Re: ahcisata0: reset failed
Taylor R Campbell <[email protected]> Mon, 6 Jul 2026 18:54:40 +0000
| Newsgroups | gmane.os.netbsd.current |
|---|---|
| Message-ID | <[email protected]> |
> From: Thomas Klausner <[email protected]> > Date: Mon, 6 Jul 2026 20:36:40 +0200 >=20 > I've just updated to today's -current (20260706), and most of my disks > are not found. >=20 > In dmesg, I see the following changes compared to about three weeks ago: >=20 > +MPACPI PCI bus number 11 configuration failed: 6 >=20 > -ahcisata0: AHCI revision 1.31, 6 ports, 32 slots, CAP 0xef36ff25<SXS,PSC= ,SSC,PMD,SPM,SAM,ISS=3D0x3=3DGen3,SCLO,SAL,SALP,SSS,SSNTF,SNCQ,S64A> > -ahcisata0: interrupting at msi4 vec 0 > -atabus0 at ahcisata0 channel 0 > -atabus1 at ahcisata0 channel 1 > -atabus2 at ahcisata0 channel 2 > -atabus3 at ahcisata0 channel 3 > +autoconfiguration error: ahcisata0: reset failed Thanks, likely caused by enabling PCI_RESOURCE, in which case the workaround is to add `no options PCI_RESOURCE' to your kernel config. But we need to get PCI_RESOURCE working out of the box. Can you file a PR with: 1. dmesg with `boot -vx' with and without PCI_RESOURCE, and 2. `pcictl pci0 dump -b <bus> -f <func> -d <dev>' with and without PCI_RESOURCE, where <bus>/<func>/<dev> are according to wherever ahcisata0 attaches at PCI?