On 9/21/07, love_code meta <[EMAIL PROTECTED]> wrote:
> Now I 'm testing mmap function, and my pagesize is 4k bytes. After mapping
> 8k virtual memory, by dtrace I find there is only one pagefault when I
> access memory like that s[0] , s[4096].
>
> The following is my test spteps.
> 1. fd = open("/var/tmp/mmap.file", O_RDWR);
> 2. I'm sure fd is validate. then
> s = (char *)mmap(NULL, 8192,
> PROT_READ | PROT_WRITE,
> MAP_PRIVATE,
> fd, 0L);
> 3. for (i =0; i < 8192; i +=4096)
> a=s[i];
>
> Based on my knowledge, the step 3 should has two pagefault traps. because
> our pagesize is 4k, correct?
What is the the size of /var/tmp/mmap.file?
--
Just me,
Wire ...
Blog: <prstat.blogspot.com>
_______________________________________________
perf-discuss mailing list
perf-discuss@opensolaris.org
