svc_rdma_re...tiple_chunks(9)

📦 version 7.3.1, in nixpkgs unstable

💅 NAME 💅

svc_rdma_read_multiple_chunks — okay bestie, this one builds RDMA Reads to pull the data item Read chunks, no cap ✨

So like, this page lives in the Kernel API Manual, and the whole thing is literally about svc_rdma_read_multiple_chunks. That's the name, that's the vibe, we move 🫶

🌸 SYNOPSIS 🌸

int svc_rdma_read_multiple_chunks (struct svc_rqst *rqstp , struct
svc_rdma_recv_ctxt *head );

🎀 ARGUMENTS 🎀

  • rqstp — the RPC transaction context, i.e. the whole situation for this call 💖
  • head — context for the ongoing I/O, keeping track of what's happening right now 🌸

🦋 DESCRIPTION 🦋

okay so here's the tea: the chunk data lands in rqstp->rq_arg as a series of contiguous pages — literally like an incoming TCP call. That's it, that's the whole plot 🧚

Return values, and bestie, memorize these, they're the whole vibe check of this function:

  • 0 — the RDMA Read WQEs were successfully built 🎉 slay
  • -EINVAL — the client provided too many chunks or segments 😭 not the client misbehaving
  • -ENOMEM — the rdma_rw context pool was exhausted 💀 we ran out
  • -ENOTCONN — posting failed (the connection is lost) 🙈 oop
  • -EIO — rdma_rw initialization failed (DMA mapping, etc) 😬 that's the DMA mapping and friends being a problem
lowkey important: these are the only outcomes listed here, so don't go making up extra ones, iykyk ✨

👑 SEE ALSO 👑

Kernel file ./net/sunrpc/xprtrdma/svc_rdma_rw.c — that's where the source literally is, go look 🔥

And the whole friend group, in order, no skips:

k bye bestie, and remember: this page is straight from ./net/sunrpc/xprtrdma and it's dated October 2026 — tagging svc_rdma_re...tiple_chunks(9) one last time so you never lose the thread 💜✨🎀

📦 where this page comes from