xe_validation_retry_on_oom(9)

💅 xe_validation_retry_on_oom — the Kernel API Manual, bestie edition ✨

okay so real talk, this one lives in the Kernel API Manual and it's called xe_validation_retry_on_oom(9) and it does exactly one dramatic thing: it gives your transaction a second chance when memory goes poof. 🫶

---

🌸 NAME 🌸

xe_validation_retry_on_oom — Retry on oom in an xe_validaton transaction

not me pointing out that "xe_validaton" is their spelling, not mine 😭 i'm just the messenger, i don't edit the lore

---

✨ SYNOPSIS ✨

xe_validation_retry_on_oom (_ctx , _ret );

---

🎀 ARGUMENTS 🎀

• _ctx — Pointer to the xe_validation_ctx 👑 like, the whole context of your validation situation
• _ret — The current error value, possibly holding -ENOMEM 💀 this is the little guy that tells you whether you ran out of memory

---

🦋 DESCRIPTION 🦋

Use this in a way similar to drm_exec_retry_on_contention. 💫

So here's the tea: if _ret contains -ENOMEM, the transaction gets restarted once — and it's restarted in a way that blocks other transactions and allows exhastive eviction. 🔥

(yes, "exhastive" is exactly how it's written in the original, she's built different, i'm not touching it 💀)

But — and this is the important part bestie — if the transaction was already restarted once, Just return the -ENOMEM. no third chances, no extra chaos, we're done. 🙈

It may also set _ret to -EINTR if it's not retrying and waits are interruptible. ⭐

🚨 and the big one: this may only be used within a drm_exec_until_all_locked loop. like, that's the rule, that's the vibe check, do not freelance this one 😭

---

🌈 SEE ALSO 🌈

• Kernel file ./drivers/gpu/drm/xe/xe_validation.h 📁
• xe_validation_lockdep(9)
• struct xe_validation_device(9)
• struct xe_val_flags(9)
• struct xe_validation_ctx(9)
• xe_validation_device_init(9)
• xe_validation_guard(9)

---

mwah 💖 that's the whole page, go forth and retry responsibly. this icon was brought to you by ./drivers/gpu/drm/xe, dated October 2026 — slay, see you in the next one 🎀🦋✨