Messages in this thread | | | Date | Sun, 15 Mar 1998 13:15:29 -0500 (EST) | From | Greg Zornetzer <> | Subject | Re: kmod fix |
| |
On Sun, 15 Mar 1998, Nahshon wrote:
> > Sorry about the reply to my own message. > > Damn, I realized after I hung up that the kmod code I presented causes > > kmod to create little zombies of itself. I'm not really that strong on > > kernel thread stuff... Could anyone know why this is happening? Thanks, > > Greg, > Another problem with your patch is in the call to interruptible_sleep_on > in request_module. > Any signal may wake the caller before modprobe has finished and a lot will > be lost. Okay - I was under the impression that only that specific queue could cause request_module to wake up, but then my only source of information is quick perusals of the code
> > About zombies. The parent process (the main kmod thread) must wait for > its children. > > I'm working on a fix (based on your patch) that will fix these problems > and also remove the limit on the number of request_modules served at the > same time. Yeah - I was toying with the idea of implementing a linked list to store the data, but the only way I could figure to do this involved another system to prevent two functions from access the data at the same time, making the code _really_ nasty. Good luck.
> > Itai >
Greg Zornetzer - gaz+@andrew.cmu.edu "Light shines brightest in the darkest night" http://www.contrib.andrew.cmu.edu/~gaz
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |