I am writing a kernel module that hooks some system calls (e.g. tcp_send() ) using jprobes and sends some information to the userspace using netlink sockets.
netlink_unicast(nlsk, skb, pid, MSG_DONTWAIT);
my callback call is:
void nl_recv(struct sk_buff *skb) {
struct nlmsghdr *nlh;
if (skb == NULL) {
return;
}
nlh = (struct nlmsghdr *) skb->data;
pid = nlh->nlmsg_pid;
debug(KERN_NOTICE "Kernel Module: Received pid from %u\n", pid);
}
I'd like to pause the execution of my kernel module after every send. relaunch on receive.
I have tried using completions and wait queues, but it seems that they push the session into a GPF.
Related
I'm developing device base on ESP32 module that have a UDP socket open only to receive broadcast packets on one port (7890 to be exact). The problem is that the data losses are high - around 90%. My test setup is:
ESP32 - connected to WiFi network with open UDP receing task (code belowe)
PC connected to the same netwer via LAN with UDP terminal set to brodacast to remote: 192.168.10.255:7890
Mobile phone connected to WiFi with UDP terminal set to brodacast to remote: 192.168.10.255:7890
When I send something from PC or mobile phone there is no data lossage between Mobile phone and PC but ESP32 receive around 10% of data that I transmit from both of senders. If I change from multicast to unicast on PC or Phone to send data to ESP32, it work without problem.
I know that UDP does not guarantee the delivery but 10% efficiency seems for me to be super low, especially when it seems that there is no problem with busy network because PC and mobile received the data all the time.
Do you have any suggestion to the code or some setting that can be changed in menu config ?
At the moment my application have only two tasks:
WiFi Task that after connection is just waiting for event
UDP Task that the code is below
Update 04.07.2018 (13:15)
Problem disappear when I don't initialize bluetooth. Sorry that I didn't mention previously about BT being initialized but I kept me initializing function from my normal program that have a lot more tasks (BT included) and totally forgot about this myself.
Anyway - do you think that there is some issue with sharing the resource or is it some physical interference ? I'm using ESP32-DevKitC that is on the breadboard, so no additional shielding is present.
#define PORT_NUMBER 7890
#define BUFLEN 100
void udp_task(void *pvParameter)
{
struct sockaddr_in clientAddress;
struct sockaddr_in serverAddress;
struct sockaddr_in si_other;
unsigned int slen = sizeof(si_other);
unsigned int recv_len;
char buf[BUFLEN];
int sock;
printf("UDP Task: Opening..\n");
int ret;
ret = UDP_List_Open(&clientAddress, &serverAddress, &sock);
if(ret == 0)
{
printf("UDP Task: Open\n");
}
else
{
printf("UDP Task: Can't open\n");
}
while(1)
{
memset(buf,0,100);
if ((recv_len = recvfrom(sock, buf, 100, 0, (struct sockaddr *) &si_other, &slen)) == -1)
{
printf("UDP error\n");
break;
}
sendto(sock, buf, recv_len, 0, (struct sockaddr *)&si_other, sizeof(si_other));
printf("UDP Task: Received packet from %s:%d\n", inet_ntoa(si_other.sin_addr), ntohs(si_other.sin_port));
printf("UDP Task: Data: %s -- %d\n" , buf, recv_len);
}
while(1)
{
vTaskDelay(100 / portTICK_RATE_MS);
}
}
int UDP_List_Open(struct sockaddr_in* clientAddress, struct sockaddr_in* serverAddress, int* sock)
{
// Create a socket that we will listen upon.
*sock = socket(AF_INET, SOCK_DGRAM, 0);
if (*sock < 0)
{
printf("UDP List Open: Socket error\n");
return 1;
}
// Bind our server socket to a port.
serverAddress->sin_family = AF_INET;
serverAddress->sin_addr.s_addr = htonl(INADDR_ANY);
serverAddress->sin_port = htons(PORT_NUMBER);
int rc = bind(*sock, serverAddress, sizeof(*serverAddress));
if (rc < 0)
{
printf("UDP List Open: Bind error\n");
return 2;
}
return 0;
}
Even though UDP is considered fire and forget, (unlike TCP), unicast UDP through WiFi is reliable because reliability is built into the WiFi protocol. But this can work for Unicast only because there is one known recipient. Multicast UDP is unreliable because there are no checks and retries.
