[ISSUE] ata3 softreset failed and slow reboot

Started by tle, November 03, 2022, 11:52:32 PM

Previous topic - Next topic

tle

I've noticed that since version 5.15, I've seen quite a number of following error:

ata3: softreset failed (1st FIS failed)

whenever I reboot the computer. This error takes at least 10 seconds til:

ata3: reset failed, giving up

Just wondering if anyone has bumped into similar error
Faithful Linux enthusiast

My Raptor Blackbird

ClassicHasClass

Yes, with my SATA optical drives. It's kept me on 5.13 because it won't see them at all.

@sharkcz have you run into this?

rheaplex


sharkcz

I have searched thru my whole journal and there is no such message. My SATA controller is Marvell 88SE9128 (RAID controller with HyperDuo) with a SATA SSD (the on-board Adaptec SAS controller didn't like it much) and an optical drive connected.

sharkcz

Also, can the start of these messages/errors be tied to a specific kernel version? What are your the SATA controllers?

ClassicHasClass

The SATA controller in this unit is another Marvell, but a different chipset (lspci says 0001:01:00.0 SATA controller: Marvell Technology Group Ltd. 88SE9235 PCIe 2.0 x2 4-port SATA 6 Gb/s Controller (rev 11)). I believe this is the one Raptor sells. It has two optical drives connected; the main volumes / and /home are NVMe SSDs.

sharkcz

I suspect the way out is to pinpoint the issue to a specific kernel version(s) and then bisect ... AFAIK the AmigaOne people are running into similar kind of issues as the mainline kernel changes and the kernel maintainers are listening to them.

ClassicHasClass

tle and rheaplex, are you using the same Marvell chipset?

rheaplex

My controller (from Raptor) is:

0001:01:00.0 SATA controller: Marvell Technology Group Ltd. 88SE9215 PCIe 2.0 x1 4-port SATA 6 Gb/s Controller (rev 11)

tle

Mine is

0002:01:00.0 SATA controller: Marvell Technology Group Ltd. 88SE9235 PCIe 2.0 x2 4-port SATA 6 Gb/s Controller (rev 11)
Faithful Linux enthusiast

My Raptor Blackbird

sharkcz

It looks to me that mine "9128" is a different generation than yours 9215/9235, it might explain the different behaviour.