si_dma_vm_copy_pages(9)
β¨ si_dma_vm_copy_pages(9) β¨
okay so bestie, welcome to the si_dma_vm_copy_pages page β it's in section 9 of the kernel API manual, and it's honestly a little baddie that just moves page table entries around π
π NAME π
si_dma_vm_copy_pages β update PTEs by copying them from the GART π«Ά
πΈ SYNOPSIS πΈ
void si_dma_vm_copy_pages (struct radeon_device *rdev , struct radeon_ib
*ib , uint64_t pe , uint64_t src , unsigned count );
like, it literally returns void, no cap β it's not coming back to tell you how it went π
π ARGUMENTS π
rdevβ aradeon_devicepointer, the whole device vibe check πibβ the indirect buffer you're filling with commands, it's giving "the actual to-do list" β¨peβ addr of the page entry, that's where we're writing, periodtsrcβ src addr where to copy from, the origin story π¦countβ number of page entries to update, so like how many we're doing π
π₯ DESCRIPTION π₯
Update PTEs by copying them from the GART using the DMA (SI). π§
That's the whole thing bestie, that's the assignment β copy the page table entries straight out of the GART with the DMA on SI. iconic, lowkey efficient, ate π
β SEE ALSO β
- Kernel file
./drivers/gpu/drm/radeon/si_dma.c si_dma_is_lockup(9)si_dma_vm_write_pages(9)si_dma_vm_set_pages(9)si_copy_dma(9)
go read those too, iykyk π
ngl the source lives in./drivers/gpu/drm/radeonand this one's dated October 2026 β so like, future girlie things fr fr. anyway that's the wholesi_dma_vm_copy_pages(9)page, mwah π π
π¦ where this page comes from
- nixpkgs unstable: linux-manual 7.3.1
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on : a language model's rewrite of the original, under its terms