pause, hence is usable in most cases, or user can also use breakpoints to step multiple times.
Intel® System Debugger - System Trace
- Rocky Linux 8 Support
Issue: Trace does not support Rocky Linux 8 because of GLIBC version mismatch (/lib/x86_64-linux-gnu/libm.so.6: version `GLIBC_2.29 not found).
Workaround: Update to Rocky Linux 9
- Elkhart Lake target system boot flow hangs after connecting the Trace Tool
Issue: After connecting the Trace Tool to some variants of Elkhart Lake, the target system hangs during boot. The system remains unbootable until the Trace Tool is disconnected.
Workaround: No workaround available. Contact Intel support if you face the issue.
- CentOS* 7 and OpenSuse* 15 host OS are not supported (Linux* host only)
Issue: CentOS* 7 and OpenSuse* 15 host OS are not supported for System Trace use cases.
Workaround: N/A
- Various error messages referring to Interview_Decoder.dll or visa64.dll are displayed (Windows* host only)
Issue: If the host system has IVI tools installed, the tool erroneously tries to load visa64.dll from C:WindowsSystem32 or Interview_Decoder.dll from <install-dir>system_debugger_2020_ndasystem_tracebin.
Workaround: Do any of the following:
Temporarily remove
C:\Windows\System32\visa64.dll or uninstall the IVI tools.Check the information on the Dynamic-Link Library Search Order but be aware of the potential security implications.
- Cannot run TDE due to an error creating inter-process shared memory (Windows* host only)
Issue: TDE fails to run with an error message: “Error creating inter-process shared memory for RC: boost_sysdbg::interprocess”
Workaround: Reboot the Windows* host system.
- On Glacier Falls targets, the tool reports errors when tracing to memory over an Intel® DCI USB DbC connection
Issue: Trace extraction over DMA is not supported on this platform. When the tool stops a trace to memory operation (Trace Profile: BIOS Reserved Trace Memory or Intel® Trace Hub Memory), the console displays a list of errors starting with
[Error]: npk::TargetConfig::configure Error configuring CfgTask 'DbC.DMA Init' Reason: Failed tap2iosf cmd . This error does not impact any functionality: trace data will still be downloaded and decoded by automatically falling back to an alternate extraction method.Workaround: To disable the failing DMA method and prevent these errors from being shown, go to Window > Preferences > System Trace and uncheck the Attempt fast USB trace extraction box in the top-level System Trace preferences pane.
- On Glacier Falls targets, the tool does not retrieve any trace messages over an Intel® DCI USB DbC connection on some firmware configurations
Issue: When selecting the “Intel® DCI USB Debug Class” connection method and starting/stopping a trace capture operation using the “Streaming to Intel® DCI USB 3.x Debug Class” trace profile, no trace events are captured. The “MIPI STP Decoder Trace Statistics” in the console reports 0, 0, 0 packets.
Workaround: Intel® DCI USB DbC streaming is not supported. Use a non-streaming Trace Profile: BIOS Reserved Trace Memory or Intel® Trace Hub Memory.
- Ice Lake targets crash after warm reset when connected via the Intel® DCI USB DbC3 debug cable
Issue: When Intel® DCI USB DbC3 debug cable is used to capture over warm reset, stopping capture right after the reset causes target system to crash.
Workaround: Connect to target using Intel® DCI USB DbC2 debug cable, either by enforcing USB2 on a port in the Intel® FIT tool or using Type-C UFP cable.
- No traces observed when multiple USB endpoints are enumerated by host
Issue: For targets with multiple trace endpoints (CPU and PCH, e.g. for Ice Lake platforms), using the “Streaming to Intel® DCI USB 3.x Debug Class” trace profile may result in the wrong trace endpoint being selected.
Workaround: Specify the correct trace mode manually in the “Streaming Decode Preferences” page. This page can be opened either via the menu bar by going Window > Preferences > System Trace > Streaming Decode or directly from the “Trace Capture” view by extending the popup menu in the upper right corner of the view and then opening the “Streaming Decode Preferences” dialog. In the “Decoder Parameters” table, scroll down to the entry for Decoder “IPC_Connector” and Name “mode”. Change the value field to either
usb-streaming-pch to capture traces from the PCH endpoint orusb-streaming-cpu to capture traces from the CPU endpoint. Click “Apply and Close” to apply the changes.- Trace does not work with Comet Lake CPU / Comet Lake PCH-V target when BIOS system memory is used
Issue: Trace destination for BIOS system memory is not applied correctly. Users see only time sync packets being streamed.
Workaround: Use streaming to Intel® SVT CCA or tracing to ITH buffer.
- On Cooperlake (CPX), streaming trace to DbC does not work
Issue: When selecting USB DbC as connection method and starting a stream capture, the capture process launches but no trace events are captured. The following error message is displayed in the Eclipse* console:
[ERROR] [tracehub_api] MTB_LBAR value is not available. Unable to configure DBC trace.
In TraceCLI, the error looks as follows:
[ERROR] npk::TargetConfig::configure:638 Error configuring CfgTask 'DbC Trace Config' Reason: Failed tap2iosf cmd: [Write] dbc_trace_in_payload_bp_low: (0x00000050) = 0xfe180000
Trace to memory (DRAM and Trace Hub internal memory) is not affected.
Workaround: Use a CCA connection for streaming trace.
- On Discrete Graphic platforms (DG1/DG2), errors occur during connection attempt of trace capture start
Issue: Due to security policies in DG1 and DG2, access to North Peak Test Access Port over Lauterbach Processor Trace Interface connection is blocked, An error message “Failed Intel® Trace Hub hardware detection check!” appears.
Workaround: Currently, you can configure DG1/DG2 tracing only through firmware running on the target. To enable this tracing capability, launch Intel® Flash Image Tool (Intel® FIT), open Debug Settings, and configure the “Emergency mode”. To start tracing, configure the firmware as described above, connect the debug probe, deselect “Configure target during capture start and stop”, and press the “Start Capture” button in the Trace Tool. The error message “Error in connecting to the target. Target does not have power.” will appear but you can safely ignore it.
Issue: Using TraceCLI to capture traces fails with “Unknown mode specified, falling back to port list” error message (https://jira.devtools.intel.com/browse/DEBGGR-11176).
Workaround: Use Eclipse-based System Trace instead of TraceCLI.
- Trace Agent CLI does not prompt for username and password when connecting to a SSH target
Issue: When you try to connect to an SSH target, you are not prompted for a username or a password. As a result, if a username and a password are required for the connection, it fails with the following message: “Failed to authenticate user ‘None’ without password”.
Workaround:
If you use the command line, explicitly pass the the username and the password as command line parameters.
If you are within a Python* environment, use the connect() method as follows:
agent.connect(<ip>, <port>, user=<username>, passw=<password>)
- Collecting early boot trace in TraceCLI may require manually setting OpenIPC TargetOS attribute
Issue: Collecting early boot trace data (including pre-DBC) in TraceCLI using the trace.ext.earlyboot.run() function may result in an incomplete trace. Only a small fraction of the streaming trace data is captured at the beginning of the boot flow, later boot stages are missing from the trace.
Workaround: Run the following commands in the same TraceCLI session before starting an early boot capture:
import ipccli; itp = ipccli.baseaccess() itp.config.root.TargetOS = "Windows" itp.forcereconfig()
- No Timeline Viewer templates available
Issue: Timeline viewer templates are not available when creating a new Trace project.
Workaround: Close Eclipse and reopen the same workspace with a new Trace project.
- System Trace does not use the updated port for Simics® simulator connection
Issue: After updating the Simics® simulator port, System Trace keeps using the old port.
Workaround: Go to Window > Preferences > System Trace > Streaming Decode Preferences and manually specify the URL and the port:
decoder name:
TCF_Connect decoder parameter:
tcf-agent-url
Intel® Debug Extensions for WinDbg*
- Compatibility issues of Intel® Debug Extensions for WinDbg* for Windows* 10 SDK
Issue: Windbg* over Intel® DCI plugin is not compatible for Windows* SDK versions 1809 (10.0.17763) and 1903 (10.0.18362.1).
Workaround: Install the latest Windows* SDK version 22000.
- Windbg* over Intel® DCI Extensions command
breakin(VMMEXIT) is not breaking Issue: Windbg* over Intel® DCI Extensions command
breakin(VMMEXIT) is not breaking in VM exit.Workaround: N/A
Intelligent Debug & Validation Tool (IDV Tool)
- Captures are reported as starting later than the actual trigger
Issue: Sporadically captures are reported as starting later than the actual trigger.
Workaround: Trigger the same capture again.
- Some files are not removed after uninstalling
Issue: When you try to uninstall the IDV tool, some files are not removed automatically.
Workaround: Delete the
C:\IntelSWTools\system_debugger\2142-nda\idv_tool folder manually.