Re: Speed specific Link-Layer Header Types for USB 2.0

[Tomasz Moń via tcpdump-workers <tcpdump-workers () lists tcpdump org>]

> --- Begin Message ---
>
>
> From: Tomasz Moń <desowin () gmail com>
>
> Date: Mon, 09 May 2022 16:21:36 +0200
>
> On Mon, 2022-05-09 at 07:47 +0200, Tomasz Moń wrote:
> > The problematic case (and the reason why full/low-speed bus is
> > mentioned) is the LINKTYPE_USB_2_0_FULL_SPEED. It is the case when
> > you capture at the connection between full speed hub and the host
> > (and possibly full speed device connected to a full speed hub if
> > there arelow speed devices connected to the full speed hub).
>
> The low speed packets (after the PRE packet) intended for low-speed
> device connected to full-speed hub downstream port, do appear on the
> link between full-speed hub and full-speed device.
>
> > If there is low speed device connected to downstream hub port, then
> > when the host wants to send packets to the low speed device, these
> > will be sent at low speed to the hub. However, there will be PRE
> > packet (sent at full speed) before every low speed transaction.
> >
> > Hopefully I have recently found full-speed only hub (thanks to Linux
> > kernel bug report [1]), so I have all the equipment necessary to
> > check how the OpenVizsla deals with the problematic case.
> >
> > [1] https://bugzilla.kernel.org/show_bug.cgi?id=213839
>
> OpenVizsla does not handle this case. While the PRE packet is correctly
> captured, the low-speed data is just exported by gateware as invalid
> packet. Similarly, the LambdaConcept USB2 sniffer does not handle this.
> LCSniff displays the "ERR/PRE Packet" colored in red (with no
> indication of the low-speed data).
>
> When OpenVizsla is configured to capture low-speed connection with the
> capture cable connected between host and full-speed hub, the data is
> just garbled. The only way to capture the low-speed data (with
> OpenVizsla configured to capture low-speed) is to capture it between
> the full-speed hub downstream port and low-speed device.
>
> I am not really surprised that both OpenVizsla and LambdaConcept USB2
> sniffer fail at this problematic case. This is a corner case really,
> and it is not that much relevant nowadays where majority of hubs are
> high-speed capable. When the high-speed hub is connected to high-speed
> host, there is no problem with full/low speed packets as each
> downstream port receives data at target device speed.
>
> I am not sure if the capture problem can be solved with current
> OpenVizsla and LambdaConcept USB2 sniffer design. If possible, then the
> necessary changes would require gateware modifications.
>
> --- End Message ---
_______________________________________________
tcpdump-workers mailing list
tcpdump-workers () lists tcpdump org
https://lists.sandelman.ca/mailman/listinfo/tcpdump-workers