minghust / ford Goto Github PK

The example briefly illustrates the situation of one computing node. I want to know how to modify the configuration if FORD is used on multiple computing nodes. At the same time I want to know if the code needs to be modified?

输出 Operation not supported
// 0x80003 = MAP_SHARED_VALIDATE | MAP_SYNC. Some old kernel does not have linux/mman.h: MAP_SHARED_VALIDATE, MAP_SYNC
hash_buffer = (char*)mmap(NULL, hash_buf_size, PROT_READ | PROT_WRITE, MAP_SHARED_VALIDATE | MAP_SYNC, pm_file_fd, 0);
if (hash_buffer == MAP_FAILED) {
printf("mmap failed, %s\n", strerror(errno));
assert(0);
}
请问使用fsdax代码应该做出怎么样的修改？

Hello, I'm using the rlib library in ford for cross-node experiments, but after the qp connection, the RDMA read always gives an error:
[qp_impl.hpp:140] poll till completion error: 12 transport retry counter exceeded
So, has Ford encountered the problems? And if so, hope to know how to solve it. Thanks!