Re: [libusb] Issues with low CDC_ACM throughput on STM32 USB-FS (sync/async)

[Jerri L. <Jer...@pr...>]

On Tuesday, April 23rd, 2024 at 7:47 PM, Tim Roberts <ti...@pr...> wrote:

> Jerri Lipp via libusb-devel wrote:
>
>> Undoubetdly true, though as it happens, I'm running this on an old junk box that is USB2/EHCI only, so that's not it at least in this case.
>
> How old? Many of the early non-Intel EHCI controllers were crap, and had problems sustaining high bandwidths.

Following up on your suggestion that the controller may be wonky, I did a test on a newer old laptop (circa 2018 sunrise-point LP) which has an xhci controller.... and it was even worse. even with dd I got around 400kb/s max. The added IN transfer which made such a difference on the old machine (with an async-API based test program) made only a minor difference. It probably doesn't surprise you, but it surprised the hell outta me. I tried ratcheting up the in-flight transfers for the OUT direction, and that made no difference. I tried various ports, type-c (native to my STM32 board), or Type-A with an adapter, it made no difference. Looked at the UEFI setup, nothing jumps out.

I'm just about at the point of giving up on finding an answer. It's hard to believe that such an old and common protocol is this broken, across hardware platforms that span 10+ years. I'm a little disgusted.

I hope the problem is actually a bug in the ST USB stack after all, which interacts weirdly (and differently) with various hardware. That would make more sense to me.

Thanks for the suggestions,

Jerri

>