I had the same problem when I was trying to use multicast UDP with the ESP8266. It caused me to dig deeper into the issue. In the end I use UDP multicast for discovery but then switch to Unicast UDP for subsequent transfers.
See Multicast Wifi Problem Statement
https://tools.ietf.org/id/draft-mcbride-mboned-wifi-mcast-problem-statement-01.html
I am writing a kernel module that should receive messages from user-space and send response back via socket.
When program and module are on the same machine and I use IP 127.0.0.1, everything works fine. But when I try it on different machines and use real network IP, something like 192.168.3.146 it works only in one way.
I receive message from user-space, but I can not receive it from kernel. I use sock_sendmsg function for sending message from kernel and it's not return any error. Also I am not get any messages from firewall that something is came up from another machine, from kernel module.
Here were similar questions and examples, but they were not useful enough for me or examples were used too old kernel version.For skeleton I used this one,from UDP sockets: http://people.ee.ethz.ch/~arkeller/linux/multi/kernel_user_space_howto-3.html. Any help?
Kernel module code for sending:
void send_data(unsigned char *data)
{
if(!IS_ERR_OR_NULL(data))
{
int ret;
mm_segment_t oldfs;
struct msghdr message;
struct iovec ioVector;
struct sockaddr_in sendAddr;
sendAddr.sin_family = AF_INET;
sendAddr.sin_addr.s_addr = INADDR_ANY;
//sendAddr.sin_addr.s_addr = in_aton("192.168.1.75");
//here I get port from sk_buff structure that I received.
sendAddr.sin_port = *((unsigned short*)skBuffer->data);
memset(&message, 0, sizeof(message));
message.msg_name = &sendAddr;
message.msg_namelen = sizeof(sendAddr);
/* send the message back */
ioVector.iov_base = data;
ioVector.iov_len = strlen(data);
message.msg_iov = &ioVector;
message.msg_iovlen = 1;
message.msg_control = NULL;
message.msg_controllen = 0;
oldfs = get_fs();
set_fs(KERNEL_DS);
ret = sock_sendmsg(sendSocket, &message, strlen(data));
set_fs(oldfs);
}
}
I found an alternative solution, using netpoll sockets. It is more easier than sockets, I used before and it works. The answer and proper code is here, on another StackOverflow question.
I have multiple threads who have a socket open with a client application each. Each of those threads receive an instruction from a main thread to send commands to the client (commands could be run test, stop test, terminate session, exit....). Those threads are generic, they just have a socket per client and just send a command when the main thread asks it to.
The client could exit or crash, or network could be bad.
I have been trying to see how to figure out that my TCP session has ended per client. Two solutions that I have found that seem appropriate here.
1) Implement my own heartbeat system
2) Use keepAlive using setsockopt.
I have tried 2) as it sounds faster to implement, but I am not sure of one thing: Will SO_KEEPALIVE generate a SIGPIPE when connection is interrupted please? I saw that it should be the case but never received a SIGPIPE.
This is how my code looks:
void setKeepAlive(int sockfd) {
int optval;
optval = 1;
setsockopt(sockfd, SOL_SOCKET, SO_KEEPALIVE, &optval, sizeof(optval));
optval = 1;
setsockopt(sockfd, SOL_TCP, TCP_KEEPIDLE, &optval, sizeof(optval));
optval = 1;
setsockopt(sockfd, SOL_TCP, TCP_KEEPCNT, &optval, sizeof(optval));
optval = 1;
setsockopt(sockfd, SOL_TCP, TCP_KEEPINTVL, &optval, sizeof(optval));
}
And my code that accepts connection is as follows:
for (mNumberConnectedClients = 0; mNumberConnectedClients < maxConnections; ++mNumberConnectedClients) {
clientSocket = accept(sockfd, (struct sockaddr *) &client_addr, &clientLength);
// set KeepAlive on socket
setKeepAlive(clientSocket);
pthread_create(&mThread_pool[mNumberConnectedClients], NULL, controlClient, (void*) &clientSocket);
}
signal(SIGPIPE, test);
....
