Re: [Linux-cachefs] Data Redundancy

[gaurav <gaurav1207 gmail com>]

thanks...
But i was talking in context with Linux 2.6 kernel. Anyways can i have a gud
article to read on this topic.
also there came another doubt in my mind while reading a paper on "unfied
buffering and caching mechanism" by Vivek Pai that :
if the page cache is shared between different domains and user level
programs so will the different programs use the same copy of page cache to
read or they wud copy the data to different page cache and then access.
If they use the same copy then what will happen if any particular process
want to update the data of the page cache.

On 11/20/06, David Howells <dhowells redhat com> wrote:
> gaurav <gaurav1207 gmail com> wrote:
>
> > sorry but i wanted to know exactly two points :
> > 1) Is default page cache(means not useing MMAP_SHARED()) is shareable
> among
> > different user level process .
>
> The pagecache is shared between all processes.  If MAP_PRIVATE is used,
> then
> pages in a private mapping *may* be shared up until those pages are
> written to.
>
> > 2)Is default page cache(means not useing MMAP_SHARED()) is shareable
> among
> > diffrent protection domain(means kernel and user adrress space)
> > thank you
>
> I'm not sure your question makes any sense.
>
> The kernel does things *to* the pagecache directly, but doesn't make use
> of it,
> unless bits of it are passed as data to various system calls.  Also, in
> general, there's no way of stopping the kernel from trampling on the
> pagecache
> at will if it wants to.
>
> I believe it differs to the M$ Windows kernel in that respect, if that's
> what
> you're thinking of.
>
> David
>
> --
> Linux-cachefs mailing list
> Linux-cachefs redhat com
> https://www.redhat.com/mailman/listinfo/linux-cachefs



--
Gaurav....