Re: NetBSD 10 on Cobalt doesn't like IDE DMA

[Havard Eidnes <he%NetBSD.org@localhost>]

> There's one big regression: viaide0 can't do DMA. Whereas the Raq has
> run for years with no problems, with the new kernel I get a steady
> stream of:
>
> [ 1370.2451615]         type: ata tc_bcount: 32768 tc_skip: 0
> viaide0:0:0: bus-master DMA error: missing interrupt, status=0x20
> [ 1370.2584021] wd0e: DMA error reading fsbn 922830144 of
> 922830144-922830207 (wd0 bn 927046976; cn 919689 tn 7 sn 23), xfer
> 1f88, retry 0
> [ 1370.7551615] wd0: soft error (corrected) xfer 1f88

I also have a Cobalt Raq2, and it runs 9.3_STABLE dated May 10
2023.  However, it also suffers from "degradation to PIO" shortly
after boot:

wd0 at atabus0 drive 0
wd0: <Kingchuxing 256GB>
wd0: drive supports 1-sector PIO transfers, LBA48 addressing
wd0: 238 GB, 496149 cyl, 16 head, 63 sec, 512 bytes/sect x 500118192 sectors
wd0: 32-bit data port
wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133), WRITE DMA FUA
wd0(viaide0:0:0): using PIO mode 4, Ultra-DMA mode 2 (Ultra/33) (using DMA), WRITE DMA FUA EXT
boot device: wd0
root on wd0a dumps on wd0b
autoconfiguration error: viaide0:0:0: lost interrupt
        type: ata tc_bcount: 8192 tc_skip: 0
autoconfiguration error: viaide0:0:0: bus-master DMA error: missing interrupt, status=0x20
wd0: autoconfiguration error: transfer error, downgrading to Ultra-DMA mode 1
wd0(viaide0:0:0): using PIO mode 4, Ultra-DMA mode 1 (using DMA), WRITE DMA FUA EXT
wd0a: DMA error reading fsbn 16 of 16-31 (wd0 bn 6289936; cn 6240 tn 0 sn 16), xfer f70, retry 0
wd0: soft error (corrected) xfer f70
root file system type: ffs
kern.module.path=/stand/cobalt/9.3/modules
autoconfiguration error: viaide0:0:0: lost interrupt
        type: ata tc_bcount: 32768 tc_skip: 0
autoconfiguration error: viaide0:0:0: bus-master DMA error: missing interrupt, status=0x20
wd0: autoconfiguration error: transfer error, downgrading to PIO mode 4
wd0(viaide0:0:0): using PIO mode 4, WRITE DMA FUA EXT
wd0a: DMA error reading fsbn 81800256 of 81800256-81800319 (wd0 bn 88090176; cn 87391 tn 0 sn 48), xfer f70, retry 0
wd0: soft error (corrected) xfer f70

After that initial spew, it's stable, but fairly slow I/O-wise.
(...despite me having an mSATA module in an adapter as a disk.)

Regards,

- Havard