Adam Cartwright
|
io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/25/2008 4:58 AM
post17241
|
io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
Hello QNX user,
We are using the new Momentics Version 4.5 (qnxsdp-6.4.0-200810211530-win32.exe) with QNX 6.4.0. We included the BSP for
QNX 6.3.2 (bsp-freescale-mgt5200b-1.0.4-20070910-binsrc.zip) for our evaluation board MPC5200b-Lite. We did recompile "
devn-mpc5200.so" from the BSP-source-package; for bestcomm5200 and libdma-bestcomm5200.so we use the BSP-binaries from
QNX 6.3.2.
Instead of "io-net" we use the new network manager "io-pkt-v4".
In the *.bsh file we call the following command:
io-pkt-v4 -dmpc5200 mac=zzzzzzzzzz,verbose -ptcpip waitfor /dev/io-net/en0 10 ifconfig en0 zz.zz.zz.zz netmask zz.zz.zz.
zz if_up -a en0
After that, the network runs fine and the connection via qconn to the IDE is working, too. Unfortunately, "io-pkt" is
shutting down after a while (this could be a few minutes or a few hours) with the following message:
Process 4102 (io-pkt-v4) terminated SIGSEGV code=2 fltno=5 ip=480881a8(io-pkt-v4@main+0x904) ref=00000000
Has anybody an idea how to solve this problem?
|
|
|
Sean Boudreau(deleted)
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Sean Boudreau(deleted)
11/25/2008 8:58 AM
post17256
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
On Tue, Nov 25, 2008 at 04:58:08AM -0500, Adam Cartwright wrote:
> Hello QNX user,
>
> We are using the new Momentics Version 4.5 (qnxsdp-6.4.0-200810211530-win32.exe) with QNX 6.4.0. We included the BSP
for QNX 6.3.2 (bsp-freescale-mgt5200b-1.0.4-20070910-binsrc.zip) for our evaluation board MPC5200b-Lite. We did
recompile "devn-mpc5200.so" from the BSP-source-package; for bestcomm5200 and libdma-bestcomm5200.so we use the BSP-
binaries from QNX 6.3.2.
> Instead of "io-net" we use the new network manager "io-pkt-v4".
> In the *.bsh file we call the following command:
>
> io-pkt-v4 -dmpc5200 mac=zzzzzzzzzz,verbose -ptcpip waitfor /dev/io-net/en0 10 ifconfig en0 zz.zz.zz.zz netmask zz.zz.
zz.zz if_up -a en0
>
> After that, the network runs fine and the connection via qconn to the IDE is working, too. Unfortunately, "io-pkt" is
shutting down after a while (this could be a few minutes or a few hours) with the following message:
>
> Process 4102 (io-pkt-v4) terminated SIGSEGV code=2 fltno=5 ip=480881a8(io-pkt-v4@main+0x904) ref=00000000
>
> Has anybody an idea how to solve this problem?
Please tar up your io-pkt-v4, devn-mpc5200.so,
devnp-shim.so, /lib/libc.so.3 and /var/dumps/io-pkt-v4.core
and I'll have a look.
-seanb
|
|
|
Adam Cartwright
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/25/2008 11:16 AM
post17282
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
See attached the requested files.
|
|
|
Robert Craig
|
RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Robert Craig
11/25/2008 11:24 AM
post17284
|
RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
Sounds very similar to this post in Networking Drivers
http://community.qnx.com/sf/go/post15246
Does this help?
Robert.
-----Original Message-----
From: Adam Cartwright [mailto:community-noreply@qnx.com]
Sent: Tuesday, November 25, 2008 11:17 AM
To: momentics-community
Subject: Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
See attached the requested files.
_______________________________________________
QNX Momentics Community Support
http://community.qnx.com/sf/go/post17282
|
|
|
Adam Cartwright
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/25/2008 11:37 AM
post17285
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
We already tried that but it didn't have any success.
io-pkt is still shutting down.
|
|
|
Sean Boudreau(deleted)
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Sean Boudreau(deleted)
11/25/2008 11:37 AM
post17286
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> Sounds very similar to this post in Networking Drivers
>
> http://community.qnx.com/sf/go/post15246
>
> Does this help?
