Synchronous vs asynchronous

Questions and discussions about the Xillybus IP core and drivers

Synchronous vs asynchronous

Postby andylgh »

In all demo bundles, the xillybus read * and xillybus write * streams are asynchronous.xillybus mem 8 is seekable and therefore synchronous.

is it Only at the host driver level ?

How to know the user application layer, The open file is block or non block?

For example in memwrite.c: how to know _write() function is block or non block

void allwrite(int fd, unsigned char *buf, int len) {
int sent = 0;
int rc;

while (sent < len) {
rc = _write(fd, buf + sent, len - sent);

if ((rc < 0) && (errno == EINTR))
continue;

if (rc < 0) {
perror("allwrite() failed to write");
exit(1);
}

if (rc == 0) {
fprintf(stderr, "Reached write EOF (?!)\n");
exit(1);
}

sent += rc;
}
}
andylgh
 
Posts: 24
Joined:

Re: Synchronous vs asynchronous

Postby support »

Hello,

Please refer to section 2 of the Xillybus host application programming guide for Windows, for an explanation on synchronous vs. asynchronous streams. It has nothing to do with blocking vs. non-blocking calls to _write() or _read().

http://xillybus.com/downloads/doc/xilly ... indows.pdf

I'm aware of the confusion given Microsoft's terminology for the same words, but they have several meanings in different contexts.

Regards,
Eli
support
 
Posts: 802
Joined:

Re: Synchronous vs asynchronous

Postby andylgh »

# coding: cp936
if __name__ == '__main__':


test_dat = bytearray()
for i in range(0,2048):
test_dat.append(i%256)
print(len(test_dat))


with open(r'\\.\xillybus_write_32','wb',buffering=-1) as wfd:
with open(r'\\.\xillybus_read_32','rb',buffering=-1) as rfd:
with open(r'test.bin','wb',buffering=-1) as fd:

#write file
w_num = wfd.write(test_dat)
wfd.flush()
print('w_num = %0d'%(w_num))

#read file
rdata = rfd.read(2048)
print(len(rdata))

w_num = fd.write(rdata)
fd.flush()

the bundle demo k7,there are 2KB fifo in FPGA write32 and read32 are loopback。the example above i first write 2KB to fpga_fifo and then read 2KB ,the test is success
now i change : rdata = rfd.read(2048) -》 rdata = rfd.read(2048 + 4) , the test is block at read()
According to my understanding, it should be able to return to 2048. It should not be blocked ?
andylgh
 
Posts: 24
Joined:

Re: Synchronous vs asynchronous

Postby support »

Hello,

You're actually asking a question on how Python handles I/O, so this isn't the right place.

As a side note -- in most cases where people have tried to test loopbacks with a single-thread program with a write followed by a read etc. it somehow failed because it got stuck in either of the two. I warmly recommend writing one program to write the data and another to read. Saves a lot of headache.

Also, in case you're heading towards making a bandwidth test soon, I suggest taking a look on this page:

http://xillybus.com/doc/bandwidth-guidelines

In particular, it begins with explaining why a bandwidth test shouldn't be done with a loopback.

Regards,
Eli
support
 
Posts: 802
Joined:

Re: Synchronous vs asynchronous

Postby andylgh »

i only want to test the stream function,not the speed
andylgh
 
Posts: 24
Joined:


Return to Xillybus

cron