I've been running Linux since about 1996. My principal machine I/O systems have
always been SCSI. I'm abandoning SCSI now -- not because I want to -- but because I'm increasingly getting priced out of the peripheral market. It seems like little (particularly U320/U160) is backwards compatible with my dead & dying Adaptec controllers, and the "known working" combinations of drives, cables, and controllers ain't cheap enough any more to keep my family fed.
Hardware:
Soyo KT400 Dragon (Black)
AMD Athlon XP 2800+, 333 FSB
Memory: 2 GB PC2700
I/O: Adaptec 2940UW, 6 IBM 9.1 GB Ultra SCSI DGHS (going out the door if I can ever get it done), and a bunch of SCSI devices I guess I'm gonna have to live without now (like a DDS3 DAT drive
, and a nice HP CD-RW)
Video: ATI Radeon 9500/256 MB
Distro/System:
Red Hat Linux 9.0
2.4.20-30.9 #1 Wed Feb 4 20:45:39 EST 2004 i686 athlon i386 GNU/Linux
Yeah, I know. My kernel/distro are old. I'll upgrade if I see I need to, and I'm wondering if this doesn't have something to do with my saga. Instinct tells me not, though.
Ran down to Circuit City and picked up an Western Digital WD800JB (see review in the HCL) for $80 to see if I could hook it up on the on-mobo IDE controller. 80 GB is more than I need for my purposes (I could get away with 45 GB if I had to). Cabled the drive up, partitioned it (into 7 or 8 partitions), ran
mke2fs -j on all of them (except for the intended swap partition). Everything looked good to that point.
Started copying off the SCSI mount points onto the IDE. This looked good, too. I hadn't yet tweaked DMA settings or anything else in the BIOS (other than turning on IDE channel 0 and having the BIOS detect the drive parameters).
I got all the way through the copying of about 20 GB of data without any complaint from the IDE subsystem, although I was getting the now usual SCSI subsystem complaints about unexpected busfrees, etc. These SCSI error messages have been growing in number, severity, and frequency; which is why I'm in a panic to get off of them NOW.
Next step in the verification: swap out a SCSI mount point for an IDE mount point and see how it looks ... try to run with it. OK,
umount /home;mount /dev/hda5 /home. Now I start getting serious messages from the ide subsystem ... to the order of things like this:
Dec 2 08:25:08 rutabaga kernel: hda: dma_timer_expiry: dma status == 0x21
Dec 2 08:25:18 rutabaga kernel: hda: error waiting for DMA
Dec 2 08:25:18 rutabaga kernel: hda: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }
Dec 2 08:25:18 rutabaga kernel:
Dec 2 08:25:38 rutabaga kernel: hda: dma_timer_expiry: dma status == 0x21
Dec 2 08:25:48 rutabaga kernel: hda: error waiting for DMA
Dec 2 08:25:48 rutabaga kernel: hda: dma timeout retry: status=0x58 { DriveReady SeekComplete DataRequest }
At that point, I can use
hdparm and
smartctl to review the drive status: DMA's been shut off all the way, I'm down to PIO mode 4, and I'm
still taking I/O ("uncorrectable error", followed by a sector number) errors.
OK, after calming down, I figured out it's
gotta be all about bad blocks. Here's where I find out that IDE has no SCSI-like low-level format utility
, no hardware sector remapper (at least that I can find).
I guess I'm stuck to having the Linux filesystem drivers do what they can do with bad block remapping. The first -cc run with
mke2fs took almost three full hours (!!!). Oh boy. So I get all the way through the formatting run with -cc on ...
dumpe2fs -b ... guess what?
NO BAD BLOCKS reported. Ay yi yi. I'm gonna toy around with this disk for an hour or two more, but it's headed back to Circuit City for an exchange, methinks.
Here's the
lspci output that the thread top asked for:
00:00.0 Host bridge: VIA Technologies, Inc. VT8377 [KT400 AGP] Host Bridge
00:01.0 PCI bridge: VIA Technologies, Inc. VT8235 PCI Bridge
00:08.0 Multimedia video controller: Brooktree Corporation Bt878 Video Capture (rev 11)
00:08.1 Multimedia controller: Brooktree Corporation Bt878 Audio Capture (rev 11)
00:09.0 Multimedia audio controller: Ensoniq ES1370 [AudioPCI]
00:0a.0 SCSI storage controller: Adaptec AHA-2940U/UW/D / AIC-7881U
00:0c.0 Ethernet controller: National Semiconductor Corporation DP83815 (MacPhyter) Ethernet Controller
00:10.0 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.1 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.2 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc. VT82C586/B/686A/B PIPC Bus Master IDE (rev 06)
01:00.0 VGA compatible controller: ATI Technologies Inc: Unknown device 4153
01:00.1 Display controller: ATI Technologies Inc: Unknown device 4173
Have tried turning APIC off at the BIOS level and as a kernel parm ... no difference. Have used
hdparm to switch DMA, UDMA, and PIO modes ... some settings are better than others, but nothing is a savior. I think I've cost Google megabucks over the last couple of days in machine cycles ... in desperation, I'm coming here (a newbie to this forum, but you guys have a good rep out there). The answer's probably so simple that I can't see it.
Anybody got any ideas for me? Is this what I have to look forward to living with IDE?
Thanks, folks ...
--Jim--