blk_mq_try_issue_directly(9)

✨ NAME ✨

blk_mq_try_issue_directly — straight from the Kernel API Manual, this is the function that tries to send a request directly to the device driver. That's the whole name vibe, bestie. 💅

🎀 SYNOPSIS 🎀

void blk_mq_try_issue_directly (struct blk_mq_hw_ctx *hctx , struct request *rq );

🫶 ARGUMENTS 🫶

• hctx — pointer of the associated hardware queue. literally the hardware queue's pointer.
• rq — pointer to the request to be sent. the request in question, periodt.

🔥 DESCRIPTION 🔥

Okay so here's the tea: if the device has enough resources to accept a new request now, this function sends the request directly to the device driver. Slay, direct path. ✨

Otherwise? It inserts the request at the hctx->dispatch queue, so we can try to send it another time in the future. Requests inserted at this queue have higher priority. No cap. 👑

🌈 SEE ALSO 🌈

The kernel file is ./block/blk-mq.c, and these are the besties in the same orbit:

• blk_mq_wait_quiesce_done(9)
• blk_mq_quiesce_queue(9)
• blk_update_request(9)
• blk_mq_complete_request(9)
• blk_mq_start_request(9)
• blk_execute_rq_nowait(9)
• blk_execute_rq(9)
• blk_mq_delay_run_hw_queue(9)
• blk_mq_run_hw_queue(9)
• blk_mq_run_hw_queues(9)
• blk_mq_delay_run_hw_queues(9)
• blk_mq_request_bypass_insert(9)
• blk_mq_submit_bio(9)
• blk_insert_cloned_request(9)
• blk_rq_unprep_clone(9)
• blk_rq_prep_clone(9)
• blk_mq_destroy_queue(9)

Cute little sign-off from the source/version footer: ./block, October 2026, blk_mq_try_issue_directly(9) — mwah 💖