[prev in list] [next in list] [prev in thread] [next in thread]
List: linux-kernel
Subject: Kernel page count management
From: Mathias Waack <mathias () atoll-net ! de>
Date: 2000-02-07 9:36:14
[Download RAW message or body]
Hi all,
sorry for bothering you, but I've asked this question in the usenet
without any answers.
So, that's my problem:
If I request a page by calling get_free_page, the kernel
sets the usage count for this page to 1. So far so good, but if
I request more pages by calling __get_free_pages(GFP...,order)
it sets only the usage count of the first returned page.
Some sample code:
#include <linux/module.h>
#include <linux/kernel.h>
#include <asm/io.h>
int init_module(void)
{
int order;
unsigned long mem;
for (order = 3; order >= 0; --order) {
printk(KERN_INFO "checking for order %d\n",order);
mem = __get_free_pages(GFP_ATOMIC, order);
if (!mem) printk(KERN_INFO "got no mem\n");
else {
int i;
for (i=0;i < 1<<order; ++i) {
printk("page %d hast count %d\n",i, atomic_read(&mem_map[MAP_NR(mem + PAGE_SIZE*i)].count));
}
free_pages(mem,order);
}
}
return -ENOMEM;
}
void cleanup_module(void)
{
}
I've compiled the code with
cc -D__SMP__ -D__KERNEL__ -DMODULE -DSIMU -O -Wall -I/usr/include -c main.c -o main.o
Can anybody explain this behaviour to me? Where is my mistake?
As far as I understand the kernel, the usage count for all pages should
be set to 1. Should I do it by hand after allocation?
If I do nothing but simple use this memory (for mmapping it to user space later), the
system crashes from time to time. I think (or better: I hope) my other code is correct...
I'm using a dual Pentium running a 2.2.14 kernel. The kernel is configured correct and
runs stable (if I do not use my own code, of course;)
Thanks for your answers
Mathias
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic