STRUCT CMD_BUF_DESC(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’… struct cmd_buf_desc(9) πŸ’…

ok so first things first, this is from the Kernel API Manual β€” yes bestie, the kernel, we're in the big leagues now 🫢


✨ NAME ✨

struct cmd_buf_desc β€” the descriptor squad that manages legacy SEV command address parameters, the ones that point at buffers firmware might write into 😭


🌸 SYNOPSIS 🌸

struct cmd_buf_desc {
    u64 *paddr_ptr;
    u64 paddr_orig;
    u32 len;
    bool guest_owned;
 };

πŸŽ€ Members πŸŽ€

  • paddr_ptr β€” pointer to the address parameter inside the command buffer, the one that might need saving/restoring depending on whether a bounce buffer is in play πŸ’«. When there is a bounce buffer, the command buffer has to get updated with the address of that shiny new bounce buffer that snp_map_cmd_buf_desc allocated specifically for it, iykyk. Must be NULL if this descriptor is only an end-of-list indicator β€” no cap, that's the rule πŸ™ˆ
  • paddr_orig β€” storage for the original address parameter πŸ’–. You use it to restore the original value back into paddr_ptr when it got swapped out for the address of a bounce buffer. literally the save file for your address, periodt.
  • len β€” the length of the buffer sitting at the address originally stored in paddr_ptr πŸ“
  • guest_owned β€” true if the address maps to guest-owned pages, in which case bounce buffers are not needed πŸŽ‰ (slay, one less thing).

πŸ¦‹ SEE ALSO πŸ¦‹

Kernel file ./drivers/crypto/ccp/sev-dev.c πŸ‘‘


ok that's the whole page, you ate that πŸ’…βœ¨

source: ./drivers/crypto/ccp β€” dated October 2026 πŸ“πŸ§šπŸ’œ

πŸ“¦ where this page comes from