Don't receive packets immediately when usb packet equals CFG_TUD_ENDPOINT0_SIZE

roel0 commented 7 months ago

Operating System

Linux

Board

STM32U575

Firmware

938cae818

What happened ?

I dont receive USB packets when they are the maximum packet size CFG_TUD_ENDPOINT0_SIZE immediately. I do receive when the host sends more bytes. (The host is a linux PC). So it's seem that the incoming ZLP isn't processed.

If I trigger handle_epout_irq manually, the packet is still not processed. So the packet is for some reason not provided yet by the hardware.

How to reproduce ?

USB device with dualport CDC. Send exactly 64 bytes on FS connection

Debug Log as txt file (LOG/CFG_TUSB_DEBUG=2)

Will add later

Screenshots

No response

I have checked existing issues, dicussion and documentation

[X] I confirm I have checked existing issues, dicussion and documentation.

HiFiPhile commented 7 months ago

Are you sure a zlp was sent ? Normally it has to be done explicitly.

roel0 commented 7 months ago

Well,

If I use the the official STM32 USB library (yikes), it works fine and I get 3 usb interrupts when receiving a 64 bytes from my linux host
If I use tinyusb I only get one usb interrupt, which handles the 64 bytes packet. But no second interrupt to complete the transfer

-> So it's seems that the the stm32u575 is not correctly initialised by tinyusb? Still need to take a deeper look

roel0 commented 7 months ago

The issue appears when CFG_TUD_CDC_RX_BUFSIZE CFG_TUD_CDC_TX_BUFSIZE and CFG_TUD_CDC_EP_BUFSIZE have a different definition than CFG_TUD_ENDPOINT0_SIZE.

It's still a mystery to me why that is the case

HiFiPhile commented 7 months ago

Please enable debug print which gives you much more information.

HiFiPhile commented 4 months ago

USB 2.0 specification:

 A bulk transfer is complete when the endpoint does one of the following:
• Has transferred exactly the amount of data expected
• Transfers a packet with a payload size less than wMaxPacketSize or transfers a zero-length packet
When a bulk transfer is complete, the Host Controller retires the current IRP and advances to the next IRP.
If a data payload is received that is larger than expected, all pending bulk IRPs for that endpoint will be
aborted/retired.

So it's a intended behavior.

If you need to receive packet immediately:

Set CFG_TUD_CDC_EP_BUFSIZE to endpoint size, which is 64 for FS or 512 for HS
Send ZLP explicitly, you need low level control of the port, something like serial.write(b"") in Python won't work

ST's lib works as it queue data length equal to EP size:

/* CDC Endpoints parameters: you can fine tune these values depending on the needed baudrates and performance. */
#define CDC_DATA_HS_MAX_PACKET_SIZE                 512U  /* Endpoint IN & OUT Packet size */
#define CDC_DATA_FS_MAX_PACKET_SIZE                 64U  /* Endpoint IN & OUT Packet size */
#define CDC_CMD_PACKET_SIZE                         8U  /* Control Endpoint Packet size */

#define CDC_DATA_HS_IN_PACKET_SIZE                  CDC_DATA_HS_MAX_PACKET_SIZE
#define CDC_DATA_HS_OUT_PACKET_SIZE                 CDC_DATA_HS_MAX_PACKET_SIZE

#define CDC_DATA_FS_IN_PACKET_SIZE                  CDC_DATA_FS_MAX_PACKET_SIZE
#define CDC_DATA_FS_OUT_PACKET_SIZE                 CDC_DATA_FS_MAX_PACKET_SIZE

hathach / tinyusb