Memory stick won't open under new USB2

I upgraded to the latest HardDisc4 version of USB2 recently, with good results until I tried to open a 128MB Teradata memory stick. It would mount (and dismount) but clicking on the drive icon produced an internal error at &FC2C18CC and killed the Filer, forcing a reboot. This memory stick worked ok under WinXP and MacOS 10.4, and used to work under the old RISC OS USB2 module. 3 successive retries all had the same result (not sure about the abort address as I only noted the last failure). This is the first problem I’ve had reading a USB device with the new version: has anyone had a similar experience? George

Hi just plugged in a 256MB memory stick to my Iyonix 5.16. Haven’t used it since reflashing. Am getting the same Abort on Data Transfer message described by George Help, Guys!

Sorry just to clarify, it is the memory stick I haven’t used since reflashing. Now I plug it in again it’s giving me grief! It didn’t cause problems under previous versions of RISC OS 5. Thanks.

Hi. My suggestion would be that we maybe need to look further at the USB changes that went into 5.16. One avenue for investigation would be to build standalone versions of the various modules involved and try forcing them to load in your boot sequence. When things start working again, we know where to look for the problem.

Sadly, I don’t really have time for this. Maybe someone out there with a build environment and copy of the CVS logs for the USB modules might oblige?

Can either (or both!) of you two try getting a bit more information about the crash? Specifically:

• The address of the first abort
• The output of *showregs afterwards
• The versions of the different modules you’re using – USBDriver, EHCIDriver, OHCIDriver, the SCSI modules, DOSFS, and RTSupport (if you have it)
• The output of *usbdevices with the stick connected (should let us know whether it’s using EHCIDriver or OHCIDriver)
• Also, what are the capacities of the other sticks you’ve tried? (It might be down to something like the small sticks using FAT12 but bigger, working sticks using FAT32)

I had a quick look at the problem last night and tracked down the source code that corresponds to the address George reported (somewhere around line 1471 of DOSFSops.c in DOSFS), which doesn’t give much of an indication of where the problem lies. If I have the time I’ll try playing around with my Iyonix tonight (I’m still running 5.12, and a very quick test on my BeagleBoard didn’t suggest anything was wrong)

Sadly, I don’t really have time for this. Maybe someone out there with a build environment and copy of the CVS logs for the USB modules might oblige?

I’ll have a go at this sometime; it shouldn’t be too hard to bundle them all together with a script to softload them. If we can’t reproduce the bug ourselves (and it’s not just down to incompatible module versions) then we’ll probably have to produce a set of softloadable modules with debug logging enabled.

Update: I recently installed FAT32FS 1.27, it seems to have solved the problem with the original memory stick which afterwards opened fine and allowed reading and copying of data without any further intervention. I would recommend trying this upgrade to Dr Whichuster.

Have installed FAT32FS 1.27 and problem solved. Thank you.