Networking/Discussion/Networking Drivers/Driver shutdown during ping/List of Posts
 
Forum Topic - Driver shutdown during ping: (4 Items)
   
Update
Roberto Braidotti(deleted)
09/10/2010 5:04 AM
post66680
Driver shutdown during ping   
I have a strange problem with a driver recompiled under QNX 6.4.1.
It works well but when I do a ping from another PC the driver goes in shutdown after 3 send/receive cycles.

The driver is devn-vortex.so for the internal DM&P Vortex86DX SoC LAN chip.
The 6.3.3 sources are supplied by DM&P.

Any idea to solve/debug this problem?

Regards,
Roberto Braidotti


Here the logs....

C:\Documents and Settings\BraidottiR>ping 20.20.20.21

Pinging 20.20.20.21 with 32 bytes of data:

Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Request timed out.

Ping statistics for 20.20.20.21:
    Packets: Sent = 4, Received = 3, Lost = 1 (25% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 0ms, Average = 0ms

C:\Documents and Settings\BraidottiR>



# io-pkt-v4 -d vortex verbose -p tcpip
# ifconfig en0 20.20.20.21
# sloginfo
Time             Sev Major Minor Args
Sep 10 11:04:15    5    14     0 tcpip starting
Sep 10 11:04:15    3    14     0 Using pseudo random generator.  See "random" op
tion
Sep 10 11:04:15    2    10     0 devn-vortex: Mapping device IO
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x3
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x5
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x13
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x14
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x15
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x16
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1b
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1c
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1e
Sep 10 11:04:16    5    10     0 devn-vortex: Using internal Phy detected at add
ress 0x1e
Sep 10 11:04:16    5    10     0
Sep 10 11:04:16    5    10     0 Vendor .............. 0x17f3
Sep 10 11:04:16    5    10     0 Device .............. 0x6040
Sep 10 11:04:16    5    10     0 Revision ............ 0x0
Sep 10 11:04:16    5    10     0 I/O port base ....... 0xdf00
Sep 10 11:04:16    5    10     0 Memory base ......... 0xfefffc00
Sep 10 11:04:16    5    10     0 Interrupt ........... 0xb
Sep 10 11:04:16    5    10     0 MAC address ......... 000060 000001
Sep 10 11:04:16    5    14     0 io-pkt shim
Sep 10 11:04:16    5    14     0 Vendor .............. 0x0
Sep 10 11:04:16    5    14     0 Device .............. 0x0
Sep 10 11:04:16    5    14     0 Revision ............ 0x0
Sep 10 11:04:16    5    14     0 MAC address ......... 000060 000001
Sep 10 11:04:19    5    10     0 devn-vortex: HPNA link up (Unknown Link)
Sep 10 11:04:19    5    10     0 devn-vortex: HPNA link up (Unknown Link)
Sep 10 11:04:28    5    10     0 devn-vortex: Ethernet link up (Unknown Link)
Sep 10 11:04:33    5    10     0 devn-vortex: Starting Shutdown1(hdl=080ed540)
Sep 10 11:04:33    5    10     0 devn-vortex: Finished Shutdown1(hdl=080ed540)
Sep 10 11:04:33    5    10     0 devn-vortex: Starting Shutdown2(hdl=080ed540)
Andrew Boyd(deleted)
09/10/2010 7:34 AM
post66683
RE: Driver shutdown during ping   
Start the driver with "verbose=100" - anything in
the sloginfo after the failure?
 
Also, compile the driver "-g -O0" and when it sigsegv's
dumper will create and io-pkt-vX.core file in /var/dumps
which you can look at with
 
  gdb /sbin/io-pkt-vX /var/dumps/io-pkt-vX.core
 
A quick bt will generally show you where it's going
off the rails.
 
--
aboyd

________________________________

From: Roberto Braidotti [mailto:community-noreply@qnx.com]
Sent: Fri 9/10/2010 2:04 AM
To: drivers-networking
Subject: Driver shutdown during ping



I have a strange problem with a driver recompiled under QNX 6.4.1.
It works well but when I do a ping from another PC the driver goes in shutdown after 3 send/receive cycles.

The driver is devn-vortex.so for the internal DM&P Vortex86DX SoC LAN chip.
The 6.3.3 sources are supplied by DM&P.

