I ported LvGL V9 latest version to my STM32H735IG-DK. Everything is working at I setup. However, once I launch the benchmark demo, the LvGL crashes and stops on this part of the benchmark test:
What MCU/Processor/Board and compiler are you using?
STM32H735IG-DK
What LVGL version are you using?
V9.3 Dev
What do you want to achieve?
Want to get the full benchmark of the system.
What have you tried so far?
Increased LV_MEM_SIZE from 60 to 80.
Increased the freeRTOS task to 4096 word size.
Increased head and stack to 16KB each.
Removed RTOS operation and went bare metal, still same issue.
If you look into the example itself, it says this:
#if LV_USE_STDLIB_MALLOC == LV_STDLIB_BUILTIN && LV_MEM_SIZE < 128 * 1024
#warning "It's recommended to have at least 128kB RAM for the benchmark"
#endif
Hence your issue. Great that you resolve and increase the RAM to 128kB.
when i increase the RAM to 128KB ， i got this trouble:
.\Objects\STM32H743IIT6-LVGL.axf: Error: L6407E: Sections of aggregate size 0x60d4 bytes could not fit into .ANY selector(s).
I have no idea about where to adjust my RAM allocation. can u help me ?
When you increase RAM allocation (e.g., to 128 KB), you need to specify not just “RAM = 128K” overall, but explicitly map each memory bank in your scatter file or linker script. If you previously configured only, say, DTCM to 128 KB and neglected the remaining SRAM regions, the linker can’t correctly allocate aggregated sections—hence the .ANY selector overflow.
To resolve this Open your linker script (e.g., .ld if using GCC/STM32CubeIDE, or .sct/scatter file for Keil). Declare all memory regions you want to use. Check your section placement, e.g., .data, .bss, .heap, .stack—make sure they are assigned to available RAM regions through proper SECTIONS definitions the rebuild.
