Date: Mon, 16 Dec 1996 20:20:55 -0500 (EST) From: "Marc G. Fournier" <scrappy@hub.org> To: hackers@freebsd.org Subject: Almost have mmap() figured, I think... Message-ID: <Pine.NEB.3.95.961216200508.283F-100000@hub.org>
next in thread | raw e-mail | index | archive | help
Evening...
Altho I really hate to ask, I think I almost have this licked, but
there is something I'm still missing, so the code isn't *quite* working the
way it should.
Essentially, to simplify things while I "understand it better",
I've created n mmap()'d regions, each representing a frame. There are
n corresponding mmap()'d regions, each of which correspond to the
*size* of the frame. (I know, inelegant, but I can make it more
elegant *after* I know better that which I'm doing...)
The regions are being created as:
======
while(ii < FRAMES) {
sprintf(framestr, "/tmp/frame-%02d", ii);
/* Setup Video Stream Buffer */
if((fd = open(framestr, O_RDWR|O_CREAT, 0666)) < 0) {
fprintf(stderr, "cannot open %s\n", framestr);
exit(-1);
}
ftruncate(fd, FRAMESIZE);
if((frame[ii] =
mmap(0, FRAMESIZE, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0))
== (caddr_t) -1) {
fprintf(stderr, "cannot mmap frame buffer #%02d\n", ii);
exit(-1);
}
close(fd);
}
=====
Where frame[ii] is declared as "caddr_t frame[FRAMES];":$
Now, the information that is being written to each frame is
coming from a socket, which all worked before I switched over to mmap(),
and is read in as:
-----
void
readn(int fd, char *buf, size_t bytes)
{
int n;
char *p = buf;
while (bytes) {
n = read(fd, p, bytes);
if (n < 0)
fprintf(stderr, "read failed in send-photo.cgi");
if (n == 0 && bytes != 0)
fprintf(stderr, "short file on read, %d bytes left", bytes);
bytes -= n;
p += n;
}
}
----
And is being called as:
=====
readn(sockfd, frame[ii], datasize);
=====
Now, this is where I think I'm making my mistake, because the
results of running this code is:
hub> ./client
Frame == 000, ii == 00, datasize == 1944, Size == 1944
Frame == 001, ii == 01, datasize == 1981, Size == 1981
Frame == 002, ii == 02, datasize == 1972, Size == 1972
Frame == 003, ii == 03, datasize == 1990, Size == 1990
Frame == 004, ii == 04, datasize == 2006, Size == 2006
Frame == 005, ii == 05, datasize == 1985, Size == 1985
Frame == 006, ii == 06, datasize == 2021, Size == 2021
Frame == 007, ii == 07, datasize == 2034, Size == 2034
Frame == 008, ii == 08, datasize == 1984, Size == 1984
Frame == 009, ii == 09, datasize == 2023, Size == 2023
Frame == 010, ii == 10, datasize == 2047, Size == 2047
datasize is what I got from the remote server, through the socket,
Size is what the 'size' mmap()'d region sees, so that mapping is working
fine. The 'client' that is reading the mmap()'d region is seeing the
exact same size value as the server is writting to it...so that aspect
is working right.
The problem comes up when I try and print out the frame itself,
using 'printf("%s\n", frame[ii]);'...the value printed out each time
is exactly the same: <FF><D8><FF><E0>
Ack, okay...just tried something quickly, and changed %s to %d, and
the results are:
hub> ./client
Frame == 000, ii == 00, datasize == 1944, Size == 1944
134840320
Frame == 001, ii == 01, datasize == 1981, Size == 1981
134860800
Frame == 002, ii == 02, datasize == 1972, Size == 1972
134881280
Frame == 003, ii == 03, datasize == 1990, Size == 1990
134905856
Frame == 004, ii == 04, datasize == 2006, Size == 2006
134930432
So...am I accurate in assuming that what I'm getting is the memory
location? I still don't understand this fully, so please bear with me...I'm
*purely* guessing here :(
So, I'm making the assumption that either the way I'm calling, or
using, readn() is incorrect for use with mmap()'d regions...would I be
better off to read in the socket, then memcpy() the data, or strncpy()? Or
is what I'm missing even more simple then that?
Thanks again...
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.3.95.961216200508.283F-100000>