And the test function:
void test(int n) {
printf("Socket broken: %d\n", n);
}
test() never gets triggered. Is it my understanding that is wrong please? I am not sure if SIGPIPE gets generated or not. Thanks a lot.
If a keep-alive fails, the connection will simply be invalidated by the OS, and any subsequent read/write operations on that socket will fail with an appropriate error code. You need to make sure your reading/writing code is handling errors so it can close the socket, if it is not already doing so.
I've ran into a problem with a simple TCP Client implemented using select.
The problem is that,at the second printf it only displays before it gets to the connect() function then waits for user input. Does connect block the rest of the program until i send something? (The TCP server is also implemented using select but i didn't find anything wrong with it)
I've searched on the web and couldn't find a cause or maybe i didn't search for the right thing..
#include <includes.h>
int main()
{
int sfd;
fd_set rset;
char buff[1024]=" ";
char playerName[20]="";
int nameSet=0;
struct sockaddr_in server;
sfd= socket(AF_INET,SOCK_STREAM,0);
if(sfd<0)
{ printf("socket not created\n"); return 0; }
bzero(&server,sizeof(struct sockaddr_in));
server.sin_family=AF_INET;
server.sin_port=htons(2020);
inet_aton("127.0.0.1",&server.sin_addr);
//here is the problem after %d which calls the connect() function
printf("Conexion returned:%d \n Name:",connect(sfd,(struct sockaddr *)&server,sizeof(server)));
for(;;)
{
bzero(buff,1024);
FD_ZERO(&rset);
FD_SET(0,&rset);
FD_SET(sfd,&rset);
if(select(sfd+1,&rset,NULL,NULL,NULL)<0)
{
printf("con-lost!\n");
break;
}
if(FD_ISSET(0,&rset))
{
printf("Talk: \n");
scanf("%s",buff);
if(nameSet==0)
{
strcpy(playerName,buff);
nameSet=1;
printf("Hi:%s\n",playerName);
}
if(write(sfd,buff,strlen(buff)+10)<0)
{
break;
}
}
if(FD_ISSET(sfd,&rset)>0)
{
if(read(sfd,buff,1024)<=0)
{
printf("con is off!\n");
break;
}
printf("msg rcd %s\n",buff);
}
} //endfor
close(sfd);
return 0;
} //endmain
The connect function, on a blocking socket, blocks until the connect operation succeeds or fails.
You should be warned that using select with a blocking socket, which is what your program does, does not ensure that your program will not block. When you get a select hit, that does not guarantee that a future operation will not block.
strlen(buff)+10
What's the reasoning behind the +10?
Here's a minimal test case , I tried to start a local domain server , with QThread , so the UI shouldn't stuck. But when it's starting , i saw Listening output from qDebug() , but the widgets added from form editor totally disappeared , everything went slow (e.g resizing the window) , if i remove thread.start() , the UI shows up and functions well.
#include "mainwindow.h"
#include "ui_mainwindow.h"
MainWindow::MainWindow(QWidget *parent) :
QMainWindow(parent),
ui(new Ui::MainWindow)
{
ui->setupUi(this);
connect (&thread , SIGNAL(started()) , SLOT(setupServer()));
thread.start();
}
MainWindow::~MainWindow()
{
delete ui;
}
void MainWindow::setupServer()
{
struct sockaddr_un address;
int socket_fd, connection_fd;
socklen_t address_length;
// create socket .. and create socket file ..
// bind ...
// listen ..
qDebug() << "Listening ..";
while((connection_fd = ::accept(socket_fd,
(struct sockaddr *) &address,
&address_length)) > -1)
{
qDebug() << "Got an connection.";
::close (connection_fd);
}
// close socket and remove the socket file
}
The accept(2) syscall is by default blocking. You should take advantage of the multiplexing syscall poll(2) or select(2) used by the QApplication's exec event loop.
See this question and use the QtNetwork module.
First of all, your setupServer is always called in gui thread. Deliver your own QThread based class and reimplement run method. Put your "setupServer" code inside