SBL Crashdumps

Introduction

In some crash scenarios in SBL, it may be required to debug without the use of a JTAG device. This can be accomplished with the crashdump feature in SBL.

Note

currently only supports graceful errors.

Detailed Operation

To collect and debug crashdumps, use the following procedure:

  1. If a crash occurs, the device screen should show “Crashdump Mode”.

  2. Collect crashdumps by opening the “QPST Configuration” tool you should see one COM port with a green progress bar. Dumps are usually collected to the following location (depending on your QPST installation):

    C:\ProgramData\Qualcomm\QPST\Sahara\Port_COMXX

  3. Start a T32 Simulator APPS window:

    <metabuild location>\common\Core\t32\t32start.pyc
Simulator -> APSS/ARM64 -> Start

  4. Run the boot_debug.cmm script (this is recommended if you do not want to load the XBLLoader symbols manually) CD.DO <build path>boot_imagesQcomPkgToolsboot_debug.cmm Follow the prompts and select option “10: Load SBL crash dumps”. It will prompt for path to XBLLoader.dll, then prompt for the QPST Port_COMXX folder where the dumps are located

The rampdump and symbols will be loaded and debugging related windows (source code, call stack, registers, etc.) will be displayed.

How it works

Crashes will cause a context backup to an unused DDR region. The device resets and kicks into ramdump image, where the context will be included as a separate binary (SBL_CRSH.BIN) in the default RAM dump to QPST. load_boot_crash_dump.cmm will use the binary to restore the context into a T32 instance, and the environment will now appear as if the device is paused at the moment of the crash.

Limitations

Currently only support for graceful errors.