>
> Robert.
>
It looks like it could be the same issue from the core.
-seanb
|
|
|
Adam Cartwright
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/26/2008 4:36 AM
post17340
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
> On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > Sounds very similar to this post in Networking Drivers
> >
> > http://community.qnx.com/sf/go/post15246
> >
> > Does this help?
> >
> > Robert.
> >
>
> It looks like it could be the same issue from the core.
>
> -seanb
As mentioned above, I already tried that, but without success. Still io-pkt shutdown.
Meanwhile I tried another constellation:
In the "io-pkt shutdown-cases" my target was connected to our company network (and my host PC) via LAN.
Now I unplugged the target from network and connected it to my host via cross-patch cable (no more company network!). In
this case, io-pkt was working fine for many hours without shutdown.
This could only be a workaround but it is no solution of long duration.
Adam.
|
|
|
Sean Boudreau(deleted)
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Sean Boudreau(deleted)
11/26/2008 9:27 AM
post17363
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
On Wed, Nov 26, 2008 at 04:36:12AM -0500, Adam Cartwright wrote:
> > On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > > Sounds very similar to this post in Networking Drivers
> > >
> > > http://community.qnx.com/sf/go/post15246
> > >
> > > Does this help?
> > >
> > > Robert.
> > >
> >
> > It looks like it could be the same issue from the core.
> >
> > -seanb
>
>
> As mentioned above, I already tried that, but without success. Still io-pkt shutdown.
> Meanwhile I tried another constellation:
> In the "io-pkt shutdown-cases" my target was connected to our company network (and my host PC) via LAN.
> Now I unplugged the target from network and connected it to my host via cross-patch cable (no more company network!).
In this case, io-pkt was working fine for many hours without shutdown.
>
> This could only be a workaround but it is no solution of long duration.
I still think it's the same issue. The proposed fix might
not address it.
-seanb
|
|
|
Adam Cartwright
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/26/2008 10:14 AM
post17380
|
Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
> On Wed, Nov 26, 2008 at 04:36:12AM -0500, Adam Cartwright wrote:
> > > On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > > > Sounds very similar to this post in Networking Drivers
> > > >
> > > > http://community.qnx.com/sf/go/post15246
> > > >
> > > > Does this help?
> > > >
> > > > Robert.
> > > >
> > >
> > > It looks like it could be the same issue from the core.
> > >
> > > -seanb
> >
> >
> > As mentioned above, I already tried that, but without success. Still io-pkt
> shutdown.
> > Meanwhile I tried another constellation:
> > In the "io-pkt shutdown-cases" my target was connected to our company
> network (and my host PC) via LAN.
> > Now I unplugged the target from network and connected it to my host via
> cross-patch cable (no more company network!). In this case, io-pkt was working
> fine for many hours without shutdown.
> >
> > This could only be a workaround but it is no solution of long duration.
>
> I still think it's the same issue. The proposed fix might
> not address it.
>
> -seanb
I'm not sure if I did understand you correctly. You mean the issue from the linked post, don't you? I corrected the "
transmit.c" file as described in that post. This correction is already included in the coredump which I attached
yesterday. But it didn't have any success.
In that post the described problem seems to be a little bit different from mine, because the user there only has a mutex
lock, but I have a SIGSEGV termination.
Adam.
|
|
|
Robert Craig
|
RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Robert Craig
11/26/2008 10:19 AM
post17382
|
RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
What we're saying is that the fix might not have been complete :>.
The crash is definitely due to the driver though. The reason why
disconnecting from the company LAN makes things better is simply because
you're pushing less packets through the interface and therefore giving
it less chance to hit the condition that causes the crash.
Robert.
