Why does glReadPixels seem to return garbage? - linux

I'm trying to draw offscreen with OpenGL. For this I use EGL to initialize a pbuffer surface, and then draw to it, reading the results back with glReadPixels. But the following program gives me garbage on different (Mesa-based Intel on Linux) GPUs. Namely, on Atom N550 I get zeros, while on Xeon E3-1200 v3 I have 70 00 07 44 instead of the expected 40 80 bf ff.
With LIBGL_ALWAYS_SOFTWARE=1 environment variable set, I get the expected results. Also, if I comment out the line with eglBindAPI, I get good result on Xeon, but still zeros on Atom.
Here's my program:
#include <EGL/egl.h>
#include <GL/gl.h>
#include <iostream>
#include <iomanip>
#include <cstring>
int eglPrintError(std::string const& context)
{
const GLint error=eglGetError();
std::cerr << context << ": error 0x" << std::hex << int(error) << "\n";
return 1;
}
bool checkError(std::string const& funcName)
{
const GLenum error=glGetError();
if(error!=GL_NO_ERROR)
{
std::cerr << funcName << ": error 0x" << std::hex << int(error) << "\n";
return true;
}
return false;
}
constexpr int fbW=1, fbH=1;
bool initGL()
{
if(!eglBindAPI(EGL_OPENGL_API)) return !eglPrintError("eglBindAPI");
const EGLDisplay dpy=eglGetDisplay(EGL_DEFAULT_DISPLAY);
if(!dpy) return !eglPrintError("eglGetDisplay");
if(!eglInitialize(dpy,nullptr,nullptr)) return !eglPrintError("eglInitialize");
static const EGLint cfgAttribs[]={EGL_RENDERABLE_TYPE, EGL_OPENGL_BIT,
EGL_SURFACE_TYPE, EGL_PBUFFER_BIT,
EGL_RED_SIZE, 8,
EGL_GREEN_SIZE, 8,
EGL_BLUE_SIZE, 8,
EGL_ALPHA_SIZE, 8,
EGL_NONE};
EGLConfig cfg;
EGLint cfgCount;
if(!eglChooseConfig(dpy,cfgAttribs,&cfg,1,&cfgCount))
return !eglPrintError("eglChooseConfig");
if(cfgCount==0)
{
std::cerr << "Failed to get any usable EGL configs\n";
return false;
}
const EGLContext context=eglCreateContext(dpy,cfg,EGL_NO_CONTEXT,NULL);
if(!context) return !eglPrintError("eglCreateContext");
const EGLint surfaceAttribs[]={EGL_WIDTH, fbW, EGL_HEIGHT, fbH, EGL_NONE};
const EGLSurface surface=eglCreatePbufferSurface(dpy,cfg,surfaceAttribs);
if(!surface) return eglPrintError("eglCreatePbufferSurface");
if(!eglMakeCurrent(dpy,surface,surface,context))
return !eglPrintError("eglMakeCurrent");
return true;
}
int main(int argc, char** argv)
{
if(!initGL()) return 1;
glViewport(0,0,fbW,fbH);
glClearColor(0.25,0.5,0.75,1.);
glClear(GL_COLOR_BUFFER_BIT);
glFinish();
unsigned char data[4*fbW*fbH];
std::memset(data,0xb7,sizeof data); // to see unchanged values
glReadPixels(0,0,fbW,fbH,GL_RGBA,GL_UNSIGNED_BYTE,data);
if(checkError("glReadPixels")) return 1;
std::cout << "Data read: " << std::hex << std::setfill('0');
for(auto datum : data)
std::cout << std::setw(2) << +datum << " ";
std::cout << "\n";
return 0;
}
My question is, is there anything wrong in the above code which could lead to such behavior, or are my drivers just simply buggy?

Related

Forwarding rvalue into a thread where it is further forwarded to another function

