Who is the Process Error Message Handler?

[Eric Chen]

Specifically, is 'spawn.exe' necessarily responsible for displaying the
message? Or can some other entity, such as the kernel, handle it?

The statement in the project is not quite clear on whether 'spawn.exe'
is directly responsible or if the message only needs to be displayed
when the process limit is overrun when running 'spawn.exe' .

[Eric Chen]

Adel's comments earlier seem to allude more toward 'spawn.exe' being
directly responsible.

Consequently, depending on the final answer to cs412018's question,
'spawn.exe' would either be killed depending on the number of overruns
while printing the errors due to the syscall limit or would exit properly
before hitting that limit.

As such, I guess I'll keep it as I had and let the kernel handle it, since
an error message could be cut short the other way. Of course, maybe that's
what you want.

Post by Eric Chen
Specifically, is 'spawn.exe' necessarily responsible for displaying the
message? Or can some other entity, such as the kernel, handle it?
The statement in the project is not quite clear on whether 'spawn.exe'
is directly responsible or if the message only needs to be displayed
when the process limit is overrun when running 'spawn.exe' .

[cs412042]

I think that spawn.exe is responsible for detecting the -1 return
value from the system call and printing the appropriate error message.

The question is whether spawn.exe 10 should print two error messages or one.
The 9th and 10th new processes spawn.exe attempts to create will not be
created, since after the 8th, there will be 8 null.exe, 1 spawn.exe, and
1 shell.exe user processes running on the system. Adel posted that we should
terminate the process, but it is unclear which process he meant. The spec
says that spawn.exe should spawn P processes. So Adel either meant to
terminate the process which couldn't be spawned (and thus need not be
terminated) or to terminate spawn.exe (which will then not attempt to
spawn P processes).

David Renie

Post by Eric Chen
Adel's comments earlier seem to allude more toward 'spawn.exe' being
directly responsible.
Consequently, depending on the final answer to cs412018's question,
'spawn.exe' would either be killed depending on the number of overruns
while printing the errors due to the syscall limit or would exit properly
before hitting that limit.
As such, I guess I'll keep it as I had and let the kernel handle it, since
an error message could be cut short the other way. Of course, maybe that's
what you want.

[Philip]

from http://www.cs.umd.edu/class/fall2005/cmsc412/proj0/ :
If the limit has been reached then the process should not be created
and Sys_Spawn should fail by returning -1.
and regarding the sys calls
For the system call that exceeds the limit (i.e., the N+1st system
call), Sys_Exit should be called instead to kill the process.
I think that's pretty clear.

<blah blah>

[Eric Chen]

That's all fine, but it still doesn't state who is directly responsible for
reading the data and executing the message, which is waht the question was
asking.

[Eric Chen]

That's all fine, but it still doesn't state who is directly responsible for
interpreting the data and executing the message, which is what the question was
asking.