Hi!
I did some further research and came to the coonclusion that FAT isn't the
(major) cause for the corruption. I checked the whole SCSI-cabling, used
known good cables/terminators and couldn't get rid of thise "phase change"
messages. I get these messages with ext2, too, but appparantly this doesn't
corrupt the filesystem. Switching to normal IO or specf:n as found by
dejanews didn't help, too. Disabling tagged queueing, disconnect and/or sync
transfer had no effect (except making the box slow as ****).
I replaced the symbios with an Adaptec 2940 I took out of another box and
the problems went away.
Is the controller broken, or is there a way to teach the driver to deal with
this beast? Do I have to take the controler back to the store?
some snippets from syslog:
kernel: sym53c8xx:
setup=disc:y,specf:0,ultra:2,tags:8,sync:12,burst:7,wide:y,diff:0,rev
prob:n,buschk:0x2
kernel: sym53c8xx:
setup=mpar:y,spar:y,fsn=n,verb:2,debug:0x0,led:n,settle:2,irqm:0x0,nv
ram:0x1,pcifix:0x7
kernel: sym53c8xx: at PCI bus 0, device 9, function 0
kernel: sym53c8xx: setting PCI_COMMAND_PARITY...(fix-up)
kernel: sym53c8xx: 53c810a detected
kernel: sym53c810a-0: rev=0x23, base=0xe7000000, io_port=0xe400, irq=9
kernel: sym53c810a-0: ID 7, Fast-10, Parity Checking
kernel: sym53c810a-0: initial SCNTL3/DMODE/DCNTL/CTEST3/4/5 = (hex) 03/8e/a0/01/00/00
kernel: sym53c810a-0: final SCNTL3/DMODE/DCNTL/CTEST3/4/5 = (hex) 03/c0/00/00/08/00
kernel: sym53c810a-0: resetting, command processing suspended for 2 seconds
kernel: sym53c810a-0: restart (scsi reset).
kernel: scsi0 : sym53c8xx - version 1.3c
kernel: scsi : 1 host.
kernel: sym53c810a-0: command processing resumed
kernel: Vendor: IBM Model: DDRS-34560 Rev: S97B
kernel: Type: Direct-Access ANSI SCSI revision: 02
kernel: Vendor: YAMAHA Model: CDR102 Rev: 1.01
kernel: Type: WORM ANSI SCSI revision: 02
kernel: sym53c810a-0-<1,0>: tagged command queue depth set to 8
kernel: Detected scsi disk sda at scsi0, channel 0, id 1, lun 0
kernel: sym53c810a-0-<1,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8)
kernel: SCSI device sda: hdwr sector= 512 bytes. Sectors= 8925000 [4357 MB] [4.4 GB]
kernel: sda: sda1 < sda5 >
[...]
kernel: sym53c810a-0-<1,0>: phase change 2-3 6@034ae845 resid=2.
kernel: sym53c810a-0-<1,0>: phase change 2-7 10@034ae645 resid=4.
I currently load the driver as module with
modprobe sym53c8xx sym53c8xx="verb:2 tags:8"
I used kernels 2.2.7, 2.2.6 and 2.0.35 (with the ncr53c8xx driver). MTRR is
enabled, PCI_QUIRKS is enabled. (shall I send the complete .config?)
The controller no longer shares interrups. BIOS of this Gigabyte 686LX1
Board is updated, too.
I use 3 * "dd if=/dev/zero of=file.$i" stated with 90 seconds delay to test
stability. Time till "phase change" varies between 5 minuts and 2 hours.
Thanx
Rainer
Rainer
--
KeyID=58341901 fingerprint=A5 57 04 B3 69 88 A1 FB 78 1D B5 64 E0 BF 72 EB
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
