Often I get this message whan I attempt to link (create) an executable
file of a process that happens to be currently running. I have to shut
the process down before I can create a new one. This seems to be a
little strange. One clue is that the process starts up in the background
(it is a device driver using the io-char DDK). Also, it doesn’t happen
all the time, but often enough to warrant a query.
Geoff Roberts
Geoff <geoff@rtts.com.au> wrote in message
news:3DE710E9.FCF9DAAE@rtts.com.au…
Often I get this message whan I attempt to link (create) an executable
file of a process that happens to be currently running. I have to shut
the process down before I can create a new one. This seems to be a
little strange. One clue is that the process starts up in the background
(it is a device driver using the io-char DDK). Also, it doesn’t happen
all the time, but often enough to warrant a query.
Instead of shut the process down, you can “rm” the one first, and then
create it.
-xtang
Yes, I discovered this. But why should I have to do that in the first
place? Why should the linker care that the file it is creating already
has an instance executing and therefore not be able write a new one?
Seems a little odd to me.
Geoff.
Xiaodan Tang wrote:
Geoff <> geoff@rtts.com.au> > wrote in message
news:> 3DE710E9.FCF9DAAE@rtts.com.au> …
Often I get this message whan I attempt to link (create) an executable
file of a process that happens to be currently running. I have to shut
the process down before I can create a new one. This seems to be a
little strange. One clue is that the process starts up in the background
(it is a device driver using the io-char DDK). Also, it doesn’t happen
all the time, but often enough to warrant a query.Instead of shut the process down, you can “rm” the one first, and then
create it.-xtang
It’s because procnto has mmap()ed the file in, and has a SH_DENYWR lock
on the executable.
Geoff <geoff@rtts.com.au> wrote:
Yes, I discovered this. But why should I have to do that in the first
place? Why should the linker care that the file it is creating already
has an instance executing and therefore not be able write a new one?
Seems a little odd to me.Geoff.
Xiaodan Tang wrote:
Geoff <> geoff@rtts.com.au> > wrote in message
news:> 3DE710E9.FCF9DAAE@rtts.com.au> …
Often I get this message whan I attempt to link (create) an executable
file of a process that happens to be currently running. I have to shut
the process down before I can create a new one. This seems to be a
little strange. One clue is that the process starts up in the background
(it is a device driver using the io-char DDK). Also, it doesn’t happen
all the time, but often enough to warrant a query.Instead of shut the process down, you can “rm” the one first, and then
create it.-xtang
“Xiaodan Tang” <xtang@qnx.com> wrote in message
news:as7ss5$fk4$1@nntp.qnx.com…
Instead of shut the process down, you can “rm” the one first, and then
create it.-xtang
That brings up an interesting question. In QNX4 a file could be unlinked
while some process had it open. The process could continue to do IO to the
file even though the directory entry had been removed.
Does this still occur in QNX6?
“Bill Caroselli (Q-TPS)” <QTPS@earthlink.net> wrote:
“Xiaodan Tang” <> xtang@qnx.com> > wrote in message
news:as7ss5$fk4$> 1@nntp.qnx.com> …Instead of shut the process down, you can “rm” the one first, and then
create it.-xtang
That brings up an interesting question. In QNX4 a file could be unlinked
while some process had it open. The process could continue to do IO to the
file even though the directory entry had been removed.Does this still occur in QNX6?
Yes.
QNX Training Services
http://www.qnx.com/support/training/
Please followup in this newsgroup if you have further questions.