__flush_cache_one(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’– NAME πŸ’–

Okay so this is __flush_cache_one(9) from the Kernel API Manual, and we're diving in bestie.

  • __flush_cache_one β€” that's the name, and the original page leaves the description blank. Not me being mysterious, it's literally just the name and a dash. No cap. πŸ’…

✨ SYNOPSIS ✨

Here's the whole signature, and yes you should copy it exactly:

void __flush_cache_one (unsigned long addr , unsigned long phys ,
       unsigned long exec_offset );

πŸŽ€ ARGUMENTS πŸŽ€

  • addr β€” address in memory mapped cache array. That's the vibe: it's an address sitting in the memory mapped cache array. πŸ¦‹
  • phys β€” P1 address to flush. And bestie, listen: it has to match tags if addr has the 'A' bit set, i.e. associative write. That's a real condition, not a suggestion. πŸ™ˆ
  • exec_offset β€” set to 0x20000000 if the flush has to be executed from the P2 region, else 0x0. Two options, icon behavior only. ✨

πŸ¦‹ DESCRIPTION πŸ¦‹

The description on this page is giving "sentence that walked offstage mid-sentence," and I'm not gonna pretend otherwise. Here's the original line in all its unfinished glory:

The offset into the cache array implied by 'addr' selects the operation (purge/write-back) is selected by the lower 2 bits of

Yep. It literally ends on "of." That's the whole sentence, no extra words from me, because inventing the rest would be a choice and we don't do that here. πŸ˜­πŸ’€


🌸 SEE ALSO 🌸

  • Kernel file ./arch/sh/mm/cache-sh4.c β€” that's the kernel file the page points you at. Go give it a look, iykyk. 🫢

That's the whole page, bestie πŸ’œ β€” brought to you by ./arch/sh/mm, dated October 2026, and it's __flush_cache_one(9) forever and always. Slay. πŸ’…βœ¨

πŸ“¦ where this page comes from