GIGIA: Giga display graphics no longer work with the release of 0.52.0

[mjs513]

Describe the bug
While retesting video on the display shield I started to receiving a Imprecise data bus error. So went back to just trying the GraphicsDisplayTest sketch. This uses the updated libraries that I issued PRs for:
arduino-libraries/Arduino_GigaDisplay#34
arduino-libraries/Arduino_GigaDisplay_GFX#10

This resulted in the same type of error:

[00:00:05.023,000] <err> os: ***** BUS FAULT *****
[00:00:05.030,000] <err> os:   Imprecise data bus error
[00:00:05.038,000] <err> os: r0/a1:  0x00000000  r1/a2:  0x00000000  r2/a3:  0x00000000
[00:00:05.049,000] <err> os: r3/a4:  0x00000000 r12/ip:  0x00000000 r14/lr:  0x08092751
[00:00:05.060,000] <err> os:  xpsr:  0x41000000
[00:00:05.068,000] <err> os: s[ 0]:  0x00001397  s[ 1]:  0x24073d18  s[ 2]:  0x000001e0  s[ 3]:  0x00000320
[00:00:05.080,000] <err> os: s[ 4]:  0x000001e0  s[ 5]:  0x08092163  s[ 6]:  0x000bb800  s[ 7]:  0x00000000
[00:00:05.093,000] <err> os: s[ 8]:  0x24020ae4  s[ 9]:  0x08092751  s[10]:  0x00000000  s[11]:  0x00000000
[00:00:05.105,000] <err> os: s[12]:  0x24032f9a  s[13]:  0x2405bc7c  s[14]:  0x24073d18  s[15]:  0x08092173
[00:00:05.118,000] <err> os: fpscr:  0x24032f9a
[00:00:05.125,000] <err> os: Faulting instruction address (r15/pc): 0x08047164
[00:00:05.135,000] <err> os: >>> ZEPHYR FATAL ERROR 26: Unknown error on CPU 0
[00:00:05.145,000] <err> os: Current thread: 0x24006060 (main)
[00:00:05.154,000] <err> os: Halting system

Target board + cli verbose compilation output
Full verbose compilation output, ideally with arduino-cli invocation or from IDE 2.3.3+
CompilerOutput.txt

Output of Serial Monitor

uart:~$ Timeout while waiting for high throughput clock
SDIO firmware download error, whd_bus_sdio_init failed at 559
Interface is not up/NULL and failed in function whd_deinit at line 203
Interface is not up/NULL and failed in function whd_wifi_get_mac_address at line 3184

[00:00:04.507,000] <inf> flashdisk: Initialize device wlan
[00:00:04.515,000] <inf> flashdisk: offset 1000, sector size 4096, page size 4096, volume size 1044480
[00:00:04.528,000] <inf> flashdisk: Initialize device ota
[00:00:04.535,000] <inf> flashdisk: offset 100000, sector size 4096, page size 4096, volume size 5242880
[00:00:04.548,000] <inf> littlefs: littlefs partition at /storage
[00:00:04.557,000] <inf> littlefs: LittleFS version 2.11, disk version 2.1
[00:00:04.566,000] <inf> littlefs: FS at qspi-nor-flash@0:0x700000 is 1792 0x1000-byte blocks with 512 cycle
[00:00:04.579,000] <inf> littlefs: partition sizes: rd 16 ; pr 4096 ; ca 4096 ; la 32
[00:00:04.590,000] <inf> littlefs: Automount /storage succeeded
*** Booting Zephyr OS build v4.2.0-31-g8e71141c1b0a ***
[00:00:04.624,000] <inf> usb_cdc_acm: Device suspended
[00:00:04.849,000] <inf> usb_cdc_acm: Device configured
Zephyr
height: 480, width: 800, rotated: 1
RGB565 set
Shared multiheap
sizeof: 768000
w: 480, h:800, buf: 768000 (from ArduinoGiga_display library when setFrameDesc is called)
[00:00:05.023,000] <err> os: ***** BUS FAULT *****
[00:00:05.030,000] <err> os:   Imprecise data bus error
[00:00:05.038,000] <err> os: r0/a1:  0x00000000  r1/a2:  0x00000000  r2/a3:  0x00000000
[00:00:05.049,000] <err> os: r3/a4:  0x00000000 r12/ip:  0x00000000 r14/lr:  0x08092751
[00:00:05.060,000] <err> os:  xpsr:  0x41000000
[00:00:05.068,000] <err> os: s[ 0]:  0x00001397  s[ 1]:  0x24073d18  s[ 2]:  0x000001e0  s[ 3]:  0x00000320
[00:00:05.080,000] <err> os: s[ 4]:  0x000001e0  s[ 5]:  0x08092163  s[ 6]:  0x000bb800  s[ 7]:  0x00000000
[00:00:05.093,000] <err> os: s[ 8]:  0x24020ae4  s[ 9]:  0x08092751  s[10]:  0x00000000  s[11]:  0x00000000
[00:00:05.105,000] <err> os: s[12]:  0x24032f9a  s[13]:  0x2405bc7c  s[14]:  0x24073d18  s[15]:  0x08092173
[00:00:05.118,000] <err> os: fpscr:  0x24032f9a
[00:00:05.125,000] <err> os: Faulting instruction address (r15/pc): 0x08047164
[00:00:05.135,000] <err> os: >>> ZEPHYR FATAL ERROR 26: Unknown error on CPU 0
[00:00:05.145,000] <err> os: Current thread: 0x24006060 (main)
[00:00:05.154,000] <err> os: Halting system

