At 9:58 AM -0700 7/11/02, David Brownell wrote:
>Tom Collins wrote:
>>I'm working with some USB drives and trying to troubleshoot a few
>>problems. MIPS hardware, kernel 2.4.19-rc1 (for the EHCI driver).
>
>Out of curiousity -- big-endian or little-endian MIPS?
>(Or should I say, "MIPS or SPIM"? :) Folk have used the
>EHCI driver on PPC (big-endian) so I don't think there
>are any issues in that area, but surprises can happen.
little-endian.
>
>>I have a USB 2.0 drive that is powered off the USB bus, connected
>>to an Adaptec USB 2.0 card. ...
>>
>>Question 1 is, if the device is seen in /proc/bus/usb/devices, why
>>didn't the SCSI driver map it to a /dev/sdx device?
>
>Seems like some kind of usb-storage issue; what does the
>/proc/scsi/scsi file tell you?
'cat /proc/scsi/scsi' results in a segmentation fault. I'm a newbie
kernel hacker, so I don't know how to match the Oops with the
System.map and get useful information. This system doesn't have a
/proc/ksyms file and I'm not sure how to turn it on. ksymoops didn't
give useful information without ksyms.
I can give you the following mappings from the System.map though:
8024afd0 - inside strnlen
8024be60 - inside vsnprintf
Unable to handle kernel paging request at virtual address 0a0a0c10,
epc == 8024afd0, ra == 8024be60
Oops in fault.c:do_page_fault, line 204:
$0 : 00000000 80000000 83db7e40 0a0a0c10 0a0a0c10 ffffffff fffffffe 83db7e40
$8 : ffffffff 0000000a 83db7d48 00000000 83da6039 fffffffe 00000008 ffffffff
$16: 0a0a0c10 83da6079 83db7e44 ffffffff 00000000 ffffffff 83da606f 7c259f91
$24: 0000000a 00000000 83db6000 83db7dc0 80260044 8024be60
Hi : fffd1c38
Lo : 0000f698
epc : 8024afd0 Not tainted
Status: b000fc03
Cause : 00800008
Process cat (pid: 118, stackpage=83db6000)
Stack: 2ab82028 00000001 803f82a0 82ae97c0 0000000a 00000002 ffffffff 00000003
0000005c 83da6013 00000004 803e7aea 803e7afa 00000013 803e7a00 83da6000
8024c174 00000004 00000000 8014640c 80260053 82ae9740 8024c19c 83db7ee0
00000400 83da6000 814c3ee0 10000268 801c41cc 00000000 83da6000 80260048
0a0a0c10 83db7e40 00000000 00000000 803e7a00 00000013 814c3ee0 00000400
00000000 ...
Call Trace: [<8024c174>] [<8014640c>] [<80260053>] [<8024c19c>]
[<801c41cc>] [<80260048>]
[<801c09b8>] [<8025f8cc>] [<801479d8>] [<8015bd70>] [<8013a914>] [<80143988>]
[<80108808>] [<80113698>] [<80117f40>]
Code: 00801821 10a0000a 24a6ffff <80820000> 10400007 2405ffff
24c6ffff 10c50004 24630001
>As for why all those ports change status ... hub driver diagnostics
>are really confusing, they're talking about two different hubs:
>first the EHCI root hub, which decides the device isn't going to
>work at high speed so it hands it off, then the OHCI one (to which
>the device was handed by the EHCI driver).
It's starting to make a little sense. I now know which port the
drive is on, and just ignore the other port messages (especially
since they list change 0 and portstatus 100 -- seems to indicate to
me that not much is going on there).
>But I'm not sure why you're seeing more than one, and it looks
>like maybe the recovery in usb-storage isn't working for this
>particular device.
Actually, I've just found out that it is a power problem. The drive
needs 800mA, but the card only provides 250mA per port.
I've been having another problem with usb storage. If I use the
external power on the drive and have the system recognize it, the
system will usually hang when copying data to the drive. I have the
same problem with a separate USB 1.1 drive.
During format, it pauses while writing the inode table. It goes
smoothly up to about 23, and then pauses for a few seconds every 4 or
5 inodes written. When I copy files (cp -av /usr /mnt/storix/) it
goes smoothly for awhile, and then stops on a file.
It still responds to pings, but my telnet session and serial console
are frozen and the copy never comes back.
--
Tom Collins
[EMAIL PROTECTED]
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
PC Mods, Computing goodies, cases & more
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
