Has anyone been able to boot using a USB3 port?

I just installed FreeBSD 10.2 on a USB3 stick (Sandisk Extreme 3.0). But it tries to boot off it, it just doesn't work. I was thinking the XHCI might need to be statically-linked on the kernel but I'm using GENERIC and it's definitely included.

Has anyone been able to boot using a USB3 port? If yes, any tips?

My motherboard is AMD A85-based and all the USB3 ports are supported in the AMD chipset. Definitely not 3rd-party.

Thanks a lot
 
Hey !
What do you mean by booting on a USB3 port ?
Are trying to boot to launch an install or just live ?

Have you checked on the BIOS for an "USB emulation type" option ?

I've never had problems booting usb 3 devices on usb 3 ports
 
Hey !
What do you mean by booting on a USB3 port ?
Are trying to boot to launch an install or just live ?

Have you checked on the BIOS for an "USB emulation type" option ?

I've never had problems booting usb 3 devices on usb 3 ports
I installed FreeBSD on a USB3 stick which is plugged in a USB3 port. When I tried booting off the stick, it doesn't work. like it cannot find the partition.

I'll take a pic of the error and I'll look into that BIOS setting.

Thanks! :)
 
I have a small FreeBSD install that is meant to be portable. It is made to boot off almost any hardware, uses GPT labels to locate the filesystem, uses numerous video cards, uses the first available Ethernet card, and so on. Recently, I put that on a 32G USB 3 memory stick and was surprised to find it booted just fine on a USB 3 system. Better than just fine, really, since USB 2 boots slowly and has slow transfer rates after booting. USB 3 was very usable.

There are still some incompatibility issues with various USB 3 hardware, so it might not work on some systems. But it worked on that one, which was an MSI motherboard with an Intel Haswell chipset.
 
I just tried to do a default install 10.1 on an disk that is in an external USB3 enclosure. It didn't work because there are bugs either in the BIOS of the machine, the enclosure, loader(8) or FreeBSD kernel that causes a disk geometry mismatch and that invalidates the GPT partition table on boot resulting in "error 19" and a mountroot prompt. The installer clearly has a different idea about the disk geometry (number of LBA sectors) than the system that gets installed on the disk. I'll probably try again with MBR partitioning but the disk geometry problems just shouldn't happen, GPT is not the only system that depends on the correct geometry troughout the process.
 
Bugs in the firmware of a USB enclosure have been known to happen. Sometimes it's an off-by-one error on the disk capacity. There are some enclosures that use on-disk metadata, and hide that area by subtracting it from the total capacity. Depending on the bug, it could help to partition the drive only after it is in the enclosure.
 
I boot 10.2-RELEASE on a ThinkPad x230 model 2325RU0 from a SanDisk Ultra Fit over a USB 3.0 port.

It's the only directly-attached storage I use on the computer, and it works fine; its boot messages are here.
 
The problem was lack of labels on the GPT partitions. The installer was installing on disk da1 and the memory stick I used for booting was da0. After the installation the disk in the enclosure was naturally da0 when the USB memory stick had been removed but the /etc/fstab was refering to da1. You would think use of labels should be the default by now?
 
Back
Top