fix(uvc): improve uac compatibility and add checks for uvc #308
Conversation
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
TDA-2030
force-pushed
the
fix/improve_uac_compatibility_and_add_checks_for_uvc
branch
2 times, most recently
from
ae92d53 to
efa0fe5
Compare
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
TDA-2030
force-pushed
the
fix/improve_uac_compatibility_and_add_checks_for_uvc
branch
2 times, most recently
from
016831a to
4ef3e94
Compare
TDA-2030
changed the title
TDA-2030
force-pushed
the
fix/improve_uac_compatibility_and_add_checks_for_uvc
branch
2 times, most recently
from
fae33ed to
389afe6
Compare
TDA-2030
force-pushed
the
fix/improve_uac_compatibility_and_add_checks_for_uvc
branch
from
389afe6 to
fd1c730
Compare
tore-espressif
left a comment
tore-espressif
left a comment
There was a problem hiding this comment.
@TDA-2030 UAC changes look OK. Could you please provide more information about the bInterval!=1 issue? Maybe there is a better way to fix it.
I have more comments for the UVC part, so if you want to speed up the release, please split this PR into 2. One for UAC and one for UVC, thanks!
| @@ -0,0 +1,12 @@ | |||
| | Supported Targets | ESP32-S2 | ESP32-S3 | ESP32-P4 | | |||
There was a problem hiding this comment.
Should work also on H4
| | Supported Targets | ESP32-S2 | ESP32-S3 | ESP32-P4 | | ||
| | ----------------- | -------- | -------- | -------- | | ||
|
|
||
| # UVC driver example: Video stream |
There was a problem hiding this comment.
Outdated description.
It would be nice to provide a simple 'how-to' use this example
| if (is_full_speed && ep_desc->bInterval != 1) { | ||
| // Full-Speed isochronous endpoints must have bInterval = 1 | ||
| uint8_t *_bInterval = (uint8_t *) & (ep_desc->bInterval); | ||
| // Check if we can modify the bInterval value | ||
| if (esp_ptr_in_dram(_bInterval) || esp_ptr_in_diram_dram(_bInterval) | ||
| #if CONFIG_SPIRAM | ||
| || esp_ptr_in_psram(_bInterval) | ||
| #endif | ||
| ) { | ||
| ESP_LOGW(TAG, "UAC Full-Speed device, Endpoint %d, bInterval %d, set to 1", USB_EP_DESC_GET_EP_NUM(ep_desc), ep_desc->bInterval); | ||
| *_bInterval = 1; | ||
| } else { | ||
| ESP_LOGW(TAG, "UAC Full-Speed device, Endpoint %d, bInterval %d, can't set to 1", USB_EP_DESC_GET_EP_NUM(ep_desc), ep_desc->bInterval); | ||
| } | ||
| } |
There was a problem hiding this comment.
This is a very risky, non-standard way that deserves explanation comment.
BTW, have you found the underlaying problem? If the device needs bInterval=2 why doesn't it work correctly with 2ms polling?
Description
UAC
UVC
usb_types_uvc.htoprivate_includedirectoryRelated
Testing
Checklist
Before submitting a Pull Request, please ensure the following:
Note
Fix UAC FS isochronous interval handling and harden UVC streaming with header/SOI validation; add an audio player example; bump UAC to 1.3.3 and UVC to 2.3.2.
UAC (usb_host_uac)
bInterval = 1on Full-Speed isoch endpoints during descriptor parse to prevent playback stutter; reject Low-Speed.USB_EP_DESC_GET_EP_NUM; direction asserts use USB masks.examples/audio_player(embedded WAV playback, optional MIC loopback) with build files, Kconfig, and configs.1.3.3; updateCHANGELOG.md.UVC (usb_host_uvc)
uvc_frame_payload_header_validate()and MJPEG SOI checks; improve BULK EoF handling; validate headers in ISOC; skip tiny/invalid frames; warn on missed EoF.usb_types_uvc.htoprivate_include; update includes across sources and tests.2.3.2; updateCHANGELOG.md.Written by Cursor Bugbot for commit fd1c730. This will update automatically on new commits. Configure here.