dax_iomap_rw(9)

💖 dax_iomap_rw(9) 💖

okay so welcome to the Kernel API Manual, bestie — today’s entry is dax_iomap_rw(9), and it’s literally about I/O to a DAX file. let’s get into it 🫶

🌸 NAME 🌸

dax_iomap_rw — the function that performs I/O to a DAX file 💅

✨ SYNOPSIS ✨

ssize_t dax_iomap_rw (struct kiocb *iocb , struct iov_iter *iter , const
       struct iomap_ops *ops );

🎀 ARGUMENTS 🎀

• iocb — the control block for this I/O. it’s giving “paperwork for the operation” ngl 📋
• iter — the addresses to do I/O from or to. literally the from/to address situation 🦋
• ops — iomap ops passed from the file system. the filesystem hands these over, no cap 🫶

💅 DESCRIPTION 💅

okay so dax_iomap_rw performs read and write operations to directly mapped persistent memory. that’s the whole serve: read + write, directly mapped persistent memory, periodt 🔥

but bestie, the callers needs to take care of read/write exclusion and evicting any page cache pages in the region under I/O. yes, that is a caller-side responsibility — this function is not doing that part for you. no cap, do not skip it 😭

🔥 SEE ALSO 🔥

the cross-reference friend group is right here, and every single one is invited:

• Kernel file ./fs/dax.c + enum dax_wake_mode(9)
• dax_folio_reset_order(9)
• dax_lock_folio(9)
• dax_layout_busy_page_range(9)
• dax_iomap_copy_around(9)
• dax_fault_iter(9)
• dax_iomap_fault(9)
• dax_finish_sync_fault(9)

mwah 💋 source/version: ./fs, date: October 2026. that’s the sign-off, bestie — bye! 💖✨