-current: very slow boot

[Thomas Klausner <wiz%NetBSD.org@localhost>]

Hi!

I see very slow boot behaviour on 8.99.42/amd64.
The first line has timestamp 1 (in /var/log/messages):

    /netbsd: [   1.0000000] Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005,

then booting proceeds at a fast pace until:

     /netbsd: [   1.0020642] nouveau0: info: disp:      0430: 0000e900
     /netbsd: [   1.0020642] timecounter: Timecounter "clockinterrupt" frequency 100 Hz quality 0
     /netbsd: [   3.2952236] autoconfiguration error: ERROR: 876 cycle TSC drift observed
     /netbsd: [   3.3384808] uhub0 at usb0: NetBSD (0000) xHCI root hub (0000), class 9/0, rev 3.00/1.00, addr 0

so there's a two-second jump without obvious activity. Then it continues:

/netbsd: [   3.3384808] uhub5: 4 ports with 4 removable, self powered
/netbsd: [   3.6354389] ahcisata0 port 5: device present, speed: 6.0Gb/s
/netbsd: [   4.4659634] autoconfiguration error: ahcisata0 port 5: clearing WDCTL_RST failed for drive 15
/netbsd: [   4.9662796] autoconfiguration error: ahcisata0 port 5: drive 15 reset failed, disabling PMP

Another 1,5s gone.

/netbsd: [   5.0663430] ahcisata0 port 5: device present, speed: 6.0Gb/s
/netbsd: [   5.0663430] uhidev0 at uhub5 port 1 configuration 1 interface 0
/netbsd: [   5.0663430] uhidev0: vendor 046a (0x46a) product 0023 (0x23), rev 2.00/2.20, addr 1, iclass 3/1
/netbsd: [   5.1163748] ukbd0 at uhidev0: 8 Variable keys, 6 Array codes
/netbsd: [   5.5266342] wskbd0 at ukbd0: console keyboard, using wsdisplay0
/netbsd: [   5.5266342] uhidev1 at uhub5 port 1 configuration 1 interface 1
/netbsd: [   5.5266342] uhidev1: vendor 046a (0x46a) product 0023 (0x23), rev 2.00/2.20, addr 1, iclass 3/0
/netbsd: [   5.5566528] uhidev1: 2 report ids
/netbsd: [   5.5566528] uhid0 at uhidev1 reportid 1: input=2, output=0, feature=0
/netbsd: [   5.5666593] uhid1 at uhidev1 reportid 2: input=1, output=0, feature=0
/netbsd: [   6.0669748] wd0 at atabus5 drive 0
/netbsd: [   6.0669748] wd0: <Samsung SSD 860 EVO 250GB>
/netbsd: [   6.0669748] wd0: drive supports 1-sector PIO transfers, LBA48 addressing
/netbsd: [   6.0669748] wd0: 232 GB, 484521 cyl, 16 head, 63 sec, 512 bytes/sect x 488397168 sectors
/netbsd: [   6.0669748] wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133), WRITE DMA FUA, NCQ (32 tags)
/netbsd: [   6.0669748] wd0(ahcisata0:5:0): using PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133) (using DMA), NCQ (31 tags)
/netbsd: [  21.5267510] IPsec: Initialized Security Association Processing.
/netbsd: [  21.5367503] WARNING: 12 errors while detecting hardware; check system log.

And here it loses 15s!?

What's worse, this doesn't match wallclock time. When the last nouveau
line is printed, there are ~40s pause without a change (except a short
blackout (1-2s) of the screen) and then another ~20s blackout of the
screen, after which the console has been switched to graphical mode
and the kernel part of the boot has finished.

The kernel config is quite simple:

--- begin quote---
include "arch/amd64/conf/GENERIC"

# Kernel Address Sanitizer
makeoptions 	KASAN=1
options 	KASAN
no options	SVS

options FONT_GO_MONO12x23
no options FONT_BOLD16x32
no options FONT_BOLD8x16

no nvme*
--- end quote ---

Any ideas?
 Thomas