Re: [PATCH net-next v2 3/3] ptp: ocp: support the WiWi revision 02 Time Card
From: Ahmad Byagowi
Date: Mon Nov 16 2026 - 13:21:34 EST
Thanks for catching the second path to the flash. Blocking devlink alone
was not enough. In the proposed v3, ptp_ocp does not register the WiWi SPI
controller or enable its interrupt. That means no spi-nor or raw MTD for
this card through ptp_ocp, including no MTD reads. Devlink continues to
reject the operation. The existing boards retain their flash devices.
The static board match is 2163:1008, subsystem 10ee:0007. The class and
programming interface are not part of that match. Probe resolves WiWi
against the static table, so a dynamic ID cannot select a different
resource layout. PCI revision 02 identifies the register interface; we
still validate BAR and MSI-X geometry before touching the FPGA. The WiWi
XIIC child now has typed platform data and its own matching size. No cast
of another controller's data or stack copy is involved.
I cannot infer GNSS2 wiring from the PCI revision. The proposed driver
therefore exposes neither a second UART nor a GNSS2 SMA output on this
board. It no longer prints GNSS2 as a fallback MAC PPS2 source. The
independent external SMA PPS2 input is still selectable. This is a
supported-feature policy, not a statement about untested builds.
I also clarified that devlink reports the build ID as fw and does not
distinguish an active image from a fallback boot. Unknown revisions are
logged with two hex digits. The revised objects compile and the
board/flash policy checks pass; this revision has not been run on the
card.
v3:
https://lore.kernel.org/r/20261116-wiwi-v3-87afa10a8-0-ahmadexp@xxxxxxxxx