Alan > The bytes of the two CBW packets are otherwise identical?
Yes the CBW packets are identicle except the value of Tag field. The difference in the value of Tag field do not bother me because it can vary. > > Is it possible that the large gap of 12ms confuses the device ? > > It should not be possible. I too thought the same way. I think,the presence of SOF should be good enough for the device to stay tunned regardless of the timegap between previous and next command. > The only reason I can think of for the difference in behavior > would be if the device didn't receive the data1 packet > correctly. But that seems very unlikely. I see that the device is receiving data1 packet. > Have you tried this with a different device? Maybe a > different brand of USB flash memory device will work better. Yes we have tried several brands of USB flash memory(ohci/ehci) and they work well with our system. We haven't seen any devices failing in a similar fashion. But only this particular USB stick do not work with our system. We want to make this rogue USB device works with our system as well. On the other hand, this rogue device works well with Linux and Windows PC -:( How do the HOST driver cope with such rogue devices ? Early on, you mentioned about blacklist table. We know that the device always fails to ack to mode sense(6) packet. Is it possible to use blacklist table to avoid sending mode sense(6) command ? Regards Vivek > -----Original Message----- > From: Alan Stern [mailto:[EMAIL PROTECTED] > Sent: Tuesday, September 12, 2006 1:21 PM > To: Vivek Dharmadhikari > Cc: [email protected]; Edward Qian > Subject: RE: [linux-usb-devel] 3SYSTEM's POCKET USB DISK does not work > > On Tue, 12 Sep 2006, Vivek Dharmadhikari wrote: > > > Alan > > > > The drive do responds to MODE SENSE when plugged into Linux PC. > > > > I used the USB analyzer and I see following commands on the wire. > > > > 1. Inquiry message - Device responds with 36 bytes of data > and returns > > success in CSW 2. Test Unit Ready - Device returns error(0x1) in CSW > > 3. Request Sense - Device responds with 18 bytes of data > and returns > > success in CSW > > 4. Test Unit Ready - Device Returns Success in CSW 5. Unknown > > Command(opcode 0x25) - Device responds with 8 bytes of data > and return > > success in CSW. > > 0x25 is READ CAPACITY. > > > The above sequence of commands and their contents are identicle for > > both our embedded system and Linuc PC. The next > command(given below) > > causes different reaction from the device on our system and PC. > > Okay, good. You've found the first place where there's a > difference between the behavior on the two systems. > > > 6 Mode Sense(6) - on our embedded system, the device do > not return > > ACK to data1 packet. However, On the PC, the device return > the ACK to > > data1 packet. > > > > 7. Our embedded system repeat the Mode Sense(6) command three times > > and do not see the ACK from the device. The USB host driver > then do reset. > > On the PC, the host proceeds with the Test Unit Ready command. > > > > There is gap of 12ms between the last EOP bit of ACK for CSW in > > command > > 5 above and the SYNC pattern for Mode Sense(6) command when > using our > > system. On the other hand,the same gap is only 1-2ms on PC. > > > > The gap of 12ms is the only difference between the trace for our > > system and PC. > > The bytes of the two CBW packets are otherwise identical? > > > Is it possible that the large gap of 12ms confuses the device ? > > It should not be possible. > > The only reason I can think of for the difference in behavior > would be if the device didn't receive the data1 packet > correctly. But that seems very unlikely. > > Have you tried this with a different device? Maybe a > different brand of USB flash memory device will work better. > > This sure seems to be a problem with the USB device and not > with your embedded Linux system. If a different USB > mass-storage device fails in the same way then I would change my mind. > > Still, you're in a better position than I am to track down > the reason for the failure. You have the USB analyzer traces > to compare, you know exactly where to look, and you have the > hardware to experiment with. > > Alan Stern > ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
