I just got myself a 2 GB SD card (label says Kingston Elite Pro 50x), which causes some problems. At first i also had hardware problems, seems both my card readers were too old for a 2 GB card, as they failed also on windows and different computers. Now i bought a new card reader, which works out of the box on a fresh windows install (copied over 1.5 gigs from/to the card to make sure to pass the 1 gig limit), so i think it's not a hardware issue with this one. Kernel: Tried with gentoo-sources linux-2.6.17-gentoo-r8 first, then linux-2.6.18-rc7-git4, seems both behave the same way. Output below is from linux-2.6.18-rc7-git4. The card has been formatted in the camera (canon ixus 800, shows a capacity of 1.8 gigs). Placing the card in the reader and connecting it results in the following (i'll attach the whole debug output, but that parts seem important to me): Oct 8 20:15:12 [kernel] SCSI device sdd: 1955328 512-byte hdwr sectors (1001 MB) Oct 8 20:15:12 [kernel] sdd: p1 exceeds device capacity Oct 8 20:15:12 [kernel] Buffer I/O error on device sdd1, logical block 3910272 (some more errors like that one) # fdisk -l /dev/sdd Disk /dev/sdd: 1001 MB, 1001127936 bytes 58 heads, 57 sectors/track, 591 cylinders Units = cylinders of 3306 * 512 = 1692672 bytes Device Boot Start End Blocks Id System /dev/sdd1 1 1183 1955263+ 6 FAT16 Partition 1 has different physical/logical beginnings (non-Linux?): phys=(0, 2, 4) logical=(0, 2, 16) Partition 1 has different physical/logical endings: phys=(969, 57, 57) logical=(1182, 51, 57) # lsusb Bus 001 Device 005: ID 0aec:3260 Neodio Technologies Corp. 7-in-1 Card Reader If i mount the card it shows up as 2 GB, but once i try to access files located above the first GB i get i/o errors. Anything else i should try/add?
Created attachment 98427 [details] debug output from usb-storage
Created attachment 98428 [details] Kernel configuration
Created attachment 99718 [details] Log from plugging in the reader with 2.6.19-rc1 Gave it a shot with 2.6.19-rc1, put the card in the reader and then plugged it in. Kernel config is same as before (apart from running make oldconfig for .19 and leaving everything new to default values).
You need to enable USB storage debugging otherwise the logs aren't interesting
Created attachment 99733 [details] Oops, that one is better Even worse, when using metalog you need to take the right file (kernel, not everything). ;-)
Nothing obvious, next step is to look at the data flying around. Please look at /usr/src/linux/Documentation/usb/usbmon.txt and capture usbmon logs *and* new debug logs from the same session in the same way as before. It's important that you provide the two logs from the same usage attempt.
Created attachment 100531 [details] Kernel messages with 2.6.19-rc3
Created attachment 100532 [details] Output from usbmon with kernel 2.6.19-rc3 Here are the logs for both things from the same plugin event of the card reader - hope i got everything right with usbmon, it's output doesn't make much sense to me. ;-)
Created attachment 103717 [details] partially annotated usbmon log
Sorry for the delay in looking at this. I've gone over the transfer and the device is reporting 1GB capacity. I have a couple of ideas where the device may be reporting things incorrectly, but the next step would be to see what windows does. Please download and run this: http://benoit.papillault.free.fr/stats/get.php?location=../usbsnoop/sniff-bin-1.8.zip Install the filter on the card reader device then plug it in with card already inserted. Wait long enough for windows to detect the media but not too long for the longs to get overly huge. THen uninstall the filter, and unplug it. Compress the log and attach it here.
Created attachment 104044 [details] usb logfile (In reply to comment #10) > Install the filter on the card reader device then plug it in with card already > inserted. Wait long enough for windows to detect the media but not too long for > the longs to get overly huge. THen uninstall the filter, and unplug it. > Compress the log and attach it here. I hope it worked, the log really seems to log quite fast. I hope the one i made is OK from both size vs. information.
Hmm... I have run over that log and I don't see any reason for windows to believe that's *not* a 1GB card. Windows is doing some extra commands. It does READ FORMAT CAPACITIES which is primarily for optical devices, but this agrees with the READ CAPACITY (0x1dd600 512 byte blocks). It also does some mode sensing without any interesting results. I guess the next step is to file a bug at http://bugzilla.kernel.org Attach the usbmon log and link to the usbsnoop one. I'll comment on my findings once you have opened the bug there.
Please test 2.6.21-rc6, file upstream, and post new URL here.