Any idea to solve/debug this problem?

Regards,
Roberto Braidotti


Here the logs....

C:\Documents and Settings\BraidottiR>ping 20.20.20.21

Pinging 20.20.20.21 with 32 bytes of data:

Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Reply from 20.20.20.21: bytes=32 time<1ms TTL=255
Request timed out.

Ping statistics for 20.20.20.21:
    Packets: Sent = 4, Received = 3, Lost = 1 (25% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 0ms, Average = 0ms

C:\Documents and Settings\BraidottiR>



# io-pkt-v4 -d vortex verbose -p tcpip
# ifconfig en0 20.20.20.21
# sloginfo
Time             Sev Major Minor Args
Sep 10 11:04:15    5    14     0 tcpip starting
Sep 10 11:04:15    3    14     0 Using pseudo random generator.  See "random" op
tion
Sep 10 11:04:15    2    10     0 devn-vortex: Mapping device IO
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x3
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x5
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x13
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x14
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x15
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x16
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1b
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1c
Sep 10 11:04:16    5    10     0 devn-vortex: PHY found: 0x1e
Sep 10 11:04:16    5    10     0 devn-vortex: Using internal Phy detected at add
ress 0x1e
Sep 10 11:04:16    5    10     0
Sep 10 11:04:16    5    10     0 Vendor .............. 0x17f3
Sep 10 11:04:16    5    10     0 Device .............. 0x6040
Sep 10 11:04:16    5    10     0 Revision ............ 0x0
Sep 10 11:04:16    5    10     0 I/O port base ....... 0xdf00
Sep 10 11:04:16    5    10     0 Memory base ......... 0xfefffc00
Sep 10 11:04:16    5    10     0 Interrupt ........... 0xb
Sep 10 11:04:16    5    10     0 MAC address ......... 000060 000001
Sep 10 11:04:16    5    14     0 io-pkt shim
Sep 10 11:04:16    5    14     0 Vendor .............. 0x0
Sep 10 11:04:16    5    14     0 Device .............. 0x0
Sep 10 11:04:16    5    14     0 Revision ............ 0x0
Sep 10 11:04:16    5    14     0 MAC address ......... 000060 000001
Sep 10 11:04:19    5    10     0 devn-vortex: HPNA link up (Unknown Link)
Sep 10 11:04:19    5    10     0 devn-vortex: HPNA link up (Unknown Link)
Sep 10 11:04:28    5    10     0 devn-vortex: Ethernet link up (Unknown Link)
Sep 10 11:04:33    5    10     0 devn-vortex: Starting Shutdown1(hdl=080ed540)
Sep 10 11:04:33    5    10     0 devn-vortex: Finished Shutdown1(hdl=080ed540)
Sep 10 11:04:33    5    10     0 devn-vortex: Starting Shutdown2(hdl=080ed540)




_______________________________________________

Networking Drivers
http://community.qnx.com/sf/go/post66680
Roberto Braidotti(deleted)
09/10/2010 8:32 AM
post66688
Re: Driver shutdown during ping   
> Start the driver with "verbose=100" - anything in
> the sloginfo after the failure?
Nothing but the same. Shutdown1, shutdown2.

> Also, compile the driver "-g -O0" and when it sigsegv's
> dumper will create and io-pkt-vX.core file in /var/dumps
> which you can look at with
Compiled with -g (make DEBUG=-g). 
It works ... no shutdown on ping.   8-O
Any idea!?

Thank you very much.

Regards,
Roberto Braidotti

>   gdb /sbin/io-pkt-vX /var/dumps/io-pkt-vX.core
>  
> A quick bt will generally show you where it's going
> off the rails.
>  
> --
> aboyd
>
Andrew Boyd(deleted)
09/10/2010 9:04 AM
post66695
RE: Driver shutdown during ping   
> Compiled with -g ... It works ... Any idea!?

Ship it!!!

j/k  :)

That in itself is an interesting clue.  Ok,
recompile it without debug, get a core file
and start disassembling to try to figure out
what's going on.

Most easily-reproducible problems - like yours
appeared to be, at first - are generally pretty
low-hanging fruit.  Null pointer reference, etc
which are generally easily diagnosed with a 
debug-compile and a core dump in gdb.

--
aboyd

Return