major problems with ataraid(4) on an Asus PSCH-SR/SATA motherboard with Intel 6300ESB

["Greg A. Woods; Planix, Inc." <woods%planix.ca@localhost>]

I recently acquired a little system with an Asus PSCH-SR/SATA  
motherboard in it for the primary purpose of building a backup server  
(and possibly also extend it eventually to use as a PVR of some  
sort).  However I'm not having very much luck accessing the SATA/RAID  
drives installed in it.

I'm running a NetBSD-4.0_STABLE GENERIC.MP kernel on it built from  
fresh sources a couple of days ago.

The following hardware relevant to this problem shows up in dmesg  
output:

piixide1 at pci0 dev 31 function 2
piixide1: Intel 6300ESB Serial ATA/RAID Controller (rev. 0x02)
piixide1: bus-master DMA support present
piixide1: primary channel configured to native-PCI mode
piixide1: using ioapic0 pin 18 (irq 10) for native-PCI interrupt
atabus2 at piixide1 channel 0
piixide1: secondary channel configured to native-PCI mode
atabus3 at piixide1 channel 1

wd1 and wd2 are configured as a RAID-1 array in the Adaptec HostRAID  
BIOS:

wd1 at atabus2 drive 0: <WDC WD2000JD-00HBB0>
wd1: drive supports 16-sector PIO transfers, LBA48 addressing
wd1: 186 GB, 387621 cyl, 16 head, 63 sec, 512 bytes/sect x 390721968  
sectors
wd1: 32-bit data port
wd1: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
wd1(piixide1:0:0): using PIO mode 4, Ultra-DMA mode 6 (Ultra/133)  
(using DMA)
wd1: mbr partition exceeds disk size
wd2 at atabus3 drive 0: <WDC WD2000JD-00HBB0>
wd2: drive supports 16-sector PIO transfers, LBA48 addressing
wd2: 186 GB, 387621 cyl, 16 head, 63 sec, 512 bytes/sect x 390721968  
sectors
wd2: 32-bit data port
wd2: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
wd2(piixide1:1:0): using PIO mode 4, Ultra-DMA mode 6 (Ultra/133)  
(using DMA)
wd2: mbr partition exceeds disk size
ataraid0: found 1 RAID volume
ld0 at ataraid0 vendtype 1 unit 0: Adaptec ATA RAID-1 array
ld0: 186 GB, 24321 cyl, 255 head, 63 sec, 512 bytes/sect x 390721536  
sectors
ld0: mbr partition exceeds disk size
ld0: mbr partition exceeds disk size
ld0: mbr partition exceeds disk size
ld0: mbr partition exceeds disk size


However whenever I try to write any significant amount of data to the  
ld0 device, eg. in order to newfs a filesystem on it, the process  
doing the writing eventually just hangs in "physio2".  Here's a newfs  
process that I started last night and which is still hung in exactly  
the same spot as it did a few moments after doing some initial I/O  
activity:


11:56 [4] $ fstat -p 553
USER     CMD          PID   FD MOUNT       INUM MODE         SZ|DV R/W
root     newfs        553   wd /          46336 drwxr-xr-x     512 r
root     newfs        553    0 /         903559 crw-------   ttyE1 rw
root     newfs        553    1 /         903559 crw-------   ttyE1 rw
root     newfs        553    2 /         903559 crw-------   ttyE1 rw
root     newfs        553    3 /         904334 crw-r-----   rld0a r
root     newfs        553    4 /         904334 crw-r-----   rld0a w
11:56 [5] $ ps -up 553
USER PID %CPU %MEM VSZ RSS TTY   STAT STARTED    TIME COMMAND
root 553  0.0  0.0 492 616 ttyE1 DL+   7:08PM 0:02.99 newfs /dev/rld0a
11:56 [6] $ ps -lp 553
UID PID PPID   CPU PRI NI VSZ RSS WCHAN   STAT TTY      TIME COMMAND
  0 553  760 19785  -5  0 492 616 physio2 DL+  ttyE1 0:02.99 newfs / 
dev/rld0a
11:56 [7] $

Note that processes hung in physio2 are not interruptible and so if  
this were a newfs during an install attempt then the only recourse is  
to hit the reset or power switches -- very annoying.  Note the  
implication here too that a uniprocessor INSTALL kernel suffers the  
same problem.

Has anyone else had any success using ataraid(4) on an Intel 6300ESB  
based implementation with NetBSD-4.0_STABLE?

Should I turn this info into a PR?

Is there any further debugging I could try?  Would a crash dump be  
helpful to anyone who might be able to fix this problem?

Note that when wd1 and wd2 are configured as JBOD (instead of the  
desired RAID-1 array) then they are fully usable by NetBSD -- I can  
install a fully functional system without any problem.  I could then  
use RAIDframe to achieve the mirroring I desire, but I think I would  
prefer to use the native RAID-1 support in order to simplify  
installation and maintenance, at least until sysinst(8) can directly  
support RAIDframe configuration and installation.


As an aside:  note this motherboard also includes a Marvell 88SX5041  
SATA controller that also implements the Adaptec HostRAID  
functionality.  It shows up as:

Marvell 88SX5041 SATA (SCSI mass storage, revision 0x03) at pci2 dev  
8 function 0 not configured

Does anyone have any clues on how hard it would be to add ataraid(4)  
support for this chip too?

Can I at least directly access any drives attached to the 88SX5041 as  
JBOD devices with some driver that could be cobbled into NetBSD  
4.0_STABLE?  It would be nice to be able to use all of the integrated  
SATA hardware to support additional drives rather than also having to  
add another SATA controller to this machine, especially since it's in  
a relatively space-restricted 2U chassis.

--
                                        Greg A. Woods; Planix, Inc.
                                        <woods%planix.ca@localhost>