pt_buffer_reset_markers(9)

💅 pt_buffer_reset_markers(9) — like, the whole page, bestie

okay so this is pt_buffer_reset_markers, and it lives in the Kernel API Manual, section 9. yes it's a real manual page, yes it's about Intel PT buffers, and yes we're gonna get through it together 🫶

---

✨ NAME ✨

pt_buffer_reset_markers — place interrupt and stop bits in the buffer 🎯

---

🌸 SYNOPSIS 🌸

int pt_buffer_reset_markers (struct pt_buffer *buf , struct
perf_output_handle *handle );

---

🎀 ARGUMENTS 🎀

• buf — the PT buffer, that's the one bb 💖
• handle — the current output handle, iykyk

---

🦋 DESCRIPTION 🦋

So what this bad boy does: it places INT and STOP marks in the buffer. why? two reasons, and both are iconic:

1. to prevent overwriting old data that the consumer hasn't yet collected 🙈 — like, rude, don't touch her stuff
2. to wake up the consumer after a certain fraction of the buffer has filled up ⏰ — a little tap on the shoulder, "hey bestie, there's stuff for you"

heads up: this is only needed and sensible for non-snapshot counters. snapshot counters? not the vibe, don't bother 💀

And ngl, this whole thing obviously relies on buf::head to figure out the buffer markers — so it has to be called after pt_buffer_reset_offsets and before the hardware tracing is enabled. get the order wrong and it's giving chaos 😭

---

🔥 SEE ALSO 🔥

Kernel file ./arch/x86/events/intel/pt.c — and the whole friend group:

• struct topa(9)
• struct topa_page(9)
• topa_alloc(9)
• topa_free(9)
• topa_insert_table(9)
• topa_table_full(9)
• topa_insert_pages(9)
• pt_topa_dump(9)
• pt_buffer_advance(9)
• pt_update_head(9)
• pt_buffer_region(9)
• pt_buffer_region_size(9)
• pt_handle_status(9)
• pt_read_offset(9)
• pt_buffer_reset_offsets(9)
• pt_buffer_fini_topa(9)
• pt_buffer_init_topa(9)
• pt_buffer_setup_aux(9)
• pt_buffer_free_aux(9)
• intel_pt_interrupt(9)

---

okay that's the page!! ✨ this lil gem comes to you from ./arch/x86/events/intel, dated October 2026, and it's filed under pt_buffer_reset_markers(9) — periodt 💅🎀