I'm trying to create a deadline class which starts a thread in its constructor. Once the deadline (a time priod) occur within a thread, a function provided as argument should be executed.
The class'es contructor takes that function as a parameter, as well as arguments of that function.
I can provide parameter by value and by reference, but when providing rvalue as parameter, I get compile error, which I don't know how to solve.
#include <thread>
#include <atomic>
#include <future>
#include <iostream>
#include <iomanip>
namespace {
std::atomic<bool> running{true};
template <typename Period>
using duration = std::chrono::duration<int64_t, Period>;
template <typename PromiseType = void>
class deadline {
std::promise<PromiseType> stopper_{};
std::thread thread_{};
public:
deadline() = delete;
template <typename Period, typename Function, typename... Args>
explicit deadline(const duration<Period> &interval, Function&& func, Args&&... args) noexcept {
if (interval.count() > 0) {
std::future<PromiseType> future{ stopper_.get_future() };
using tuple = std::tuple<std::decay_t<Args>...>;
auto decay_copied = std::make_unique<tuple>(std::forward<Args>(args)...);
thread_ = std::thread(
[](std::future<PromiseType>&& future,
const duration<Period> &interval,
Function&& func, decltype(decay_copied)&& params) {
if (future.wait_for(interval) == std::future_status::timeout) {
std::apply(func, *params);
}
}, std::move(future), std::cref(interval), std::forward<Function>(func), std::move(decay_copied));
}
}
deadline(const deadline&) = delete;
deadline(deadline&&) = delete;
deadline& operator=(const deadline&) = delete;
deadline& operator=(deadline&&) = delete;
~deadline() {
stopper_.set_value();
if (thread_.joinable()) {
thread_.join();
}
}
};
}
auto main() -> int {
std::string test1{"test 1"};
std::string test2{"test 2"};
std::string test3{"test 3"};
std::string test4{"test 4"};
using namespace std::chrono_literals;
deadline dl(5s,
[](
std::string test_1
, const std::string& test_2
, std::string& test_3
, std::string&& test_4
) {
std::cout << test_1 << '\n';
std::cout << test_2 << '\n';
std::cout << test_3 << '\n';
test_3 += " modified";
std::cout << test_4 << '\n';
running = false;
}
, test1
, std::cref(test2)
, std::ref(test3)
, std::move(test4)
//, "test 4"
);
while (running.load()) {
std::this_thread::sleep_for(std::chrono::milliseconds(500));
}
std::cout << test1 << '\n';
std::cout << test2 << '\n';
std::cout << test3 << '\n';
std::cout << test4 << '\n';
return 0;
}
The code could be checked here also:
https://godbolt.org/z/fx3a4szoY
Any help is appreciated.
Thanks!
When providing a const value (also a rvalue), it works, it doesn't work if I use std::move.

How to use a specific DNS server/nameserver for name resolve queries in C++ boost::asio?

