LPC-Link2 times out during SWD Discovery in 11.1.0 on macOS Catalina 10.15.3

[wj_ueos]

I'm trying to debug a RT1062 target (that successfully debugs under Windows 10 using the same LPC-Link2 and cabling). This is not successful on macOS 10.15.3. MCUXpresso successfully softloads the CMSIS-DAP V5.361 firmware and the probe is detected in the "Probes discovered" window as can be seen in the attached screenshot, but when debugging is about to start it gets stuck in SWD Discovery, followed by a timeout a while later.

Is this a known issue on the later macOS versions?

[lpcxpresso_supp]

I certainly don't see any such issues using LPC-Link2 on a Catalina-based Mac here (though that machine has no VM software installed). Nor do I see such problems on my own Mojave-based Mac when running Parallels VM.

To be honest, the usual problem when using VMs is when trying to debug from the VM itself rather than the host machine. In such circumstances, we have occasionally seen issues where the VM fails to spot the USB device reenumerating after the probe firmware is softloaded - which can typically be solved by detaching / reattaching the USB device within the VM software's menu system.

I can only assume that in your case, your Fusion setup has been configured such that it always grabs the LPC-Link2 device as soon as it enumerates?

Anyway, glad to hear that programming the probe firmware into the LPC-Link2 using LPCScrypt has provided a way forward for you.

Regards,

MCUXpresso IDE Support

[wj_ueos]

Thanks for the reply!

My Fusion setup was setup to ask every time a USB device was detected, and the odd thing was that the probe was detected properly and could be opened, but the (obviously longer) packets sent during SWD discovery seems to have been discarded/truncated.

By quitting Fusion and having the probe boot straight from QSPI I'm having 100% success debugging under the host Catalina OS. Booting Fusion back up and telling it to always connect the NXP LPC-Link2 device to the host also works properly, so I guess I'll leave it at that for now.

If I have time I might sniff the USB traffic between the probe and the computer to see any difference when Fusion intercepts the enumeration and when it isn't.

[Sabina_Bruce]

Hello ,

Hope you are doing well.

This is not a known issue. Could you please check how it is enumerating on your device manager?

Best Regards,

Sabina

[wj_ueos]

I noticed that I was getting Hardware interface transfer errors in Redlink when WireSWDConnect is being run (but probestatus and other identifying commands work fine).

Oddly enough using lpcscrypt to flash the CMSIS-DAP (full-featured) firmware and avoid soft-loading improved the situation somewhat. I had VMware Fusion running at the time (running a Windows 10 instance due to FPGA work) and for some reason the Redlink server isn't able to reliably communicate with the LPC-Link2 when Fusion is running (incorrect buffer handling somewhere?).

Exiting Fusion and using a NON soft-loaded LPC-Link2 makes it possible to debug on Catalina, but obviously this _should_ work straight out of the box with a soft-loaded probe and Fusion running in the background?

[wj_ueos]

This is how it looks like. As mentioned earlier, the debug probe is identified and booted, but SWD discovery never completes.