Opened 10 years ago

Last modified 10 years ago

#20 accepted defect

VIA VT6421A read issues

Reported by: Lewisr Owned by: Steven Levine
Priority: major Milestone: Release-1.8.16
Component: danis506.add Version: 1.8.15
Keywords: esata, via, vt6421, read Cc:

Description

Installed an eSATA card in my ThinkPad? dock, based on the VIA VT6421A chipset (same PCI ID as VT6421).

ChipList?.txt indicates that this chipset is supported by Dani, though only in JBOD mode (no issues; I only needed the eSATA connection, and only at 1.5Gbps).

DaniS506 & DaniATAPI load with the following statements:

BASEDEV=DANIS506.ADD /!BIOS /PCS /VL /A:0 /U:0 /APM:254 /A:1 /BAY /A:5 /BAY
BASEDEV=DANIATAP.FLT /EJ /A:1 /U:0 /TYPE:ZWL /RSJ /A:5 /U:0 /TYPE:ZL

(Note that the VT6421 card has three connections: (1) internal SATA, (1) eSATA, and (1) PATA, taking the positions of adapters 2-4. Adapter 1 is the secondary internal interface in the T43, and adapter 5 (now) is the CMD unit in the dock, to which the UltraBay? is attached.

The drive connected via eSATA is a Seagate 1TB in a Vantec NexStar? CX eSATA/USB 2.0 case. The drive works fine when connected via USB, and I am not hot-plugging the device into the eSATA port (even though ChipList?.txt indicates that hot-plug is supported).

Writing to the drive is fine. When reading, the drive stops responding. Such a condition requires a power cycle of the external case before the drive responds again, even after rebooting (or power cycling) the VIA card and booting to a different OS (Parted Magic Linux).

CHKDSK starts, gets through phase 0, and starts phase 1, at which point the activity light on the drive goes solid, CHKDSK continues for a time, then fails with our old friend:

JFS0148: CHKDSK Unrecoverable error reading M from m:. CHKDSK CANNOT CONTINUE.

(Separate ticket to be filed for the truncated JFS0148 message which has frustrated me before.)

Parted Magic (recent version, using jfsutils from 2011) runs chkdsk against the drive, and can read and write without issues, as long as the drive has not been left in its "zombie" state.

I swapped the Seagate drive for a 120GB Western Digital. Using DFSee, I repartitioned the WD for a 10GB HPFS partition, rebooted, and formatted. I then filled the drive with data. Writing succeeded. Reading (copy /s m:* c:\var\temp\test) failed after a few files. This is apparently not a filesystem problem.

Finally, in my Startup folder is a call to refresh all removable media. With this device connected, the refresh never finishes, and lvm /rediscoverPRM continues to beep. Shutting down yields a warning from Xshutdown that it is still running.

Report files attached. If helpful, I can provide hardware to duplicate the issue. :-)

Attachments (1)

rosenthal-VIA-VT6421A-read-failures.zip (15.3 KB) - added by Lewisr 10 years ago.
All requested repoirts from PROBLEM.TXT, plus PCI and a couple bldlevel ones

Download all attachments as: .zip

Change History (2)

Changed 10 years ago by Lewisr

All requested repoirts from PROBLEM.TXT, plus PCI and a couple bldlevel ones

comment:1 Changed 10 years ago by Steven Levine

Milestone: Release-1.8.16
Owner: set to Steven Levine
Status: newaccepted
Note: See TracTickets for help on using tickets.