On Thu, 30 May 2019 at 06:51, Richard Cochran [off-list ref] wrote:

On Wed, May 29, 2019 at 11:33:31PM +0300, Vladimir Oltean wrote:

quoted

I would like to avoid keeping meta frames in their own RX queue,
because then I'm complicating (or rather put, making impossible) the
association between a meta frame and the frame it holds a timestamp
of.

We have an example of how a driver can match meta time stamp packets
with received packets.  See drivers/net/phy/dp83640.c to see how it
can be done completely within the driver.

Thanks,
Richard

The meta frames generated by the SJA1105 do not contain any seqid.
They contain:
* A globally programmable DMAC
* A globally programmable SMAC
* The 0x8 EtherType
* A partial (24-bit or 32-bit) RX timestamp
* Two bytes from the initial (pre follow-up) frame's DMAC, before the
switch mangled those with the source port and switch id. The driver is
supposed to patch these bytes from the follow-up back into the initial
frame before passing them up the stack.
* The source port that generated the meta frame
* The switch id that generated the meta frame

Regards,
-Vladimir