Home | Notifications | New Note | Local | Federated | Search | Logout

Note Detail

Reply to @seuros@mastodon.bsd.cafe
Abdelkader Boudih@seuros@mastodon.bsd.cafe (2026-07-23 18:40:45)
Now let me tell you about the 87th commit.

It was about FireWire.

https://github.com/freebsd/freebsd-src/commit/e7efe2f059a474f8f8e14393713071761a633e65

Most people think FireWire died when USB became famous. It didnt.

Its still used by professional audio interfaces, DV camcorders, industrial and scientific cameras, and plenty of legacy equipment that continues to do its job every day. Audio equipment don't need upgrades if they are still operating correctly. A firewire cable is superior to a new USB cable.

That where i tested it. The users just kept using old version of MacOS that don't even have updated SSH.

Together with Adrian Chadd, we dusted off the stack, rediscovered how it worked, cleaned it up, and refactored years of accumulated code.

And we didnt stop there.

We also added three new drivers:

• fwcam - Firewire webcams
• fwisound - Apple FireWire audio interfaces.
• fwdv - DV camcorders, including many classic Sony, Canon, Panasonic, and JVC models.

That means #FreeBSD can now talk directly to equipment that many operating systems have quietly abandoned.

Someone in reddit used it to dump 50 tapes of childhood to digital format. I used it to have a zoom call with FreeBSD wifi group. :)
---Reply---
Cassandrich@dalias@hachyderm.io (2026-07-23 22:44:39)
@seuros My impression of FireWire was always that it was inherently insecure and gave the device a backdoor to the host. Was this inaccurate or later fixed or is it just something you have to accept using that ecosystem?

---Replies---

Abdelkader Boudih@seuros@mastodon.bsd.cafe (2026-07-23 22:59:26)
@dalias FireWire has direct DMA access to the host’s memory (up to the 4 GB address space on most implementations), which is why it earned that reputation. In many ways, it became the scapegoat.

You know what else has DMA? PCI Express. Every GPU, NVMe SSD, NIC, and Thunderbolt device.

The real issue was that, at the time, IOMMUs either didnt exist or werent widely deployed or enabled.

The driver will use the IOMMU after few round of refactoring. My priority was to get the driver correct and working on real hardware first. And get it out the deprecation list.