[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: concurrent access
Hello Lenonardo!
On Mon, 21 Feb 2000, Leonardo Lanzi wrote:
> Hi,
> in our lab we're using itifg-0.6a for image acquisition
> with an ADIMEC MX12P camera (1024x1024-12 bit)
> and an IC-PCI AM-DIG.
> We have 2 programs reading frames (in snap and select modes) ,
> one for displaying frame, the 2nd for image processing.
> Most important for us, is that the first process doesn't write a new
> frame
> until the 2nd has finished calculations.
> Is it possible to configure the driver in such a way that each process
> had
> a reserved physical memory area for the snapped frame?
>
I think so. With the specification of the UMB size YOU decide how many
frames reside in the upper memory area. With the GET_BMABUFFERS ->
GET_BMAADDR calls you get the pointers to different buffers into the
mmaped area.
The driver can fill every buffer you specify with SET_BMABUF. No frames
are transfered into this buffers, when SIGIO is blocked. A synchronization
between your both processes is not done by the driver.
Is this the answer to your question?
matthias
> Thanks,
>
> Leo
>
> --
> Leonardo Lanzi - lanzi@fi.infn.it
> Dipartimento di Fisica
> Universita' di Firenze
> Largo E. Fermi 2 - I 50125 - Firenze
> tel: +39 0552307764
> fax: +39 055229330
>
>
>
>