Output of readelf
Readelf.txt

Optional: attach the elf file
graphicsTestGigaDisplay.ino.zip

Optional: attach the sketch
graphicsTestGigaDisplay-251119a.zip

Additional context
Note this was working perfectly in 0.51.0!!!! was developing zephy video drivers for HM0360 and Ov7675 without a problem until now.

[mjs513]

Decided to try something - put back our 0.51.0 config file for the giga

CONFIG_USB_DEVICE_STACK=y
CONFIG_USB_DEVICE_PRODUCT="Arduino GIGA R1"
CONFIG_USB_DEVICE_MANUFACTURER="Arduino"
CONFIG_USB_DEVICE_VID=0x2341
CONFIG_USB_DEVICE_PID=0x0066

CONFIG_USB_CDC_ACM=y
CONFIG_USB_CDC_ACM_RINGBUF_SIZE=512
CONFIG_UART_LINE_CTRL=y
CONFIG_CDC_ACM_DTE_RATE_CALLBACK_SUPPORT=y

CONFIG_LLEXT_STORAGE_WRITABLE=n
CONFIG_SHARED_MULTI_HEAP=y
CONFIG_HEAP_MEM_POOL_SIZE=2048
CONFIG_SHELL_STACK_SIZE=32768
CONFIG_MAIN_STACK_SIZE=32768
CONFIG_LLEXT_HEAP_SIZE=128

CONFIG_FPU=y

CONFIG_ADC=y
CONFIG_DAC=y
CONFIG_PWM=y
CONFIG_I2C_TARGET=y

CONFIG_ICACHE=y
CONFIG_DCACHE=y
CONFIG_CACHE_MANAGEMENT=y

CONFIG_DMA=y
CONFIG_MEMC=y
CONFIG_SPI_ASYNC=y
CONFIG_SPI_STM32_INTERRUPT=y

CONFIG_ENTROPY_GENERATOR=y
CONFIG_TEST_RANDOM_GENERATOR=y

CONFIG_VIDEO=y
CONFIG_VIDEO_LOG_LEVEL_DBG=y
CONFIG_VIDEO_STM32_DCMI=y
CONFIG_VIDEO_BUFFER_POOL_NUM_MAX=3
CONFIG_VIDEO_BUFFER_POOL_SZ_MAX=614400
CONFIG_VIDEO_BUFFER_POOL_ALIGN=32
CONFIG_VIDEO_BUFFER_USE_SHARED_MULTI_HEAP=y
CONFIG_VIDEO_BUFFER_SMH_ATTRIBUTE=2
#CONFIG_VIDEO_GC2145=y
CONFIG_VIDEO_HM0360=y
#CONFIG_VIDEO_OV767X=y
#CONFIG_VIDEO_HM01B0=y

CONFIG_DISPLAY=y
CONFIG_INPUT=y
CONFIG_INPUT_QUEUE_MAX_MSGS=64
CONFIG_STM32_LTDC_FB_USE_SHARED_MULTI_HEAP=y
CONFIG_STM32_LTDC_FB_SMH_ATTRIBUTE=2

CONFIG_BT=y
CONFIG_BT_HCI_RAW=y
CONFIG_BT_HCI_RAW_H4=y
CONFIG_BT_HCI_RAW_H4_ENABLE=y
CONFIG_BT_BUF_ACL_RX_SIZE=255
CONFIG_BT_BUF_CMD_TX_SIZE=255
CONFIG_BT_BUF_EVT_DISCARDABLE_SIZE=255
CONFIG_BT_MAX_CONN=4
CONFIG_CYW4343W_MURATA_1DX=y
CONFIG_SYSTEM_WORKQUEUE_STACK_SIZE=2048
CONFIG_BT_RX_STACK_SIZE=4096

CONFIG_STM32H7_BOOT_M4_AT_INIT=n

basically removes wifi/littlefs/qspi/mbed. When it compiles with the old config I see

