USB to SATA & PATA Adapter (ID 152d:2338)

[drave40]

Just in case anyone still has this problem ( I did, that's how i landed here)

disconnect from USB
find jmicron driver. on my system it's called "pata_jmicron"
load it, "modprobe pata_jmicron"
connect to USB gain

dahdah everything is visible. well, it is here

cheers

[RedPenguin]

Thank you drave40, I never thought I needed an extra module being that it "worked" but I noticed even the minute I added this module, the IDE/Busy light actually worked on SATA, so it definitely seems to have better support.

It was driving me crazy, I was trying to clone a 200GB to my new 2TB but I kept getting USB disconnects of echi_hcd on the 200GB but so far it appears to work now.

EDIT: LoL, I got a random reset message again but yet it must have "recovered" because data is still flowing.

EDIT #2: dmesg shows like 8 resets yet the drive is still "working", so I dunno.

[reikred]

Drave40, I used your recipe once with success, but thereafter I have not been able to bring up my IDE drive (/dev/sdc, as it happens) viamy jmicron usb-ide adapter again.

Are there any other details that need to be considered? I have tried a bunch of stuff, such as

modprobe -r pata_jmicron
modprobe pata_jmicron

One observation I have made is that when I unplug/replug the Jmicron USB, there are no new messages from dmesg. However, if I unplug, then plug in some other USB device (a flash card reader in my case), then unplug flashreader and replug the Jmicron usb device, then both the card reader and then the Jmicron will appear one after the other in dmesg and also successively in lsusb.

I have tried a bunch of permutations of unplug, modprobe -r pata_jmicron, plug/unplug cardreader, modprobe pata_jmicron and replug jmicron. None of the permutations I tried so far have made /dev/sdc appear in fdisk -l. It just worked that one time, an I do not understand what I did differently.

I did boot the machine once as well, but perhaps not in combination with ALL the other permutations that I tried.

Is the anything else I need to try? Does the usb_storage.ko module play a part? Are there any other ahci or ehci or ohci things I need to worry about?

I have specifically (un)jumpered the IDE hard drive so that it is MASTER, as I have seen problems before with usb-ide adapters when drive is jumpered as cable-select or slave.

I should perhaps also mention that there is never a problem when I attach a SATA drive with the same Jmicron adapter. That always works, no problems of any kind. It is just the IDE that worked only that one time.

I'm running fedora core 14, but I don't think that matters in this context.

[reikred]

WOW, I just discovered something mind-blowingly simple and stupid that was causing 95% of my problems with usb-to-ide adapters:

IDE CONNECTOR MISALIGNMENT

Read that again: IDE connector misalignment.

Here is the story. I have a number of these no-brand (but Jmicron-based) usb/ide/sata adapters, and they all have three fundamental mechanical flaws:

1. the IDE plug is about 2mm narrower than it should be

2. the IDE guide-tab is not tall enough to block misaligned or upside-down insertion

and, MOST IMPORTANTLY,

3. the IDE pin-20 "key" position on the adapter (female side) is open, but should be filled/blocked to act as a "key" that blocks offset or upside-down insertion.

Because of these three flaws, it is very easy to offset the connector by +-1 pin, and it happened to me all the time because I tended to align the connector with the outer edge of the drive-side connector sleeve.

Note that the usb side of the adapter will still work, and dmesg will happily report that drive sdc (say) is up an running. Even if the adapter is misaligned so that the IDE side of the adapter is not talking to the drive at all. The net result is that nothing works, no partitions are reported, etc, etc.

I wonder how many people are having a mechanical alignment problem, but think they are having a software problem?

Note also that on the mini-IDE side if the adapter, the ""key" hole is properly blocked/filled. Hence the mini-IDE always works, no misalignment is possible.

I also found that there never was a need for loading the module pata_jmicron.ko. It was a red herring. I just happened to plug a drive correctly aligned at the same time I tried to load the module, so it looked like the module had an effect.

Finally, for the other 5% trouble cases, the cause is that the IDE drive was not properly jumpered as MASTER. The adapters (at least the JMicron-based ones) do NOT accept drives jumpered as SLAVE or CABLE-SELECT. The interface will then repeatedly reset and so on.

Hope this helps.