kset_onmedia_valid(9)

📦 version 7.3.1, in nixpkgs unstable

💅 kset_onmedia_valid(9) — like, the vibe check for your kset header 🎀

okay so bestie, welcome to the kset_onmedia_valid(9) page from the Kernel API Manual — yes it's a kernel manpage, no it's not boring, stay with me 🫶

✨ NAME ✨

kset_onmedia_valid — validate a kset header that got read in from the cache device. that's it, that's the job, and honestly? iconic 💅

🌸 SYNOPSIS 🌸

bool kset_onmedia_valid (struct pcache_cache_kset_onmedia *kset_onmedia
);

yeah she takes one arg and hands you back a bool. simple queen behavior 👑

🦋 ARGUMENTS 🦋

  • kset_onmedia — pointer to the kset that was copied straight off the on-media metadata. like, this is the thing you're checking, bestie, be nice to it 🥺

🔥 DESCRIPTION 🔥

okay so here's the tea, and it's kind of a lot ☕

the magic and the CRC are both attacker-computable — because the seed is a fixed public seed. so like, don't be out here trusting them with your life, they're giving "security theater" energy ngl 💀

here's the actual danger zone: a non-last kset stores key_num keys inline. and both cache_kset_crc and the replay loop read struct_size(.., data, key_num) bytes out of a buffer that's only sized for PCACHE_KSET_KEYS_MAX keys. so key_num must be bounded before any of that use happens. not after, not "eventually", before. periodt 🚨

⚠️ like seriously bestie, bound key_num first. this is not a drill, this is a "read past the end of a buffer" situation and that's the bad kind of main character energy 😭

🧚 SEE ALSO 🧚

kernel file ./drivers/md/dm-pcache/cache.h plus the whole friend group:

go read those too, they're all slay in their own way 💖

and that's the page!! brought to you by ./drivers/md/dm-pcache, dated October 2026, filed forever under kset_onmedia_valid(9) — mwah, go validate responsibly ✨🎀💜

📦 where this page comes from