When I compile my project with profiling enabled and attempt to run it on my
remote target, over ethernet, via the IDE, I receive the following error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program there it
initiates properly, however, my program cannot run on my native development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem is
perhaps with shmem or something. Can anyone give me any clues as to how to
correct for this error?
Jason
Can you try pointing TMPDIR on your target to something other
than /dev/shmem? What target CPU are you using?
Jason Farque <jasonf@pigging.com> wrote:
When I compile my project with profiling enabled and attempt to run it on my
remote target, over ethernet, via the IDE, I receive the following error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program there it
initiates properly, however, my program cannot run on my native development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem is
perhaps with shmem or something. Can anyone give me any clues as to how to
correct for this error?
Jason
–
cburgess@qnx.com
I have tried moving the TMPDIR to / and other locations without success.
The target is x86, the dev machine is x86. Could this be a permissions
issue?
Jason
“Colin Burgess” <cburgess@qnx.com> wrote in message
news:ajtu6m$m3$1@nntp.qnx.com…
Can you try pointing TMPDIR on your target to something other
than /dev/shmem? What target CPU are you using?
Jason Farque <> jasonf@pigging.com> > wrote:
When I compile my project with profiling enabled and attempt to run it
on my
remote target, over ethernet, via the IDE, I receive the following error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program there
it
initiates properly, however, my program cannot run on my native
development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem is
perhaps with shmem or something. Can anyone give me any clues as to how
to
correct for this error?
Jason
\
cburgess@qnx.com
Jason Farque <jasonf@pigging.com> wrote:
I have tried moving the TMPDIR to / and other locations without success.
The target is x86, the dev machine is x86. Could this be a permissions
issue?
Could be. Does qconn have permissions to create directories in /tmp on your
target?
Jason
“Colin Burgess” <> cburgess@qnx.com> > wrote in message
news:ajtu6m$m3$> 1@nntp.qnx.com> …
Can you try pointing TMPDIR on your target to something other
than /dev/shmem? What target CPU are you using?
Jason Farque <> jasonf@pigging.com> > wrote:
When I compile my project with profiling enabled and attempt to run it
on my
remote target, over ethernet, via the IDE, I receive the following error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program there
it
initiates properly, however, my program cannot run on my native
development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem is
perhaps with shmem or something. Can anyone give me any clues as to how
to
correct for this error?
Jason
\
cburgess@qnx.com
–
cburgess@qnx.com
The target is an embedded system that runs as root, so I don’t see why not.
Of course on the development machine I’m not root so I’m wondering if that
has something to do with it.
Jason
“Colin Burgess” <cburgess@qnx.com> wrote in message
news:aju34k$68i$2@nntp.qnx.com…
Jason Farque <> jasonf@pigging.com> > wrote:
I have tried moving the TMPDIR to / and other locations without success.
The target is x86, the dev machine is x86. Could this be a permissions
issue?
Could be. Does qconn have permissions to create directories in /tmp on
your
target?
Jason
“Colin Burgess” <> cburgess@qnx.com> > wrote in message
news:ajtu6m$m3$> 1@nntp.qnx.com> …
Can you try pointing TMPDIR on your target to something other
than /dev/shmem? What target CPU are you using?
Jason Farque <> jasonf@pigging.com> > wrote:
When I compile my project with profiling enabled and attempt to run
it
on my
remote target, over ethernet, via the IDE, I receive the following
error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program
there
it
initiates properly, however, my program cannot run on my native
development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem
is
perhaps with shmem or something. Can anyone give me any clues as to
how
to
correct for this error?
Jason
\
cburgess@qnx.com
Jason Farque <jasonf@pigging.com> wrote:
The target is an embedded system that runs as root, so I don’t see why not.
Of course on the development machine I’m not root so I’m wondering if that
has something to do with it.
Just to check - you do have a /tmp directory on your target, don’t you?
Jason
“Colin Burgess” <> cburgess@qnx.com> > wrote in message
news:aju34k$68i$> 2@nntp.qnx.com> …
Jason Farque <> jasonf@pigging.com> > wrote:
I have tried moving the TMPDIR to / and other locations without success.
The target is x86, the dev machine is x86. Could this be a permissions
issue?
Could be. Does qconn have permissions to create directories in /tmp on
your
target?
Jason
“Colin Burgess” <> cburgess@qnx.com> > wrote in message
news:ajtu6m$m3$> 1@nntp.qnx.com> …
Can you try pointing TMPDIR on your target to something other
than /dev/shmem? What target CPU are you using?
Jason Farque <> jasonf@pigging.com> > wrote:
When I compile my project with profiling enabled and attempt to run
it
on my
remote target, over ethernet, via the IDE, I receive the following
error
dialog:
Profiler Error: Open Not enough memory
If I run qconn on my localhost and attempt to profile the program
there
it
initiates properly, however, my program cannot run on my native
development
machine due to the absence of required hardware.
There is quite a bit of RAM on both machines so I suspect the problem
is
perhaps with shmem or something. Can anyone give me any clues as to
how
to
correct for this error?
Jason
\
cburgess@qnx.com
–
cburgess@qnx.com
Hmmm very odd. I did have a /tmp set up as a symlink to /dev/shmem, but
when I went into it and did a ‘ls’ it said “no such directory .” (note the
dot). I removed the symlink in System Builder and created a physical /tmp
directory and things are starting to work now. Still not configured
properly but definitely further along than I was.
Jason
Just to check - you do have a /tmp directory on your target, don’t you?
Colin Burgess <cburgess@qnx.com> wrote:
Jason Farque <> jasonf@pigging.com> > wrote:
The target is an embedded system that runs as root, so I don’t see why not.
Of course on the development machine I’m not root so I’m wondering if that
has something to do with it.
Just to check - you do have a /tmp directory on your target, don’t you?
Also, make sure that /tmp is not a link to /dev/shmem since you can’t make
dirs in /dev/shmem.
chris
–
Chris McKillop <cdm@qnx.com> “The faster I go, the behinder I get.”
Software Engineer, QSSL – Lewis Carroll –
http://qnx.wox.org/
Jason…
When I create a file system on an embedded system, I literally create
the partitions manually, otherwise the OS will create them for you with
links. This is my experience, but I do not know the correct details of
the why/how questions. Maybe someone else can explain.
regards…
Miguel.
Jason Farque wrote:
Hmmm very odd. I did have a /tmp set up as a symlink to /dev/shmem, but
when I went into it and did a ‘ls’ it said “no such directory .” (note the
dot). I removed the symlink in System Builder and created a physical /tmp
directory and things are starting to work now. Still not configured
properly but definitely further along than I was.
Jason
Just to check - you do have a /tmp directory on your target, don’t you?