[Linux-cachefs] (no subject)

[gaurav]

thats rite but wen all processes have different address spaces so cant we
say that each process have duplicated copy of same data?

On 1/17/07, Peter Staubach <staubach at redhat.com> wrote:
>
> gaurav wrote:
> > i am asking this doubt with respect to block devices files.... with
> > respect
> > to cache only
> > as we know each process have different address spaces so thus this means
> > that if any general process access block device data ,which already is
> > present in page cache, will allocate new buffer for the data...  or
> > indeed
> > uses the same buffer which is already there to access the data.Jus i
> > want to
> > know that new copies of data is created on every access of data ??
>
> This question isn't particularly relevant to this group, but we can pursue
> it a little further.
>
> Access to a block device, via the read(2) and write(2) calls, results in
> cached blocks from the block device being shared.  When a process reads
> a block from the block device, the page containing the data is located in
> the page cache and is then copied into the buffer, which was specific in
> the read(2) or write(2) arguments, in the address space of the process.
> If another process running on the same system wants to access the same
> block of the block device, then the same page in the page cache will be
> used to satisfy requests from this second process.
>
>       ps
>
> --
> Linux-cachefs mailing list
> Linux-cachefs at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cachefs
>



-- 
Gaurav....