5.4 vs 5.3.2 difference (IDFGH-14216)

[dannybackx]

Answers checklist.

• I have read the documentation ESP-IDF Programming Guide and the issue is not addressed there.
• I have updated my IDF branch (master or release) to the latest version and checked that the issue is present there.
• I have searched the issue tracker for a similar issue and not found a similar issue.

IDF version.

Any 5.4 version ("5.4" as well as "5.4 beta 2") compared to 5.3.2

Espressif SoC revision.

Warning: ESP32-S3 has no Chip ID. Reading MAC instead. MAC: 48:27:e2:e7:4a:88

Operating System used.

Linux

How did you build your project?

Command line with idf.py

If you are using Windows, please specify command line type.

None

Development Kit.

ZX3D95CD01S

Power Supply used.

External 5V

What is the expected behavior?

I expected the image to be the same

What is the actual behavior?

Visual output shifted

Steps to reproduce.

https://sourceforge.net/p/esp32-s3-86-box-project-base/code/HEAD/tree/trunk/alarm

Compile with the two versions mentioned.

Debug Logs.

No response

More Information.

No response

[Lzw655]

Hi @dannybackx,

I'm sorry you encountered this issue and not sure what changes ESP-IDF 5.4 beta has compared to 5.3.2. But this is an RGB LCD screen drift issue that occurs with a high probability when using WIFI or erasing the flash. I reviewed the source code you provided, and the issue can be resolved with the following modifications:

• Add new configuration options in sdkconfig to improve system performance:
  • CONFIG_ESP32S3_DATA_CACHE_LINE_64B=y
  • CONFIG_COMPILER_OPTIMIZATION_PERF=y
  • CONFIG_FREERTOS_HZ=1000
• Set .bounce_buffer_size_px = 10 * 480, in the GC9503 LCD initialization configuration esp_lcd_rgb_panel_config_t (if the issue persists, try increasing it to 20 * 480).

[dannybackx]

Thanks but these don't help

[Lzw655]

Sorry, I missed something. Please also add the CONFIG_SPIRAM_XIP_FROM_PSRAM=y configuration.

[dannybackx]

Please note that using the 5.4 (beta1, beta2, rc1) trees also causes a shift in a demo application (on the same hardware) that uses a completely different video software stack.
The one I referenced earlier uses esp-idf components for hardware drivers.

This one - see https://sourceforge.net/p/esp32-s3-86-box-project-base/code/HEAD/tree/trunk/minimal/ - uses the software stack provided by the hardware manufacturer. Same behaviour : with 5.3 the UI looks normal, with any of the 5.4 distributions, the screen is shifted.

[ckirmse]

I am running into the same situation as dannybackx and can confirm it's there even in very simple apps--something important has changed in the esp-idf 5.4 release for rgb displays and so far I have not been able to work around it in any way.

[dannybackx]

Same thing on a Spotpear module.