On Tue, 1 Mar 2005, Gerd v. Egidy wrote: > Hi, > > > There's a description of max_sectors in the FAQ at www.linux-usb.org. > > Can you verify that it is still set to 240? Maybe something we're not > > aware of has changed it. > > [EMAIL PROTECTED] cat /sys/block/sda/device/max_sectors > 256 > > direct after plugging in. don't know where that comes from. maybe some > hotplugging scripts. I can take a look where that is from if that matters.
I'm a little curious. It might be important for users with other sorts of drives. There are some which don't like max_sectors to be as large as 256. > With dd (on the pure disk, just as in your example) the first reset happens a > lot faster than with my method which was running rdiff-backup to a disk with > reiserfs ontop of a dm-crypto. Maybe the dd generates more stress on the usb. Maybe... > I tried reducing max_sectors to 128 and bs=32k: I think the bs value doesn't make much difference. The block layer will do its own read-ahead and will generate requests with sizes up to max_sectors, no matter how much dd asks for -- just so long as there is a rapid set of sequential transfers. > usb-storage: usb_stor_bulk_transfer_sglist: xfer 65536 bytes, 8 entries > usb-storage: Status code -121; transferred 64525/65536 > > max_sectors 64, bs=16k: > > usb_stor_bulk_transfer_sglist: xfer 32768 bytes, 8 entries > usb-storage: Status code -121; transferred 31757/32768 > > but even if I set max_sectors to 8 (bs=2k or lower, doesn't matter) the > lowest > amount of bytes transmitted seems to be always 4k. The block layer likes to transfer full memory pages (on x86 a page is 4 KB). > The errors look like this: > > usb_stor_bulk_transfer_sglist: xfer 4096 bytes, 1 entries > usb-storage: Status code -121; transferred 3085/4096 > > So in case of an error there are always 1011 bytes missing in the transfer. In other words, two 512-byte blocks are missing and the 13-byte CSW is sent instead. Clearly this isn't directly related to the transfer length. And I don't see how it could possibly be caused by a problem with the EHCI driver or hardware. That leaves nothing but the Cypress controller. I haven't gotten anything back yet from Cypress on the technical support request. In your most recent tests, have you run across any more examples where the device got disconnected? Or did the hub driver patch fix that particular problem? Alan Stern ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