-----Original Message-----
From: Adam Cartwright [mailto:community-noreply@qnx.com]
Sent: Wednesday, November 26, 2008 10:14 AM
To: momentics-community
Subject: Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
> On Wed, Nov 26, 2008 at 04:36:12AM -0500, Adam Cartwright wrote:
> > > On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > > > Sounds very similar to this post in Networking Drivers
> > > >
> > > > http://community.qnx.com/sf/go/post15246
> > > >
> > > > Does this help?
> > > >
> > > > Robert.
> > > >
> > >
> > > It looks like it could be the same issue from the core.
> > >
> > > -seanb
> >
> >
> > As mentioned above, I already tried that, but without success. Still
> > io-pkt
> shutdown.
> > Meanwhile I tried another constellation:
> > In the "io-pkt shutdown-cases" my target was connected to our
> > company
> network (and my host PC) via LAN.
> > Now I unplugged the target from network and connected it to my host
> > via
> cross-patch cable (no more company network!). In this case, io-pkt was
> working fine for many hours without shutdown.
> >
> > This could only be a workaround but it is no solution of long
duration.
>
> I still think it's the same issue. The proposed fix might not address
> it.
>
> -seanb
I'm not sure if I did understand you correctly. You mean the issue from
the linked post, don't you? I corrected the "transmit.c" file as
described in that post. This correction is already included in the
coredump which I attached yesterday. But it didn't have any success.
In that post the described problem seems to be a little bit different
from mine, because the user there only has a mutex lock, but I have a
SIGSEGV termination.
Adam.
_______________________________________________
QNX Momentics Community Support
http://community.qnx.com/sf/go/post17380
|
|
|
Christoph Nemmaier
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Christoph Nemmaier
11/28/2008 4:50 AM
post17509
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
Sounds exactly the same as our problem.
You think that devn-mgt5200b.so is responsible?
So, is there the possibility to get a preliminary version of
devnp-mgt5200b.so from the freescale-lite5200b-BSP, which is coming soon?
C.W.
> What we're saying is that the fix might not have been complete :>.
>
> The crash is definitely due to the driver though. The reason why
> disconnecting from the company LAN makes things better is simply because
> you're pushing less packets through the interface and therefore giving
> it less chance to hit the condition that causes the crash.
>
> Robert.
>
> -----Original Message-----
> From: Adam Cartwright [mailto:community-noreply@qnx.com]
> Sent: Wednesday, November 26, 2008 10:14 AM
> To: momentics-community
> Subject: Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
>
> > On Wed, Nov 26, 2008 at 04:36:12AM -0500, Adam Cartwright wrote:
> > > > On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > > > > Sounds very similar to this post in Networking Drivers
> > > > >
> > > > > http://community.qnx.com/sf/go/post15246
> > > > >
> > > > > Does this help?
> > > > >
> > > > > Robert.
> > > > >
> > > >
> > > > It looks like it could be the same issue from the core.
> > > >
> > > > -seanb
> > >
> > >
> > > As mentioned above, I already tried that, but without success. Still
>
> > > io-pkt
> > shutdown.
> > > Meanwhile I tried another constellation:
> > > In the "io-pkt shutdown-cases" my target was connected to our
> > > company
> > network (and my host PC) via LAN.
> > > Now I unplugged the target from network and connected it to my host
> > > via
> > cross-patch cable (no more company network!). In this case, io-pkt was
>
> > working fine for many hours without shutdown.
> > >
> > > This could only be a workaround but it is no solution of long
> duration.
> >
> > I still think it's the same issue. The proposed fix might not address
>
> > it.
> >
> > -seanb
>
>
> I'm not sure if I did understand you correctly. You mean the issue from
> the linked post, don't you? I corrected the "transmit.c" file as
> described in that post. This correction is already included in the
> coredump which I attached yesterday. But it didn't have any success.
> In that post the described problem seems to be a little bit different
> from mine, because the user there only has a mutex lock, but I have a
> SIGSEGV termination.
>
> Adam.
