Hello All,
On our mips based hardware, we are using linux-2.4.30.
When we do some tests on storage devices,
command_abort() is called frequently. As we knew that
error recovery in Linux 2.6 is much better than in 2.4
tree, I am wondering if anybody out there has
backported USB stack from 2.6.x
I saw an entry in your log where the status was
0xcd. It was for the
second READ_10 command.
You are right. It happens from time to time especially
when read the partition table and mount the device.
However, md5sum a file will give the correct result
always.
Or maybe something fishy
len=0xd actual length=0x0 status=0xff7d
I don't know what those messages are supposed to
mean. Are those status
values the error codes from your HCD? What does
0xff7d = -131 =
-EUNRECOVERABLE mean?
Those messages were printed out in function
usb_stor_bulk_msg by myself. status
Alan,
Why is status set to -131 instead of -ECONNRESET?
ECONNRESET is defined as 131 in linux-2.4.30 mips
tree.
EUNRECOVERABLE isn't defined.
The log shows everything worked okay (if you ignore
the problem of the
illegal status codes) until the computer tried to do
a 127 KB write.
Hello all,
I am working on a mips based SOC which has usb 2.0
core integrated in. After taking care of memory
mapping and other issues, host controller and usb
device are able to be recognized. However, I got some
SCSI errors so that partition table of the device
cannot be read.
I have
After I set VT6212L USB2.0 controller into D3hot
mode,
connecting/disconnecting usb drive doesn't
generate
any PME# assertion.
No, it won't do any of that unless
CONFIG_USB_SUSPEND
is enabled. I've certainly observed the PME# bit be
set
in the PCI config space after forcing the
Hello all,
After I set VT6212L USB2.0 controller into D3hot mode,
connecting/disconnecting usb drive doesn't generate
any PME# assertion.
According to spec, the controller supports D0, D1, D2,
D3hot and D3cold PME(reg82=0xffc2). the only thing
need to do is setting PME_En bit in the