heck, got the same nvidia copy vbios issue on a thinkpad P51 that has a Quadro M2200 which works in my Dell Precision 7570

"...nothing works and no one knows why"

exactly

on the Dell everything works...EXCEPT: the pointer/touchpad

on the newer Dell it even locks up on boot when it tries to load the mouse driver

Is planetillumos.org moribund? (wiedi?)

Given the lack of content, is it worth continuing with it?

it's not much effort to keep running on my part. Of course I'd prefer if more people blogged like you do :)

I suspect it only shows "recent" items, so much of the time it'll be completely blank (as it was up to a few minutes ago)

Yes, I can increase that time

has anyone tripped illumos.org/msg/SUNOS-8000-J0 ? I think this is either something wrong with my CPU, my motherboard, my PCIEx16 to 4x u.2 breakout card (mobo already configured for slot speed of 4x4x4x4x) or one of the two nvme drives connected so far...

It says dev:////pci@0,0/pci1022,1483@3,4/pci15b7,5030@0 faulted and taken out of service

and diskinfo is now hanging (this is SmartOS)

I have a crash from while I was trying to zfs send/receive a pool from some SATA drives over onto a pool on the two new nvme drives.

any advice welcome. I can probably also provide access to the crash dump.

I also kind of want to panic the box right now while diskinfo is hanging, but ipmitool power diag doesn't seem to be triggering a panic and I can't remember a way to do it from the CLI.

nahamu with uadmin?

uadmin 5 1 I think did it

tsoome: yes, I just figured that out from the man pages. Thank you!

:)

I did need the hint to use uadmin.8

`reboot -d` is another way if you can get in to run it.

Uh, pci1022,1483 shows up in pci-ids.ucw.cz as a Gimltet Baseboard?! I have a X570D4U-2L2T-BCM not a Gimlet.

I clicked the wrong thing, nevermind.

Gimlet Baseboard is under there. This is the Starship/Matisse GPP Bridge

reddit.com/r/AMDHelp/comments/xl7pl…ors_from_starshipmatisse_gpp_bridge is worrying.

:(

Googling turns up illumos.org/issues/12867 with your name on it, andyf. Any ideas?

→ BUG 12867: Mis-programmed pcie bridge leaves 64-bit device unusable (Closed) | code.illumos.org/c/illumos-gate/+/732

It's the next part of the path you need, which is whatever device is under that bridge. 15b7,5030

"reboot -d" is probably what I was trying and failing to remember

Ah, so it's the SSD throwing errors...

Ugh. Thank you!

It did something that fault management didn't like - I'd stat with `fmadm faulty` and `fmdump`

yeah, I found it with fmadm faulty

I'll double check the breakout card seating and cable connections, and then see if I can isolate which SSD is sad.

Better to have it fail when freshly installed than in 6 months.

Thank you for the hand-holding, andyf!

You're welcome. There have been firmware bugs in the past in drives which have tripped fault management and have been worked around (like reporting bogus temperatures or temperature thresholds).

So next steps depend on what you discover.

I'm already a bit disappointed that the drive is branded Western Digital but has a PCI ID assigned to Sandisk.

I guess the downside of plugging storage directly into the PCIe lanes is that they can do bad things if/when they misbehave.

[illumos-gate] 15661 rtld: unused variable -- Toomas Soome <tsoome⊙mc>

For posterity, ls -l /dev/dsk/*d0 | grep 'pci@0,0/pci1022,1483@3,4/pci15b7,5030@0'

that helped me correlate the problem to be /dev/dsk/c5t001B448B4ABFFF21d0

and I can use that get the serial number from nvmeadm list

well, offline the problematic drive, attempt to do the send/receive, fmadm gets mad about the remaining drive. I guess I really do need to shut this thing down and poke at things.

Did the update FTDI driver floating around somewhere ever get integrated into gate?

I think it was jclulow's ?

sjorge: No, not yet.