On 3/19/23 21:02, Elmar Psilog wrote:

Am 19.03.23 um 18:21 schrieb Marek Vasut:
On 3/19/23 11:43, Elmar Psilog wrote:
Without that patch it lost track to the node to scan
speed and duplex.
Patch was created by Marek Vasut, just tested by me.

Signed-off-by: Elmar Psilog <epsi at gmx.de>
Reviewed-by: Marek Vasut <marex at denx.de>
---
   changes v2: fix format issues
   changes v3: remove {} around single if, add reviewd..

  drivers/net/dwc_eth_qos.c | 18 +++++++++++++++---
  1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/net/dwc_eth_qos.c b/drivers/net/dwc_eth_qos.c
index 112deb546d..b9de205b8a 100644
--- a/drivers/net/dwc_eth_qos.c
+++ b/drivers/net/dwc_eth_qos.c
@@ -788,9 +788,21 @@ static int eqos_start(struct udevice *dev)
       */
      if (!eqos->phy) {
          int addr = -1;
-        addr = eqos_get_phy_addr(eqos, dev);
-        eqos->phy = phy_connect(eqos->mii, addr, dev,
-                    eqos->config->interface(dev));
+        ofnode fixed_node;
+
+        if (IS_ENABLED(CONFIG_PHY_FIXED)) {
+            fixed_node = ofnode_find_subnode(dev_ofnode(dev),
+                             "fixed-link");
+            if (ofnode_valid(fixed_node))
+                eqos->phy = fixed_phy_create(dev_ofnode(dev));
+        }
+
+        if (!eqos->phy) {
+            addr = eqos_get_phy_addr(eqos, dev);
+            eqos->phy = phy_connect(eqos->mii, addr, dev,
+                        eqos->config->interface(dev));
+        }
+
          if (!eqos->phy) {
              pr_err("phy_connect() failed");
              goto err_stop_resets;
--
2.34.1


I would give a short reminder about the patch. Reviewed by Marek, ok for
Ramon. Would be glad to see it in 2023.04.

We're already in 2023.04-rc4 , this is material for v2023.07 , i.e. for u-boot/next branch .

Ok, but does it change anything for the patch itself?

No, it will be applied to u-boot-net / next and then proceed via PR to u-boot / next .

As source isn't touched in last weeks it is still needed. Other way asked: Any action (from my side) needed?

No, it is up to Ramon now.

I just afraid your (@Marek) and my work gets lost. I would expect either an "accepted" or "rejected" (for a reason).

It's just overloaded/busy maintainers problem, sadly, this does happen from time to time. The patch is tracked by patchwork, so it shouldn't be lost.

Reply via email to