Hi all, Got two simple questions about the Linux kernel ........... 1. Why doesn't the kernel let all the processes use the CLONE_PID flag? What is the reason for allowing only the process with PID == 0 to use this feature. If this was allowed then threads of a single process could have the same PID as required by pthreads. Currently each thread gets a separate PID even though it is spawned from the same process (not what POSIX asks for). 2. When a process is running IT IS possible that an interrupt is received. Does the time needed for handling the interrupt taken from the process time slice? If yes , is it justified that the time slice of a process may be consumed by the kernel for handling interrupts even though the process has nothing to do with them. I know this is very unlikely but can't a process be starved because interrupts are arriving only during its time slot (or maybe the frequency of interrupts is very high). Ofcourse here I am talking about interrupt handlers that may be time consuming (due to errors or bad code). Please note - I do understand the need for interrupts that help in time accounting (ie the timer interrupt). I know these problems are a bit too newbie-ish (???) but I can't seem to proceed with uncertainties and doubts. I may have got the problems/concepts all wrong ....... but then thanx for reading anyway. -mohit -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ IRC Channel: irc.openprojects.net / #kernelnewbies Web Page: http://www.kernelnewbies.org/
