| Message ID | 20221103210650.2325784-9-sean.anderson@seco.com |
|---|---|
| State | New |
| Headers | show |
| Series | [net-next,v2,01/11] arm64: dts: Add compatible strings for Lynx PCSs | expand |
On 11/7/22 15:22, Vladimir Oltean wrote: > On Mon, Nov 07, 2022 at 02:10:10PM -0600, Rob Herring wrote: >> On Thu, Nov 03, 2022 at 05:06:47PM -0400, Sean Anderson wrote: >> > This adds device link support for PCS devices. Both the recommended >> > pcs-handle and the deprecated pcsphy-handle properties are supported. >> > This should provide better probe ordering. >> > >> > Signed-off-by: Sean Anderson <sean.anderson@seco.com> >> > --- >> > >> > (no changes since v1) >> > >> > drivers/of/property.c | 4 ++++ >> > 1 file changed, 4 insertions(+) >> >> Seems like no dependency on the rest of the series, so I can take this >> patch? > > Is fw_devlink well-behaved these days, so as to not break (forever defer) > the probing of the device having the pcs-handle, if no driver probed on > the referenced PCS? Because the latter is what will happen if no one > picks up Sean's patches to probe PCS devices in the usual device model > way, I think. Last time [1], Saravana suggested to move this to the end of the series to avoid such problems. FWIW, I just tried booting a LS1046A with the following patches applied 01/11 (compatibles) 05/11 (device) 08/11 (link) 09/11 (consumer) =================== ============== ============ ================ Y N Y N Y Y Y Y Y Y Y N N Y Y N N N Y N and all interfaces probed each time. So maybe it is safe to pick this patch. --Sean [1] https://lore.kernel.org/netdev/CAGETcx97ijCpVyOqCfnrDuGh+SahQCC-3QrJta5HOscUkJQdEw@mail.gmail.com/
On Tue, Nov 08, 2022 at 12:56:15PM -0800, Saravana Kannan wrote: > > > Last time [1], Saravana suggested to move this to the end of the series to > > > avoid such problems. FWIW, I just tried booting a LS1046A with the > > > following patches applied > > > > > > 01/11 (compatibles) 05/11 (device) 08/11 (link) 09/11 (consumer) > > > =================== ============== ============ ================ > > > Y N Y N > > > Y Y Y Y > > > Y Y Y N > > > N Y Y N > > > N N Y N > > > > > > and all interfaces probed each time. So maybe it is safe to pick > > > this patch. > > > > Maybe? Just take it with the rest of the series. > > > > Acked-by: Rob Herring <robh@kernel.org> > > Let's have Vladimir ack this. I'm not sure if it's fully safe yet. I > haven't done the necessary fixes for phy-handle yet, but I don't know > how pcs-handle and pcsphy-handle are used or if none of their uses > will hit the chicken and egg problem that some uses of phy-handle hit. I can confirm that on today's net-next, the driver owning the pcs-handle will probe even if the PCS driver is missing. With the mention that it only does so after the driver_deferred_probe_timeout, which also in today's net-next is by default 10 seconds if CONFIG_MODULES=y.
diff --git a/drivers/of/property.c b/drivers/of/property.c index 967f79b59016..ec360a616d17 100644 --- a/drivers/of/property.c +++ b/drivers/of/property.c @@ -1318,6 +1318,8 @@ DEFINE_SIMPLE_PROP(pinctrl6, "pinctrl-6", NULL) DEFINE_SIMPLE_PROP(pinctrl7, "pinctrl-7", NULL) DEFINE_SIMPLE_PROP(pinctrl8, "pinctrl-8", NULL) DEFINE_SIMPLE_PROP(remote_endpoint, "remote-endpoint", NULL) +DEFINE_SIMPLE_PROP(pcs_handle, "pcs-handle", NULL) +DEFINE_SIMPLE_PROP(pcsphy_handle, "pcsphy-handle", NULL) DEFINE_SIMPLE_PROP(pwms, "pwms", "#pwm-cells") DEFINE_SIMPLE_PROP(resets, "resets", "#reset-cells") DEFINE_SIMPLE_PROP(leds, "leds", NULL) @@ -1406,6 +1408,8 @@ static const struct supplier_bindings of_supplier_bindings[] = { { .parse_prop = parse_pinctrl7, }, { .parse_prop = parse_pinctrl8, }, { .parse_prop = parse_remote_endpoint, .node_not_dev = true, }, + { .parse_prop = parse_pcs_handle, }, + { .parse_prop = parse_pcsphy_handle, }, { .parse_prop = parse_pwms, }, { .parse_prop = parse_resets, }, { .parse_prop = parse_leds, },
This adds device link support for PCS devices. Both the recommended pcs-handle and the deprecated pcsphy-handle properties are supported. This should provide better probe ordering. Signed-off-by: Sean Anderson <sean.anderson@seco.com> --- (no changes since v1) drivers/of/property.c | 4 ++++ 1 file changed, 4 insertions(+)