Fix #507 - add missing sdmmc_host_get_dma_info for sdcard driver

[fangpenlin]

Submission Checklist 📝

• I have updated existing examples or added new ones (if applicable).
• I have used cargo fmt command to ensure that all changed code is formatted correctly.
• My changes were added to the CHANGELOG.md in the proper section.
• I have used cargo clippy command to ensure that all changed code passes latest Clippy nightly lints.

Pull Request Details 📖

Fix #507 by adding the missing sdmmc_host_get_dma_info function pointer for the SD card host driver

Description

The get_dma_info function pointer here:

esp-idf-hal/src/sd.rs

Lines 384 to 390 in 8a11ef4

	#[cfg(not(any(
	esp_idf_version_major = "4",
	all(esp_idf_version_major = "5", esp_idf_version_minor = "0"),
	all(esp_idf_version_major = "5", esp_idf_version_minor = "1"),
	all(esp_idf_version_major = "5", esp_idf_version_minor = "2"),
	)))] // For ESP-IDF v5.3 and later
	get_dma_info: None,

supposes to be pointing to sdmmc_host_get_dma_info based on what I saw from the SDMMC_HOST_DEFAULT macro. Without this func pointer, the first time calling it will crash due to calling a null pointer from here:

https://github.com/espressif/esp-idf/blob/d7d3eb3f0a99f069a7814b3c0824626383e4a511/components/sdmmc/sdmmc_sd.c#L95

Testing

Adding patch pointing hal repo to my PR this problem is gone:

[patch.crates-io]
esp-idf-hal = { git = "https://github.com/LaunchPlatform/esp-idf-hal", rev = "ec432eae61fdbeecaaf05a7bd33b8ffa738911ab" }

output:

I (416) sleep: Configure to isolate all GPIO pins in sleep state
I (423) sleep: Enable automatic switching of GPIO sleep configuration
I (431) main_task: Started on CPU0
I (441) main_task: Calling app_main()
I (441) gpio: GPIO[36]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0
I (451) gpio: GPIO[35]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0 
I (461) gpio: GPIO[37]| InputEn: 0| OutputEn: 0| OpenDrain: 0| Pullup: 1| Pulldown: 0| Intr:0
I (531) esp_idf_svc_ws_client_cross_thread_bug: File File { fd: 9 } created
I (531) esp_idf_svc_ws_client_cross_thread_bug: File File { fd: 9 } written with [72, 101, 108, 108, 111, 44, 32, 119, 111, 114, 108, 100, 33]
I (541) esp_idf_svc_ws_client_cross_thread_bug: File File { fd: 9 } seeked
I (551) esp_idf_svc_ws_client_cross_thread_bug: File File { fd: 9 } opened
I (561) esp_idf_svc_ws_client_cross_thread_bug: File File { fd: 9 } read: Hello, world!
I (561) esp_idf_svc_ws_client_cross_thread_bug: Entry: "DCIM"
I (571) esp_idf_svc_ws_client_cross_thread_bug: Entry: "FOO.TXT"
I (571) esp_idf_svc_ws_client_cross_thread_bug: Entry: "SYSTEM~1"
I (581) esp_idf_svc_ws_client_cross_thread_bug: Entry: "TEST.TXT"

you can also see the branch I created in the reproducing repo:

https://github.com/LaunchPlatform/esp-idf-hal-sd-get-info-dma-bug/tree/with-fix

[ivmarkov]

Is it not working without this callback provided, with V5.3?

[fangpenlin]

Probably won't work with v5.3. The get_dma_info function is called from multiple places in

https://github.com/espressif/esp-idf/blob/d7d3eb3f0a99f069a7814b3c0824626383e4a511/components/sdmmc/sdmmc_sd.c#L95

Any function relies on it will crash. I am not sure if there any cases this function won't be called (like DMA not available?)

[fangpenlin]

The PR is ready for review, by the way I tried my best to run clippy, but there are many build errors. I have no time to resolve them all just for running clippy. This is just one line change, there shouldn't be anything new code smell added.

[fangpenlin]

The platform I am using for testing this is esp32-s3-usb-otg

https://espressif-docs.readthedocs-hosted.com/projects/esp-dev-kits/en/latest/esp32s3/esp32-s3-usb-otg/user_guide.html

by the way

[ivmarkov]

Thanks - merging.

BTW, if you are successfully running cargo build, there is no reason whatsoever cargo clippy to fail, as it is a subset of cargo build so to say.

[ivmarkov]

... just reporting more stuff.