dmaengine_p...nfig_sg_safe(9)

💅 NAME 💅

dmaengine_prep_config_sg_safe — bestie, this one prepares a scatter-gather DMA transfer and does an atomic slave configuration update at the same time ✨

---

✨ SYNOPSIS ✨

struct dma_async_tx_descriptor * dmaengine_prep_config_sg_safe (struct
dma_chan *chan , struct scatterlist *sgl , unsigned int sg_len , enum
dma_transfer_direction dir , unsigned long flags , struct
dma_slave_config *config );

---

🎀 ARGUMENTS 🎀

• chan — the DMA channel. your main character 🫶
• sgl — the scatterlist for the transfer. it's giving "here's all the places the data lives" 🌈
• sg_len — literally how many entries are in sgl. count 'em, don't guess 💅
• dir — the DMA transfer direction 🔥
• flags — transfer preparation flags ⭐
• config — the DMA slave configuration for this transfer 🦋

---

🌸 DESCRIPTION 🌸

okay so picture this: you wanna prepare a DMA scatter-gather transfer, and you need the corresponding slave configuration update to happen with it — and you need the whole thing to be re-entrant and race-safe. that's the whole vibe. no cap 💖

DMA engine drivers can implement the optional device_prep_config_sg callback to do both the slave configuration and the descriptor preparation atomically. if they do, the operation is fully handled by the DMA engine driver. iconic behavior, honestly 👑

but if the DMA engine driver doesn't implement device_prep_config_sg, it falls back to calling dmaengine_slave_config followed by dmaengine_prep_slave_sg. and listen — that fallback path is protected by a per-channel spinlock, so concurrent callers can't interleave configuration and descriptor preparation on the same DMA channel. the way it just locks it down 😭🔒

---

⭐ RETURN ⭐

You get a pointer to a prepared DMA async transaction descriptor on success — slay. Or NULL if the transfer could not be prepared. we don't love that, but we respect it 🙈

---

🦋 SEE ALSO 🦋

All of this lives in the kernel file ./include/linux/dmaengine.h, and it lowkey has the whole friend group with it:

• dma_cookie_t(9), enum dma_status(9), enum dma_transaction_type(9), enum dma_transfer_direction(9), struct data_chunk(9), struct dma_interleaved_template(9), struct dma_vec(9)
• enum dma_ctrl_flags(9), enum sum_check_bits(9), enum sum_check_flags(9), enum dma_desc_metadata_mode(9), struct dma_chan_percpu(9), struct dma_router(9)
• struct dma_chan(9), struct dma_chan_dev(9), enum dma_slave_buswidth(9), struct dma_slave_config(9), enum dma_residue_granularity(9), struct dma_slave_caps(9)
• dma_filter_fn(9), struct dma_async_tx_descriptor(9), struct dma_tx_state(9), enum dmaengine_alignment(9), struct dma_slave_map(9), struct dma_filter(9), struct dma_device(9)
• dmaengine_prep_peripheral_dma_vec(9), dmaengine_prep_config_single_safe(9), dmaengine_prep_dma_memset(9)
• dmaengine_terminate_all(9), dmaengine_terminate_async(9), dmaengine_synchronize(9), dmaengine_terminate_sync(9)
• dma_async_issue_pending(9), dma_async_is_tx_complete(9), dma_async_is_complete(9)

---

and that's the page, bestie 🎀 this lil entry comes straight out of the ./include/linux tree, October 2026, Kernel API Manual era — go forth and prep your scatter-gather transfers safely, ily, mwah 💖✨🦋