xpcs_create_mdiodev(9)

💅 xpcs_create_mdiodev — from the Kernel API Manual, bestie

okay so real talk, this is a page from the Kernel API Manual about xpcs_create_mdiodev, and honestly? she's a little baddie. the whole job is making a DW xPCS instance and slapping it onto the MDIO bus at addr. that's it, that's the vibe. ✨

---

🎀 NAME 🎀

xpcs_create_mdiodev — literally just creating a DW xPCS instance with the MDIO addr, no cap 🫶

---

🌸 SYNOPSIS 🌸

struct dw_xpcs * xpcs_create_mdiodev (struct mii_bus *bus , int addr );

like, that's the whole signature. she's compact, she's clean, she ate 💅

---

🦋 ARGUMENTS 🦋

• bus — a pointer to the MDIO-bus descriptor for the device you want looked at 👀
• addr — the device's MDIO-bus ID, i.e. which spot on the bus we're talking to 🎯

---

💖 RETURN 💖

so on a good day you get back a pointer to the DW XPCS handle and everyone cheers 🎉

but if it goes sideways:

• you get -ENODEV if the PCS device genuinely couldn't be found on the bus 😭
• or you get other negative errno values, and those ones are related to the data allocation and the MDIO-bus communications. so like, allocation drama or bus-chatter drama. either way, negative errno = not slaying 🙈

ngl bestie, always check the return before you start celebrating. a pointer-shaped thing that's actually an errno is giving betrayal 💀

---

🌈 SEE ALSO 🌈

• Kernel file ./drivers/net/pcs/pcs-xpcs.c — go read the source, iykyk 🔥
• xpcs_config_eee_mult_fact(9) — her cousin
• xpcs_create_fwnode(9) — her other cousin 👑

---

that's the whole page, periodt. source/version situation is ./drivers/net/pcs, dated October 2026, and it lives in the Kernel API Manual alongside her friends. mwah, go forth and create your xPCS instances 💜🧚‍♀️