ATA over Ethernet (AoE) client for Coraid EtherDrive storage
Linux-side support for ATA over Ethernet, a lightweight layer-2 protocol that lets a machine use disks in a remote box as if they were local ATA drives. Its main commercial home was Coraid's EtherDrive SAN appliances, popular in budget and lab storage setups through the 2000s and early 2010s.
recommendation
Worth keeping but flagging as a niche: Coraid, the company behind the EtherDrive appliances this targets, went under in 2015 and the hardware is no longer sold new, so deployments today are limited to legacy SAN installs and labs. However, the code is not abandoned, with real maintenance work in 2024 and a stable backport in 2025, and because AoE is its own protocol with no drop-in in-tree replacement, removing it would strand existing users.
repository signals
sources
- lore.kernel.org
AoE still receives upstream bug-fix traffic, including a 2025 stable backport for a potential deadlock in set_capacity.
- lore.kernel.org
AoE saw non-trivial maintenance work in 2024, including net_device lifetime/reference tracking changes.
- en.wikipedia.org
EtherDrive was the Coraid AoE storage product line; the page records Coraid's 2015 foreclosure/IP transfer, indicating the original hardware line is long out of its mainstream commercial era.
- en.wikipedia.org
AoE is a niche layer-2 storage protocol with native Linux support and historically limited vendor ecosystem centered on Coraid/EtherDrive.
codex reasoning notes (technical)
Real driver: Kconfig/Makefile in-tree identify it as 'ATA over Ethernet' support for Coraid EtherDrive (found via local rg). Lore evidence came from mcp__lore_http__.lore_activity on drivers/block/aoe/aoeblk.c after lore_file_timeline on the directory returned no hits; the 2024-2025 lore URLs show ongoing fixes/adaptation, so removal/deprecation would be premature. Web URLs were obtained via web.search_query; they are historical/product-context evidence, not proof of current retail availability. I infer 'not still sold new in 2025' and last widespread availability around 2015 from the dated public product history plus absence of current mainstream vendor evidence in the search budget. Deployments are likely low: legacy/lab/industrial Linux SAN niches rather than new broad deployments. No natural in-kernel replacement exists for the same AoE protocol, so annotate rather than deprecate/remove.