Hi,
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
this occurs only when I work with photon (either locally or remote via
phindows). and it occurs repeatedly.
when I do a coldstart, this message does not occur for a while 
)
Does anybody know what might the reason for such a message (I assume
a piece of hardware, but do not know to identify 
)
tia for help
–
Joerg Kampmann
IBK-Consult - (embedded Systems)
WWW: http://www.ibk-consult.de
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- maschoen@pobox.com
I’d bet this is not in QSSL stuff.
If you want to be sure, there is maybe a way but it’s dedious.
Do a sin and note every program that is running. Then load
every file in vedit and search for the string"mem/pmem.c".
I use vedit cause I don’t know how to make grep work on binaries
files. Do not forget to check the share librairies as well
You could also redirect the output of the program started in the sysinit
to various console to help pin poin the program.
But I would lot at third party stuff or at your own first.
“Mitchell Schoenbrun” <maschoen@tsoft.com> wrote in message
news:snpl84ld3j132@corp.supernews.com…
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- > maschoen@pobox.com
Mario Charest wrote:
I’d bet this is not in QSSL stuff.
If you want to be sure, there is maybe a way but it’s dedious.
Do a sin and note every program that is running. Then load
every file in vedit and search for the string"mem/pmem.c".
I use vedit cause I don’t know how to make grep work on binaries
files. Do not forget to check the share librairies as well
How about:
strings file | grep “mem/pmem.c”
You could also redirect the output of the program started in the sysinit
to various console to help pin poin the program.
But I would lot at third party stuff or at your own first.
“Mitchell Schoenbrun” <> maschoen@tsoft.com> > wrote in message
news:> snpl84ld3j132@corp.supernews.com> …
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- > maschoen@pobox.com
This message comes from Proc… from the section on pysical memory
allocation.
The error message is coming from the function used to free memory
pages…
I don’t much more than this, someone more qualified would need to give
specifics.
Heather
Dean Douthat <ddouthat@faac.com> wrote:
Mario Charest wrote:
I’d bet this is not in QSSL stuff.
If you want to be sure, there is maybe a way but it’s dedious.
Do a sin and note every program that is running. Then load
every file in vedit and search for the string"mem/pmem.c".
I use vedit cause I don’t know how to make grep work on binaries
files. Do not forget to check the share librairies as well
How about:
strings file | grep “mem/pmem.c”
You could also redirect the output of the program started in the sysinit
to various console to help pin poin the program.
But I would lot at third party stuff or at your own first.
“Mitchell Schoenbrun” <> maschoen@tsoft.com> > wrote in message
news:> snpl84ld3j132@corp.supernews.com> …
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- > maschoen@pobox.com
Hi,
thanks for all the messages. I was a few days off
I’ll investigate that stuff …
Jörg
Heather Johnstone schrieb:
This message comes from Proc… from the section on pysical memory
allocation.
The error message is coming from the function used to free memory
pages…
I don’t much more than this, someone more qualified would need to give
specifics.
Heather
Dean Douthat <> ddouthat@faac.com> > wrote:
Mario Charest wrote:
I’d bet this is not in QSSL stuff.
If you want to be sure, there is maybe a way but it’s dedious.
Do a sin and note every program that is running. Then load
every file in vedit and search for the string"mem/pmem.c".
I use vedit cause I don’t know how to make grep work on binaries
files. Do not forget to check the share librairies as well
How about:
strings file | grep “mem/pmem.c”
You could also redirect the output of the program started in the sysinit
to various console to help pin poin the program.
But I would lot at third party stuff or at your own first.
“Mitchell Schoenbrun” <> maschoen@tsoft.com> > wrote in message
news:> snpl84ld3j132@corp.supernews.com> …
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- > maschoen@pobox.com
–
Joerg Kampmann
IBK-Consult - (embedded Systems)
WWW: http://www.ibk-consult.de
Hi,
another thing.
It’s plain QNX 4.25
see attachment - hopefully it’s readable …
thanks again
Jörg
Heather Johnstone schrieb:
This message comes from Proc… from the section on pysical memory
allocation.
The error message is coming from the function used to free memory
pages…
I don’t much more than this, someone more qualified would need to give
specifics.
Heather
Dean Douthat <> ddouthat@faac.com> > wrote:
Mario Charest wrote:
I’d bet this is not in QSSL stuff.
If you want to be sure, there is maybe a way but it’s dedious.
Do a sin and note every program that is running. Then load
every file in vedit and search for the string"mem/pmem.c".
I use vedit cause I don’t know how to make grep work on binaries
files. Do not forget to check the share librairies as well
How about:
strings file | grep “mem/pmem.c”
You could also redirect the output of the program started in the sysinit
to various console to help pin poin the program.
But I would lot at third party stuff or at your own first.
“Mitchell Schoenbrun” <> maschoen@tsoft.com> > wrote in message
news:> snpl84ld3j132@corp.supernews.com> …
sometimes I get a strange error message on my QNX 4 console (/dev/con1)
“mem/pmem.c(188) inserting [fffff,1] between [3271,d] and [fffff,1]”
This is just a wild guess. It looks like a diagnostic from a
a call to free() that is failing. The reason for the failure might
be that it thinks you are trying to free a value that is already
on the chain. The ffffff value many be suspicious, but it is hard
to tell.
I’d be surprised to find a diagnostic like this turned on in a
QNX release. Was this a beta, or are you using your own allocator?
Mitchell Schoenbrun --------- > maschoen@pobox.com
–
Joerg Kampmann
IBK-Consult - (embedded Systems)
WWW: http://www.ibk-consult.de