hal: renesas: ra: Fix zero length packets not being sent for pipe0

[jeremydick]

Fix zero length packets not being sent for pipe0

This fixes samples/subsys/usb/webusb failing to enumerate properly

[thenguyenyf]

Hello @jeremydick . According to RA8M1 HWM, the executing for ZLP transfer at control status stage doesn't need a CFIFOCTR:BVAL writing. Could you provide more detail about the issue you met? A console output or a USB packet capture maybe helpful.

[jeremydick]

I am running samples/subsys/usb/webusb on the EK-RA6M3G eval kit

The webusb sample prj.conf was updated with these additional configs for logging

CONFIG_USBD_LOG_LEVEL_DBG=y
CONFIG_UDC_DRIVER_LOG_LEVEL_DBG=y
CONFIG_LOG_BUFFER_SIZE=32768

west build -p always -b ek_ra6m3 samples/subsys/usb/webusb

Without the fix, enumeration on the host fails, the device is getting stuck where it should be sending a ZLP because the last packet length is exactly the max packet length.

At line 124 in the logs:

[00:00:00.233,000] <dbg> usbd_ep: usbd_ep_ctrl_set_zlp: add ZLP, wLength 255 buf length 64
[00:00:00.233,000] <dbg> udc: udc_ep_enqueue: Queue ep 0x80 0x1ffeb4bc len 64
[00:00:00.233,000] <dbg> udc_renesas_ra: udc_renesas_ra_ep_enqueue: 0xdd58 enqueue 0x1ffeb4bc
[00:00:05.838,000] <dbg> usbd_core: usbd_event_handler: RESET event

With the patch:

[00:00:00.286,000] <dbg> usbd_ep: usbd_ep_ctrl_set_zlp: add ZLP, wLength 255 buf length 64
[00:00:00.286,000] <dbg> udc: udc_ep_enqueue: Queue ep 0x80 0x1ffeb4bc len 64
[00:00:00.286,000] <dbg> udc_renesas_ra: udc_renesas_ra_ep_enqueue: 0xdd70 enqueue 0x1ffeb4bc
[00:00:00.286,000] <dbg> udc: udc_ctrl_update_stage: s-in->(status)

webusb_sample_log.txt
webusb_sample_with_fix.txt