>
> _______________________________________________
> QNX Momentics Community Support
> http://community.qnx.com/sf/go/post17380
|
|
|
Adam Cartwright
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Adam Cartwright
11/28/2008 8:37 AM
post17530
|
Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
> Sounds exactly the same as our problem.
> You think that devn-mgt5200b.so is responsible?
> So, is there the possibility to get a preliminary version of
> devnp-mgt5200b.so from the freescale-lite5200b-BSP, which is coming soon?
> C.W.
Hello Christoph,
could that be, that you mean the "devn-mpc5200.so" respectively "devnp-mpc5200.so"?
I think so, because the "devn-mgt5200b.so" seems not to exist.
;-)
Adam.
|
|
|
Robert Craig
|
RE: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
|
Robert Craig
11/28/2008 11:32 AM
post17553
|
RE: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
It definitely looks like a driver problem.
I believe that the initial BSP port will simply re-use the io-net
driver, but obviously with appropriate bug fixes included. I'm not sure
when this is scheduled to be done though.
There's an internal PR tracking this issue, but there's no feedback
available just yet.
Robert.
-----Original Message-----
From: Christoph Nemmaier [mailto:community-noreply@qnx.com]
Sent: Friday, November 28, 2008 4:51 AM
To: momentics-community
Subject: Re: RE: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
Sounds exactly the same as our problem.
You think that devn-mgt5200b.so is responsible?
So, is there the possibility to get a preliminary version of
devnp-mgt5200b.so from the freescale-lite5200b-BSP, which is coming
soon?
C.W.
> What we're saying is that the fix might not have been complete :>.
>
> The crash is definitely due to the driver though. The reason why
> disconnecting from the company LAN makes things better is simply
> because you're pushing less packets through the interface and
> therefore giving it less chance to hit the condition that causes the
crash.
>
> Robert.
>
> -----Original Message-----
> From: Adam Cartwright [mailto:community-noreply@qnx.com]
> Sent: Wednesday, November 26, 2008 10:14 AM
> To: momentics-community
> Subject: Re: io-pkt-v4 terminated SIGSEGV (QNX 6.4 MPC5200b)
>
> > On Wed, Nov 26, 2008 at 04:36:12AM -0500, Adam Cartwright wrote:
> > > > On Tue, Nov 25, 2008 at 11:24:56AM -0500, Robert Craig wrote:
> > > > > Sounds very similar to this post in Networking Drivers
> > > > >
> > > > > http://community.qnx.com/sf/go/post15246
> > > > >
> > > > > Does this help?
> > > > >
> > > > > Robert.
> > > > >
> > > >
> > > > It looks like it could be the same issue from the core.
> > > >
> > > > -seanb
> > >
> > >
> > > As mentioned above, I already tried that, but without success.
> > > Still
>
> > > io-pkt
> > shutdown.
> > > Meanwhile I tried another constellation:
> > > In the "io-pkt shutdown-cases" my target was connected to our
> > > company
> > network (and my host PC) via LAN.
> > > Now I unplugged the target from network and connected it to my
> > > host via
> > cross-patch cable (no more company network!). In this case, io-pkt
> > was
>
> > working fine for many hours without shutdown.
> > >
> > > This could only be a workaround but it is no solution of long
> duration.
> >
> > I still think it's the same issue. The proposed fix might not
> > address
>
> > it.
> >
> > -seanb
>
>
> I'm not sure if I did understand you correctly. You mean the issue
> from the linked post, don't you? I corrected the "transmit.c" file as
> described in that post. This correction is already included in the
> coredump which I attached yesterday. But it didn't have any success.
> In that post the described problem seems to be a little bit different
> from mine, because the user there only has a mutex lock, but I have a
> SIGSEGV termination.
>
> Adam.
>
> _______________________________________________
> QNX Momentics Community Support
> http://community.qnx.com/sf/go/post17380
_______________________________________________
QNX Momentics Community Support
http://community.qnx.com/sf/go/post17509
|
|
|
|