On Thu, 19 Jun 2003, Vedran Rodic wrote: > Hi > > Unfortunaltely, I had a couple of problems with ehci on usb-storage since > my last report. I've applied ehci-hcd micro-patch from David. I use > 2.4.21-rc8-ac1 (basicaly equivalent to 2.4.21-ac1 + ehci-hcd micro patch USB > wise). > > Also there was a patch from David with the subject line: > "[patch 2.5.70] ehci, fix qh re-activation problem" > Does this problem exist in some form in 2.4? > > The problems: > > 1. > usb_control/bulk_msg: timeout > > You probably know what happens when you get this on usb-storage, but anyway: > > I was running cdrecord/mkisofs to record some files from vfat disk on USB > 2.0 to ide cd-r device. I was playing some mp3s from that disk too. > cdrecord stopped, xmms stopped, and kernel panicked in vfat kernel module > with a lot of I/O errors. > > I had to umount the disk, rmmod/insmod usb-storage to make it work again.
In itself that timeout isn't necessarily bad, but in practice it usually seems to indicate a problem with EHCI. More extensive debugging information would help here. If it's not a host controller problem then we should be able to recover from the error. But without knowing what the cause is and what's happening, it's hard to say anything definite. > 2. > One time, it happened immediately after powering the disk on, without > me even mounting it: > > hub.c: new USB device 00:09.2-1, assigned address 2 > usb.c: USB device 2 (vend/prod 0x5e3/0x702) is not claimed by any active driver. > Initializing USB Mass Storage driver... > usb.c: registered new driver usb-storage > scsi1 : SCSI emulation for USB Mass Storage devices > usb_control/bulk_msg: timeout > last message repeated 2 times > scsi: device set offline - command error recover failed: host 1 channel 0 id 0 lun 0 > > After I reconnected the disk and rmmod/insmod usb-storage, it worked fine. Again, it might be a problem in the host controller driver. > 3. > I've noticed that usb-storage doesn't remove the disk from /proc/scsi/scsi > (the representation of kernel scsi device list) > when the device is disconnected. Because of that, I cannot connect a > different hard disk to usb, without rmmod/insmod usb-storage first, and that can > be impossible when I have another device on usb-storage that I don't want to > stop using. Can this be fixed for 2.4? That was by design. The behavior has been changed for 2.5, but it's not going to be changed for 2.4 since the SCSI layer for 2.4 can't handle hot-unplugging. However, depending on what sort of disk you're using this shouldn't affect you. Go ahead and leave the old device entry in /proc/scsi/scsi; when you plug in a different disk it should get its own new entry. > 4. > > About possible performance improvement on 2.5 ehci/usb-storage: > I've tested with 2.5.71, and I don't see a significant improvement over 2.4. > (Both are arround 7.7 MB/s for read, 10 MB/s for write). Disk can do arround > 40 MB/s reads on IDE. I don't know what's going on there. Alan Stern ------------------------------------------------------- This SF.Net email is sponsored by: INetU Attention Web Developers & Consultants: Become An INetU Hosting Partner. Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission! INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
