JAWS Internal Page

15.12.2006, 19:00 Mustafa Yuecel

• reflash the whole network with jaws from CVS HEAD. The unstable behavior while code distributing is no more observed. Note: node 01:86 has no bootloader...

29.11.2006, 16:00 Mustafa Yuecel

• while the code is distributed, the network is very unstable. I set CDIST_PKT_SLEEP to 70 ms (I calculated (5 KB/s, 192 Bytes per ACL packet) 40 ms per packet).

27.11.2006, 17:00 Mustafa Yuecel

• node 01:fe reflashed with bootloader and jaws from CVS HEAD

24.11.2006 Mustafa Yuecel

• revert back to 2-way ping (network is with 3-way ping more unstable)

23.11.2006, 13:00 Mustafa Yuecel

• try to set up a net from CVS HEAD with enabled 3-way ping feature, set watchdog timer to 5 min (3 min is too short to upload a large image). Because the EEPROM location is changed for the mode (gui or dsn), I have to set up all dsn nodes apartly with jaws set mode dsn command...

17.11.2006, 14:00 Mustafa Yuecel

• node 01:c1 reflashed with bootloader and jaws from CVS HEAD

21.10.2006, 16:45 Kevin Martin

• yesterday evening, 20:30pm, I flashed the DSN with the software 'jaws.20061020-1919.btnode3.hex'. The DSN seemed to be quite stable during night. Unfortunately, I set the watchdog timer to 20min, that's why there are some longer periods where not all of the nodes have been reachable (see cacti, btw: usually it was the 4 nodes around ETZ G60.1/G96 that have not been reachable). Nevertheless, the watchdog timer seems to do it's job.
• This morning, 10:52, I flashed the DSN with the software 'jaws.20061020-1052.btnode3.hex', which sets the watchdog to 3min. During code distribution, the GUI node hung up and was not responding until the watchdog timer got active. That's why there was a total downtime of about half an hour (see cacti). After the GUI node was running again, the remaining nodes of the network were reachable immediately again.

11.10.2006, 16:45 Mustafa Yuecel

• node 01:86, 01:01 and 00:dd reflashed with new bootloader (CVS HEAD with version tag)
• NOTE: 01:01 and 00:dd are moved from G82 to G81 cause of missing USB power source

06.10.2006, 16:45 Mustafa Yuecel

• node 00:d4 reflashed with new bootloader (CVS HEAD) and flash image jaws.tg_btnode.20060924-2106.btnode3.hex

05.10.2006, 17:15 Mustafa Yuecel

• node 00:db (with a80 target) had no valid flash image (even the bootloader was non-working) --> reflash

04.10.2006, 11:00 Jan Beutel

• nodes that have problems are 00:cb (baumann), 00:d7 (moser) and 01:01 and 00:dd (plessl) as well as the two at the printer

03.10.2006, 22:00 Jan Beutel

• fixed Node 01:20
• network worked with 20060914, but i turned off all target nodes and flashed them with senso to be sure nothing happens
• will now try to distribute 20061003-2141

03.10.2006, 16:00 Jan Beutel

• Node 01:20 down (broken)
• network seems very shaky (connectivity)
• will try to distribute code 20060914

25.09.2006, 11:00 Jan Beutel

• Uploaded 20060924-2106
• Distributed 40 Nodes on ETZ G
• Set all locations on nodes

09.06.2006, 16:00 Matthias Dyer

• Uploaded 20060609-1558 (current CVS + class 5 logs)

30.05.2006, 22:00h Jan Beutel

• Uploaded 20060530-2147 with bt uart 115200 and fixed rts/cts mapping to a few nodes for testing.

23.05.2006, 20:00 Matthias Dyer

Uploading jaws.20060523-1916.btnode3.hex (= actual CVS version) resolved conflicting connection request deadlock. Nodes sometimes still reset themselfs, but reapear in network.

18.05.2006, 17:00h Jan Beutel

• Changed GUI node to 00:DB because 00:BE has errors.

15.05.2006, 10:50 Matthias Dyer

Uploading jaws.20060515-1046.btnode3.hex (= actual CVS version + debug class 4,5,6,7,9 compiled)

• target control is working again, however, some nodes need to be reconfigured with jaws set mode dsn.

12.05.2006, 17:00 Jan Beutel

• The new JAWS version does not function with target control whereas the old version 20060501-0917 still works fine (no GUI error).
• The GUI node 00:BE has HCI OUT OF SYNC errors quite often. We should monitor this.

12.05.2006, 10:45 Matthias Dyer

Uploading new file jaws.20060512-1049.btnode3.hex with following changes:

• async inquiries/page from KM
• cdist has 2nd own thread fixing semaphore deadlock
• disabled reading of con-mgr-tree eeprom tuning params
• log classes >= 4 (mhop) are compiled, but logmask initially off (can be enabled on runtime with e.g. log set logmask 5 ff)
• moved log class target from 16 to 17 to avoid conflict with jaws app
• new logging classes added in gui_applet

10.05.2006, 12:00 Matthias Dyer

Uploaded new gui_applet (should fix the hang on :SL messages - test pending)
Plan for this afternoon: deploy new .hex for DSN nodes with asynchronous inq./page and fixed eeprom param.
After resetting some nodes -> 10 / 10 online. 2 nodes removed form manuall inspection

05.05.2006, 12:17 Jan Beutel

Only 5 nodes visible.

04.05.2006, 09:31 Jan Beutel

Only 6 nodes visible. Reset didn't change anything.

03.05.2006, 19:44 Jan Beutel

7 nodes online with jaws-20060501-0917

00:d4 00:d7 00:68 00:74 01:01 00:d0 00:d2

03.05.2006, 16:00 Kevin Martin

This morning, 10:00am, only 5 nodes were connected. Two of them (00:dc and 01:01) did not reset properly, i.e. they were not able to boot (not even a single LED was blinking). Thus, it seems that moving back to the old bootloader does not solve the reset/bootloader problem.

Node 01:0b had an older version of the tree con mgr which is not compatible to the newer version. Node 00:f0 did have an XTC connection manager which seemed to have confused node 00:68. Thus, i removed node 00:f0 from the network, rebooted node 00:68 and reprogrammed node 01:0b.

01.05.2006 Matthias Dyer

I have collected all blue boxes and programmed the old bootloader and the tree alg (COD 376). There still seems to be a mix between XTC and Tree. We should not use the same COD for both.