Jetson AGX HDMI Display is not working after applying RT Patch

Robotics & Edge Computing Jetson & Embedded Systems Jetson AGX Xavier
,
24

Could you see if

apply patch
defconf
build kernel

give you different result?

25

To try multiple options I tried below sequence>

root@Prasanna:/home/Build/sources/kernel/kernel-4.9# make clean
root@Prasanna:/home/Build/sources/kernel/kernel-4.9# make mrproper
root@Prasanna:/home/Build/sources/kernel/kernel-4.9# ./scripts/rt-patch.sh revert-patches
The PREEMPT RT patches have been successfully reverted!
root@Prasanna:/home/Build/sources/kernel/kernel-4.9# make ARCH=arm64 O=$TEGRA_KERNEL_OUT tegra_defconfig
make[1]: Entering directory ‘/home/prasanna/GRL_C2V/Build/sources/kernel/kernel-4.9/tegra_image’
HOSTCC scripts/basic/fixdep
HOSTCC scripts/basic/bin2c
GEN ./Makefile
HOSTCC scripts/kconfig/conf.o
HOSTCC scripts/kconfig/zconf.tab.o
HOSTLD scripts/kconfig/conf

configuration written to .config

make[1]: Leaving directory ‘/home/Build/sources/kernel/kernel-4.9/tegra_image’
root@Prasanna:/home/Build/sources/kernel/kernel-4.9# ./scripts/rt-patch.sh apply-patches
PREEMPT RT patches successfully applied for Auto!
PREEMPT RT patches successfully applied for L4T!
root@Prasanna:/home/Build/sources/kernel/kernel-4.9#
root@Prasanna:/home/Build/sources/kernel/kernel-4.9# make ARCH=arm64 O=$TEGRA_KERNEL_OUT -j24
after this I used to get the below log and select option 5
make[1]: Entering directory ‘/home/Build/sources/kernel/kernel-4.9/tegra_image’
GEN ./Makefile
scripts/kconfig/conf --silentoldconfig Kconfig
*

  • Restart config…
  • Kernel Features

Page size

  1. 4KB (ARM64_4K_PAGES)
  1. 16KB (ARM64_16K_PAGES)
  2. 64KB (ARM64_64K_PAGES)
    choice[1-3?]: 1
    Virtual address space size
  1. 39-bit (ARM64_VA_BITS_39)
  1. 48-bit (ARM64_VA_BITS_48)
    choice[1-2?]: 1
    Build big-endian kernel (CPU_BIG_ENDIAN) [N/y/?] n
    Multi-core scheduler support (SCHED_MC) [Y/n/?] y
    SMT scheduler support (SCHED_SMT) [N/y/?] n
    Maximum number of CPUs (2-4096) (NR_CPUS) [64] 64
    Support for hot-pluggable CPUs (HOTPLUG_CPU) [Y/?] y
    Numa Memory Allocation and Scheduler Support (NUMA) [N/y/?] n
    Preemption Model
  1. No Forced Preemption (Server) (PREEMPT_NONE)
  1. Voluntary Kernel Preemption (Desktop) (PREEMPT_VOLUNTARY)
  2. Preemptible Kernel (Low-Latency Desktop) (PREEMPT__LL) (NEW)
  3. Preemptible Kernel (Basic RT) (PREEMPT_RTB) (NEW)
  4. Fully Preemptible Kernel (RT) (PREEMPT_RT_FULL) (NEW)
    choice[1-5]:

