Hello Alberto, I had similar problems with USB2 hanging during file transfers. We have seismic data here with very large files (>10GB) which take a long time to transfer to disk. I found that when I accessed two or more USB2 devices on the same controller the system would hang. I think this error is caused by bus contention between the two devices. My workaround for this problem was to install two USB2 controller cards (PCI in my case) and plug one USB2 drive into each USB2 controller. When the drives are on separate controllers I can copy data to and from both drives simultaneously with no problems. This solution is somewhat wasteful but it does work... hopefully the problem will be resolved in some future kernel. I am currently running 2.4.20 on that system, 2.4.18 does not work very well and the -ac series did not work at all for me.
I hope this helps, A -----Original Message----- Message: 6 Date: Tue, 11 Feb 2003 09:19:20 -0500 From: ALBERTO D SCOTTI <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: [Linux-usb-users] System hangs up with USB 2.0 and mass storage I have been trying to add an external hard drive using USB 2.0 The box is a dual Athlon, with a Tyan mb (thunder). It runs RH 8.0 (2.4.18-14smp), but I had the same problem after upgrading to 2.4.20. I have tried all possible permutation of 2 controllers (vt6202 and NEC) and 2 kind of HD (Maxtor and generic based on a ACER IDE to USB chipset). The problem is the following: ehci-hcd recognizes the controller without a glitch. When I plug in the HD, usb-storage correctly identifies it and attach it to the appropriate scsi emulation device. I was able to create the partitions and format the disk (either ext3 of reiserfs). So far so good. Trouble starts when I start moving files to and from the disk. It works for some time (a few minutes, equivalent to 1 or 2 Gb). Then, first the disk stop working, say during a transfer. After approximatively 30 seconds, the whole system freezes and I need to to do a hard reboot. With the 2.4.20 modules compiled with the debug option on, the last two lines that I read out of dmesg are usb-controller: Bulk data transfer result 0x0 usb-controller: attempting to get CSW Things that I have tried and did not work 1) Boot in single CPU mode 2) Boot with the noapic command appended 3) Set the smp_affinity so that only one cpu services the interrupt requests of ehci-hcd 4) Set the MP version to 1.1 in the bios. This is a production system, so I am reluctant in doing any extensive testing without some guidance. Any help is appreciated. -- Alberto Scotti Asst. Prof. Dept. of Marine Sciences CB 3300 University of North Carolina Chapel Hill, NC 27599-3300 919-962-9454 (w) 919-962-1254 (f) ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