[294/295] Linking C executable zephyr/zephyr.elf
Memory region         Used Size  Region Size  %age Used
           FLASH:      248408 B       640 KB     37.90%
             RAM:      242384 B       512 KB     46.23%
          EXTMEM:          0 GB        16 MB      0.00%
           SRAM1:          0 GB       128 KB      0.00%
           SRAM2:          0 GB       128 KB      0.00%
           SRAM3:          0 GB        32 KB      0.00%
           SRAM4:          0 GB        64 KB      0.00%
          SDRAM1:          4 MB         8 MB     50.00%
        IDT_LIST:          0 GB        32 KB      0.00%

Now when I run the video sketch it is working without issue

Just for reference with the 0.52.0 config file for the giga

Memory region         Used Size  Region Size  %age Used
           FLASH:      591548 B       640 KB     90.26%
             RAM:      471776 B       512 KB     89.98%
          EXTMEM:          0 GB        16 MB      0.00%
           SRAM1:          0 GB       128 KB      0.00%
           SRAM2:          0 GB       128 KB      0.00%
           SRAM3:          0 GB        32 KB      0.00%
           SRAM4:          0 GB        64 KB      0.00%
          SDRAM1:          4 MB         8 MB     50.00%
        IDT_LIST:          0 GB        32 KB      0.00%
Generating files from /home/my_new_zephyr_folder/ArduinoCore-zephyr/build/arduino_giga_r1_stm32h747xx_m7/zephyr/zephyr.elf for board: arduino_giga_r1
[585/585] Generating llext-edk.tar.Z

My conclusion with all the stuff you are asking for its running out of memory!!!!!

[iabdalkader]

My conclusion with all the stuff you are asking for its running out of memory!!!!!

Yes, my guess is someone is malloc'ing memory and not checking if it's NULL. Can you reproduce the issue without your open PRs? If so I will look into it.

[mjs513]

My conclusion with all the stuff you are asking for its running out of memory!!!!!

Yes, my guess is someone is malloc'ing memory and not checking if it's NULL. Can you reproduce the issue without your open PRs? If so I will look into it.

@iabdalkader
Looks like #272 is related and is cleared up the same way. So the simplest sketch is probably just running this sketch:

void setup() {
  // put your setup code here, to run once:
  Serial.begin(115200);
  while (!Serial && millis() < 5000) {}
  pinMode(LED_BUILTIN, OUTPUT);

  void *buffer = malloc(51200);
  Serial.println((uint32_t )buffer, HEX);
  printk("Buffer: %p\n", buffer);
}

void loop() {
  digitalWrite(LED_BUILTIN, !digitalRead(LED_BUILTIN));
  delay(500);
}

[mjs513]

@iabdalkader
If you want a bit more complicated sketch.

GIGA_Display_GFX_test-251120a.zip

This is an early sketch test sketch before we generate changes to the graphics library and it will generate the fault as well.

[pennam]

@mjs513 the issue is this struct is never allocated and you are writing into it.

[KurtE]

@mjs513 the issue is this struct is never allocated and you are writing into it.

Good catch!

Note: currently I am not running into this, as we are currently asking for RGB565.
We are using the format: PIXEL_FORMAT_RGB_565
https://github.com/mjs513/Arduino_GigaDisplay/blob/e05eb4e8065a47ddec399962065f526f55ec9d71/src/GigaDisplay.cpp#L86-L90

But zephyr code is now using: PIXEL_FORMAT_BGR_565
https://github.com/arduino/zephyr/blob/zephyr-arduino-v4.2.0/drivers/display/display_stm32_ltdc.c#L113-L117
So the call: display_set_pixel_format is failing and we bail from the init code before we call the function that writes to the structure

https://github.com/mjs513/Arduino_GigaDisplay/blob/e05eb4e8065a47ddec399962065f526f55ec9d71/src/GigaDisplay.cpp#L109-L112

But the End up faulting as later on we try to allocate a buffer for the display which fails and then try to use it and then fault.

[pennam]

@mjs513 the issue is this struct is never allocated and you are writing into it.

fixing this makes the original example posted by @mjs513 work on release 0.52.0

[mjs513]

@mjs513 the issue is this struct is never allocated and you are writing into it.

fixing this makes the original example posted by @mjs513 work on release 0.52.0

Sorry was out doing things. But I just did the same thing and it worked for me as well with the example. Updated the library and ran our graphicstest which has more of the primitives and all worked as well from library calls with 0.52.0. In process of updating the library before I forget.

@pennam - really good catch as Kurt said we had it that way for so long and never spotted it!!!!!! One way glad we had the issue so we were able to fix it.

Another question - what are Arduino's plan to use lvgl?

EDIT: Just tested with the HM0360 camera as well and works like a charm