I have two machines, one running RH 9 and the other running RH 8.0, with all updates applied. The RH9 machine has an Intel D845PEBT2 mobo. The RH8 machine is a Toshiba Satellite 1115-S103. Both experience the following problem. If I plug a Minolta/QMS 2300 DL printer into the USB connector, I get the following messages from the kernel: hub.c: new USB device 00:1d.2-2, assigned address 2 usb.c: USB device descriptor short read (expected 18, got 0) hub.c: new USB device 00:1d.2-2, assigned address 3 usb.c: USB device descriptor short read (expected 18, got 0) Let me know what additional info you need.
There are some things I'd need, yes. 1. Please get me a full dmesg, taken after the problem happened. Please do not drop it into the comments box, but attach using the link "Create a New Attachement". 2. Tell me if you had any kernel at all working with that printer, and if yes, which one, so I can compare. The newer, the better. 3. If you didn't try 2.4.20-18, please do. Although I am not holding my breath for your specific problem, there was some brokenness in -13.
Created attachment 92131 [details] dmesg output
I've never gotten the USB port on this printer to work with Linux. I did just try 2.4.20-18, the dmesg output is from that kernel.
Rick, I'm wondering short reads may be caused by races. Please try this: ftp://people.redhat.com/zaitcev/tmp/kernel-smp-2.4.22-1.2121.2.1.nptl.i686.rpm No guarantee, though. DO NOT remove old kernels or use rpm -U! Also, it might require --force to install over RHL9.
This new kernel exhibits the same "short read" problem as before.
Upgrading this printers firmware to version 2.55 (available on the Minolta UK web site, but not on the Minolta US web site) fixed this issue. This does not explain why Windows USB subsystem had no trouble with the previous firmware revision, and Linux did. But why argue with success? -Rick
I think we *finally* know why it was happening: missing usblp->rcomplete=0 in usblp_read().
[Modified in FC2 2.6.7-1.494.2.2]