choice[1-5]: 5
Timer frequency

  1. 100 HZ (HZ_100)
  1. 250 HZ (HZ_250)
  1. 300 HZ (HZ_300)
  2. 1000 HZ (HZ_1000)
    choice[1-4?]: 2
    Memory model
  1. Sparse Memory (SPARSEMEM_MANUAL)
    choice[1]: 1
    Sparse Memory virtual memmap (SPARSEMEM_VMEMMAP) [Y/n/?] y
    Allow for memory hot-add (MEMORY_HOTPLUG) [Y/n] y
    Online the newly added memory blocks by default (MEMORY_HOTPLUG_DEFAULT_ONLINE) [Y/n/?] y
    Allow for memory hot remove (MEMORY_HOTREMOVE) [Y/n] y
    Allow for balloon memory compaction/migration (BALLOON_COMPACTION) [Y/n/?] y
    Allow for memory compaction (COMPACTION) [Y/n/?] y
    Page migration (MIGRATION) [Y/?] y
    HMM mirror CPU page table into a device page table (HMM_MIRROR) [Y/n/?] y
    Enable bounce buffers (BOUNCE) [Y/n/?] y
    Enable KSM for page merging (KSM) [Y/n/?] y
    Low address space to protect from user allocation (DEFAULT_MMAP_MIN_ADDR) [4096] 4096
    Enable cleancache driver to cache clean pages if tmem is present (CLEANCACHE) [N/y/?] n
    Enable frontswap to cache swap pages if tmem is present (FRONTSWAP) [N/y/?] n
    Contiguous Memory Allocator (CMA) [Y/n/?] y
    CMA debug messages (DEVELOPMENT) (CMA_DEBUG) [N/y/?] n
    CMA debugfs interface (CMA_DEBUGFS) [N/y/?] n
    Maximum count of the CMA areas (CMA_AREAS) [7] 7
    Common API for compressed memory storage (ZPOOL) [N/m/y/?] n
    Low (Up to 2x) density storage for compressed pages (ZBUD) [N/m/y/?] n
    Memory allocator for compressed pages (ZSMALLOC) [Y/n/m/?] y
    Use page table mapping to access object in zsmalloc (PGTABLE_MAPPING) [N/y/?] n
    Export zsmalloc statistics (ZSMALLOC_STAT) [N/y/?] n
    Enable idle page tracking (IDLE_PAGE_TRACKING) [N/y/?] n
    Device memory (pmem, etc…) hotplug support (ZONE_DEVICE) [Y/n/?] y
    Unaddressable device memory (GPU memory, …) (DEVICE_PRIVATE) [Y/n/?] y
    Addressable device memory (like GPU memory) (DEVICE_PUBLIC) [N/y/?] n
    Enable seccomp to safely compute untrusted bytecode (SECCOMP) [Y/n/?] y
    Enable paravirtualization code (PARAVIRT) [N/y/?] n
    Paravirtual steal time accounting (PARAVIRT_TIME_ACCOUNTING) [N/y/?] n
    kexec system call (KEXEC) [Y/n/?] y
    Harden the branch predictor against aliasing attacks (HARDEN_BRANCH_PREDICTOR) [Y/n/?] y
    Unmap kernel when running in userspace (aka “KAISER”) (UNMAP_KERNEL_AT_EL0) [N/y/?] n
    Harden the branch predictor against aliasing attacks (HARDEN_BRANCH_PREDICTOR) [Y/n/?] y
    Speculative Store Bypass Disable (ARM64_SSBD) [Y/n/?] y
    Emulate Privileged Access Never using TTBR0_EL1 switching (ARM64_SW_TTBR0_PAN) [Y/n/?] y
    Randomize the address of the kernel image (RANDOMIZE_BASE) [Y/n/?] y
    Tegra Early Boot Partition (TEGRA_EBP) [N/y/?] n
  • Tracers

Tracers (FTRACE) [Y/n/?] y
Kernel Function Tracer (FUNCTION_TRACER) [Y/n/?] y
Kernel Function Graph Tracer (FUNCTION_GRAPH_TRACER) [N/y/?] n
Enable trace events for preempt and irq disable/enable (PREEMPTIRQ_EVENTS) [N/y/?] n
Interrupts-off Latency Tracer (IRQSOFF_TRACER) [N/y/?] n
Preemption-off Latency Tracer (PREEMPT_TRACER) [N/y/?] n
Scheduling Latency Tracer (SCHED_TRACER) [N/y/?] n
Tracer to detect hardware latencies (like SMIs) (HWLAT_TRACER) [N/y/?] n
Missed Timer Offsets Histogram (MISSED_TIMER_OFFSETS_HIST) [N/y/?] (NEW) n

26

Ok. I will try the sequnce as mentioned and update you.
Thanks

27

If the make is always asking you to enter which feature/function needs to be enabled each time, it means tegra_defconfig does not get read.
Tegra_defconfig has the necessary config the system needs for enabling tegra. During building kernel, you don’t need to choose functions one by one.

28

Thanks. I followed the sequence:
apply patch
defconf
build kernel

It didn’t asked for options. Once build complete I will test and update.
For flashing I am using sudo ./flash.sh jetson-xavier mmcblk0p1.

Since flashing takes time. Do we have any option like Copying the image into SD card from Host PC directly. and to boot from SD ?

29

Thanks for quick responses and support.

Finally with above sequence. GUI is up with RT patch.
root@jetson:~# uname -a
Linux jetson 4.9.140-rt93-tegra #6 SMP PREEMPT RT Tue Jan 26 17:41:01 IST 2021 aarch64 aarch64 aarch64 GNU/Linux

It will select. Fully Preemptible Kernel (RT) (PREEMPT_RT_FULL) By default ?

30

Hi Wayne,

I have a query. After successful flashing of RT patch applied kernel. System booting is fine.
How to prepare linux-headers-4.9.140-rt93-tegra-ubuntu-18.04_aarch64 ?
Actually.
Now I just tried to build a a kernel module on Jetson.
If I choose kernel directory as “/lib/modules/$(shell uname -r)/build” or
/usr/src/linux-headers-$(uname -r)-ubuntu18.04_aarch64
It is not able to build since now kernel is “4.9.140-rt93-tegra”.
So I just tried to select the kernel src as “/usr/src/linux-headers-4.9.140-tegra-ubuntu18.04_aarch64/kernel-4.9”. and the module built using above source gives as insmod: ERROR: could not insert module phy_intr.ko: Invalid module format.
dmesg : disagrees about version of symbol module_layout

If I cross compile the module from Linux Host PC. I am able to insert the module in Jetson.

Linux_for_Tegra/kernel/kernel_headers.tbz2 has only
linux-headers-4.9.140-tegra-linux_x86_64
linux-headers-4.9.140-tegra-ubuntu18.04_aarch64
. How to prepare for RT. ?

Thanks,
Prasanna