Re: port-riscv/60482

On 24/08/2026 17:10, Jeff Rizzo via gnats wrote:

[...]

 
 I was unable to duplicate this with NetBSD-11.0 (release), which
 should not be different than 11.0_RC6 in this regard - so perhaps
 this is related to the specific nvme?  Here is the relevant
 section from my dmesg:
 
 [   1.0000000] pci2 at jh7110pcie1 bus 0
 [   1.0000000] ppb1 at pci2 dev 0 function 0: vendor 1556 product 1111 (rev. 0x02)
 [   1.0000000] ppb1: PCI Express capability version 2 <Root Port of PCI-E Root Complex> x1 @ 5.0GT/s
 [   1.0000000] pci3 at ppb1 bus 1
 [   1.0000000] nvme0 at pci3 dev 0 function 0: vendor 144d product a809 (rev. 0x00)
 [   1.0000000] nvme0: NVMe 1.3
 [   1.0000000] nvme0: interrupting at plic0 irq 57
 [   1.0000000] nvme0: SAMSUNG MZALQ256HAJD-000L1, firmware BL1QFXV7, serial S4YDNF0NC73147
 [   1.0000000] ld4 at nvme0 nsid 1
 [   1.0000000] ld4: 238 GB, 31130 cyl, 255 head, 63 sec, 512 bytes/sect x 500118192 sectors (0 bytes/physsect)
 
 ...and it gets to multiuser without issue.
 
 The obvious difference I see is that my NVMe is version 1.3, not 1.4 
 (like all the ones that cause the hang).

I used to see these semi-regularly when DEBUG was enabled, but it’s not disabled by default. DEBUG was doing some pretty heavyweight checks which meant timers could get blocked for too long.