Hi Oyvind,
This patch is not ready to commit yet.
Unfortunately I had several emails open at a time and I inadvertently pushed
send on the wrong message before I had it polished.
Here is the short story of this patch left out of the first email:
I have a SAM-ICE device running V8 firmware that I am using on an Atmel
AT91SAM9G20-EK evaluation board. When I execute openocd, approximately
every third attempt causes a complete communication failure, the green light
on the SAM-ICE extinguishes for about three seconds (internal watchdog
timeout in the dongle?), flickers a few times, and then resumes a solid
illumination. The console displays the following messages during the event
and the debugger is left unusable afterwards:
Jupiter:Embedded garycarlson$ openocd-jlink -f
/usr/local/openocd-config/interface/jlink.cfg -f
/usr/local/openocd-config/target/at91sam9260.cfg
.
.
.
Error: J-Link command EMU_CMD_VERSION failed (-60)
Error: J-Link command EMU_CMD_VERSION failed (-1)
Error: J-Link command EMU_CMD_VERSION failed (-32)
Info : J-Link initial read failed, don't worry
Info : J-Link JTAG Interface ready
Error: usb_bulk_read failed (requested=1, result=-32)
Error: jlink_tap_execute, wrong result -107 (expected 1)
Error: usb_bulk_read failed (requested=1, result=-32)
.
.
.
Based on the j-link code written by the original author, I suspect he may
have experienced the same or a similar problem. The code has a while loop
and three retries in the jlink_init function to get communication to work
with the dongle. The first version patch I submitted yesterday has
decreased the statistical startup failure rate in my case from 43% to 2%,
but has not completely eliminated the problem.
Unfortunately I only have one j-link device with one particular board. What
I am hoping to find is some other people who use the j-link device with
other CPUs that have also experienced the problem and are willing to test
this patch and a possible revised patch prior to making a final commitment
decision. What I don't want to do is break it for someone else through
unintended consequences.
The second version of this patch that I am working on will hopefully reduce
the failure rate from 2% to zero. I will try to get that out either today
or tomorrow if my own testing proves fruitful.
Gary Carlson
Gary Carlson, MSEE
Principal Engineer
Carlson-Minot Inc.
On 7/4/09 12:29 AM, "Øyvind Harboe" <[email protected]> wrote:
> Is this ready to commit?
>
> Could you write a few words about what it fixes for the
> commit message?
>
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development
