On di, 08 okt 2019 09:52:22 +0200, Marc Kleine-Budde wrote:
> Date: Tue, 8 Oct 2019 09:52:22 +0200
> From: Marc Kleine-Budde <mkl@xxxxxxxxxxxxxx>
> To: linux-can@xxxxxxxxxxxxxxx
> Cc: Kurt Van Dijck <dev.kurt@xxxxxxxxxxxxxxxxxxxxxx>, Joe Burmeister
> <joe.burmeister@xxxxxxxxxxxxx>
> Subject: [PATCH v3 0/4] can: c_can/rx-offload
> X-Mailer: git-send-email 2.23.0
>
> Hello,
>
> taking up Kurt's work. I've cleaned up the rx-offload and c_can patches
> a bit. Untested as I don't have any hardware at hand.
I had created equivalent code (skb_queue in isr, skb_dequeue in napi
handler) running on a 4.9 kernel since some days now. I didn't observe
any problems yet.
>
> I just looked at the c_can and d_can datasheets. Am I right, that both
> cores don't have a timestamp register to indicate when a CAN frame has
> been received?
I had the same impression.
The core clearly has not been built as a FIFO, but 'message objects'.
Hence the magic in the isr for trying to maintain proper sequence of
messages.
>
> regards,
> Marc
>
> Changes since v2:
> - added can_rx_offload_add_manual()
> - cleaned up c_can_isr()
> - remove unneeded dev_id cast
> - rename intreg -> reg_int to match the rest of the code
> - use { } on both sides of if else
> - fix return value
> - use can_rx_offload_add_manual() instead of can_rx_offload_add_fifo()
> - move can_rx_offload_add_manual()/can_rx_offload_del() to
> register_c_can_dev()/unregister_c_can_dev()
> - move include rx-offload.h to c_can.h
>
>
>
