Heap memory leak when using gpio_install_isr_service and gpio_uninstall_isr_service (IDFGH-14172)

[MmdMi]

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.

General issue report

Title: Heap memory leak when using gpio_install_isr_service and gpio_uninstall_isr_service

Description:
I encountered a heap memory leak issue while testing the gpio_install_isr_service and gpio_uninstall_isr_service functions in my ESP-IDF project. When these functions are called, 16 bytes of heap memory remain unreleased, causing a memory leak.

Steps to reproduce:

1. Call gpio_install_isr_service with default flags.
2. Call gpio_uninstall_isr_service.
3. Compare the heap size before and after using the following code snippet:

uint32_t start_heap = esp_get_free_heap_size();
ESP_ERROR_CHECK(gpio_install_isr_service(0));
gpio_uninstall_isr_service();
uint32_t end_heap = esp_get_free_heap_size();
ESP_LOGI(TAG, "Diff[%ld]", (start_heap - end_heap));

Observed behavior:
The output of the above code is:

I (459) test_heap: Diff[16]

Expected behavior:
After calling gpio_uninstall_isr_service, the heap memory should be fully released, and the difference should be 0.

ESP-IDF version:

• ESP-IDF v5.3.1

Environment:

• ESP32-WROOM-32 module
• Platform: Ubuntu 24.04

Additional notes:
This issue might accumulate over time if the gpio_install_isr_service and gpio_uninstall_isr_service functions are used repeatedly in a long-running application, leading to eventual heap exhaustion.

[ESP-Marius]

The interrupt allocator will lazy allocate some descriptors, and these will be re-used when interrupts are freed.

So these 16 bytes you see here will be used again when someone allocates a new interrupt, repeatably calling gpio_install_isr_service will not cause issues or heap exhaustion.

[Alvin1Zhang]

@MmdMi Thanks for reporting, can you please help share if any further updates for the issue? Thanks.