Date: Tue, 17 Dec 1996 19:12:25 -0500 (EST) From: "Marc G. Fournier" <scrappy@hub.org> To: hackers@freebsd.org Subject: Thanks to all (re: MMAP) Message-ID: <Pine.NEB.3.95.961217190125.8167C-100000@hub.org>
next in thread | raw e-mail | index | archive | help
Hi...
Thanks to everyone for the invaluable help with getting
this mmap() stuff understood...
The most recent problem that I was seeing was *purely*
a coding error, and nothing more :( Everything else works like a
charm :)
Terry...thanks for the pointers about how to reduce the
number of maps required...will implement that final one, of
merging sizeof(int) + FRAMESIZE, tonight...
But, have one question concerning that first.
When I do the socket and want to send the size across the
socket, I use:
write(sockfd, (char *)&datasize, sizeof(datasize));
How would I do similar with the mmap()'d region? What
I'm doing right now is, other then inefficient:
for (;;) {
for(ii = 0; ii < FRAMES; ii++) {
readn(sockfd, (char *)&datasize, sizeof(datasize));
if (datasize == 0)
break;
sprintf(vsize + (ii * sizeof(int)), "%d", datasize);
readn(sockfd, video + (ii * FRAMESIZE), datasize);
}
}
Is this as correct as any other way? I know it works, but
if I merge the two into one mmap'd region, I would think I'd be
looking at something like:
for (;;) {
for(ii = 0; ii < FRAMES; ii++) {
readn(sockfd, (char *)&datasize, sizeof(datasize));
if (datasize == 0)
break;
sprintf(vsize + (ii * (sizeof(int) + FRAMESIZE)), "%d", datasize);
readn(sockfd, video + (ii * (sizeof(int) + FRAMESIZE)) + sizeof(int),
datasize);
}
}
But I would have thought that there was a better method of doing
the sprintf() above? Somehow, it *feels* wrong...
Marc G. Fournier scrappy@hub.org
Systems Administrator @ hub.org scrappy@freebsd.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.95.961217190125.8167C-100000>