report_iommu_fault(9)

💅 report_iommu_fault — from the Kernel API Manual, bestie ✨

okay so like, this is report_iommu_fault(9) and it lives in the Kernel API Manual, and honestly? she's giving messenger. she's giving "the low-level driver texts the framework and is like bestie, something happened." 🫶 no cap.

---

🌸 SYNOPSIS 🌸

int report_iommu_fault (struct iommu_domain *domain , struct device *dev
, unsigned long iova , int flags );

---

🎀 ARGUMENTS 🎀

• domain — the iommu domain where the fault has happened. that's her home turf 💜
• dev — the device where the fault has happened. the actual culprit, lowkey 😭
• iova — the faulting address. the exact spot. iconic of it to be that specific ✨
• flags — mmu fault flags, like IOMMU_FAULT_READ/IOMMU_FAULT_WRITE/... — basically the vibe check on what kind of fault we're dealing with 🦋

---

✨ DESCRIPTION ✨

so the way this works: the low-level IOMMU implementations should call this function whenever IOMMU faults happen, so the high-level users who are interested in such events can, like, actually know about them. communication, bestie. it's everything. 💅

this event may be useful for several possible use cases, and here's the list:

• mere logging of the event
• dynamic TLB/PTE loading
• if restarting of the faulting device is required

returns 0 on success and an appropriate error code otherwise — and like, if dynamic PTE/TLB loading will one day be supported, implementations will be able to tell whether it succeeded or not according to this return value. forward thinking. we love to see it 👑

specifically, -ENOSYS is returned if a fault handler isn't installed (though fault handlers can also return -ENOSYS, in case they want to elicit the default behavior of the IOMMU drivers). yes that's the same error from two directions, it's a whole thing, don't @ me 💀

bestie pls note: that "one day be supported" bit is exactly how the original says it — I'm not making promises about dynamic PTE/TLB loading, I'm just reporting the message fr fr 🙈

---

🦋 SEE ALSO 🦋

kernel file ./drivers/iommu/iommu.c — plus the whole friend group:

iommu_device_register(9), iommu_insert_resv_region(9), iommu_group_alloc(9), iommu_group_get_iommudata(9), iommu_group_set_iommudata(9), iommu_group_set_name(9), iommu_group_add_device(9), iommu_group_remove_device(9), iommu_group_mutex_assert(9), iommu_group_for_each_dev(9), iommu_group_get(9), iommu_group_ref_get(9), iommu_group_put(9), iommu_group_id(9), device_iommu_capable(9), iommu_group_has_isolated_msi(9), iommu_set_fault_handler(9), iommu_paging_domain_alloc_flags(9), iommu_attach_device(9), iommu_get_domain_for_dev(9), iommu_driver_get_domain_for_dev(9), iommu_attach_group(9), iommu_unmap(9), iommu_unmap_fast(9), iommu_get_resv_regions(9), iommu_put_resv_regions(9), iommu_setup_default_domain(9), iommu_device_use_default_domain(9), iommu_device_unuse_default_domain(9), iommu_group_claim_dma_owner(9), iommu_device_claim_dma_owner(9), iommu_group_release_dma_owner(9), iommu_device_release_dma_owner(9), iommu_group_dma_owner_claimed(9), iommu_replace_device_pasid(9), iommu_attach_handle_get(9), iommu_attach_group_handle(9), iommu_detach_group_handle(9), iommu_replace_group_handle(9), pci_dev_reset_iommu_prepare(9), pci_dev_reset_iommu_done(9), iommu_dma_prepare_msi(9)

like. that's the whole group chat. literally all of them 🍓

---

🌈 sign-off 🌈

this little masterpiece comes from ./drivers/iommu, dated October 2026, and yes she's titled report_iommu_fault(9) because the main character always gets her name in the footer 💖 stay faultless bestie 🧚✨