I would like to bypass system configured nameserver and use my own nameserver (or list of them) configured in application. I can do in nslookup in windows. How can I do it in C++ preferably by using boost::asio? I would like to avoid using std::system("nslookup ...> output.txt") and reading the file.
I cannot see where I can specify nameserver to use for lookup in boost::asio.
#include <boost/asio.hpp>
#include <string>
#include <iostream>
using namespace boost;
int main()
{
asio::io_service io_service;
asio::ip::tcp::resolver resolver(io_service);//how to pass specific nameserver?
asio::ip::tcp::resolver::iterator itr = resolver.resolve("bbc.co.uk","");
asio::ip::tcp::resolver::iterator end;
for (int i = 1; itr != end; itr++, i++)
std::cout << "hostname #" << i << ": " << itr->host_name() << " " << itr->endpoint() << '\n';
return 0;
}
You can't. The DNS resvoling is handled by the sockets api and you can't specify the DNS servers.
You will have to directly use either a OS specific API to resolve names like DnsQueryEx in win32 or using a library like LDNS.
/*
* DNSResolver.h
*
* Created on: Jan 12, 2023
* Author: marian
*/
#ifndef DNSRESOLVER_H_
#define DNSRESOLVER_H_
#include <cstdint>
#include <cstddef>
#include <vector>
#include <string_view>
#include <string>
#include <ostream>
#include <iostream>
#include <thread>
#include <chrono>
#include <boost/asio.hpp>
namespace tools
{
template <typename Func>
std::string_view find_first_success(const std::string_view& str,const Func& func, size_t pos0=0, char delim=',')
{
size_t pos=0;
do
{
pos=str.find_first_of(delim,pos0);
std::string_view sv=str.substr(pos0,pos-pos0);
if (func(sv))
return sv;
pos0=pos+1;
} while (pos!=std::string_view::npos);
return {};
}
template<int unused=0>
uint16_t be2uint16(const uint8_t* hi)
{
uint16_t hi1=*(hi+1);
return (((uint16_t)*hi)<<8)|hi1;
}
template<int unused=0>
uint32_t be2uint32(const uint8_t* hi)
{
uint32_t v3=*(hi+3);
uint32_t v2=*(hi+2);
uint32_t v1=*(hi+1);
uint32_t v0=*hi;
return (v0<<24)|(v1<<16)|(v2<<8)|v3;
}
template<int unused=0>
void out_IP4(std::ostream& os,uint32_t IP4)
{
os << (IP4>>24) << '.' << ((IP4>>16)&0xFF) << '.' << ((IP4>>8)&0xFF) << '.' << (IP4&0xFF);
}
}
namespace DNS
{
const char *rcode_msg[]={"No error","Format error","Server failure","Name error","Not implemented","Refused"};
const char *opcode_msg[]={"Standard query","Inverse query","Status"};
const char *type_msg[]={"0","A","NS","MD","MF","CNAME","SOA","MB","MG","MR","NULL","WKS","PTR","HINFO","MINFO","MX","TX"};
/*
A 1 a host address
NS 2 an authoritative name server
MD 3 a mail destination (Obsolete - use MX)
MF 4 a mail forwarder (Obsolete - use MX)
CNAME 5 the canonical name for an alias
SOA 6 marks the start of a zone of authority
MB 7 a mailbox domain name (EXPERIMENTAL)
MG 8 a mail group member (EXPERIMENTAL)
MR 9 a mail rename domain name (EXPERIMENTAL)
NULL 10 a null RR (EXPERIMENTAL)
WKS 11 a well known service description
PTR 12 a domain name pointer
HINFO 13 host information
MINFO 14 mailbox or mail list information
MX 15 mail exchange
TXT 16 text strings
*/
template<int unused=0>
void disp_type(std::ostream&os,uint16_t type)
{
os << (int)type;
if (type<=16)
os << " " << type_msg[type];
}
typedef boost::asio::detail::socket_option::integer<SOL_SOCKET, SO_RCVTIMEO> rcv_timeout_option; //somewhere in your headers to be used everywhere you need it
template<int unused=0>
class resolver
{
public:
resolver(const std::string_view& nameservers="127.0.0.53",boost::asio::chrono::milliseconds timeout=boost::asio::chrono::milliseconds{3000})
:nameservers_{nameservers},timeout_(timeout)
{}
std::vector<uint32_t> resolve(
const std::string_view& hostname,
std::string_view& nameserver_used,
std::vector<uint32_t>* pTTLs=nullptr)
{
std::vector<uint8_t> result;
result.resize(1024);
nameserver_used=resolve(hostname,result);
if (nameserver_used.empty())
return {};
std::string qname;
std::vector<uint32_t> IP4s;
int cnt=answer_A(result,qname,IP4s,pTTLs);
(void)cnt;
if (hostname!=qname)
return {};
return IP4s;
}
void resolve_ostream(
const std::string_view& hostname,
std::string_view& nameserver_used,
std::ostream& os)
{
std::vector<uint8_t> result;
result.resize(1024);
nameserver_used=resolve(hostname,result);
if (nameserver_used.empty())
return;
os << "nameserver:" << nameserver_used << std::endl;
disp_result(os,result);
}
private:
static void out_name(std::ostream&os,const std::vector<uint8_t>& result,uint16_t& pos)
{
uint8_t sz=result[pos++];
for (;sz;)
{
if (sz>=64)
{
if ((sz>>6)!=3)
throw "label size exceed 64 character and not pointer, see RFC 1035";
uint16_t offset=(((uint16_t)(sz&63))<<8)|result[pos++];
out_name(os,result,offset);//recursion, hope not too deep
return;
}
uint16_t epos=pos+sz;
while (pos<epos)
os << result[pos++];
sz=result[pos++];
if (sz)
os << ".";
}
}
static void disp_rr(std::ostream&os,const std::vector<uint8_t>& result,uint16_t& pos)
{
if (pos>=result.size())
return;
if (result[pos]==0)
return;
out_name(os,result,pos);
if (((size_t)pos+10)>=(size_t)result.size())
{
throw "wrong packet size1";
}
uint16_t rr_type=tools::be2uint16(&result[pos]);
os << " (" ;
os << "type:" ; disp_type(os,rr_type) ;
pos+=2;
os << " class:" << tools::be2uint16(&result[pos]) ;
pos+=2;
os << " TTL:" << tools::be2uint32(&result[pos]) << " secs) ";
pos+=4;
uint16_t len=tools::be2uint16(&result[pos]);
pos+=2;
if (pos+len>result.size())
throw "wrong packet size2";
if (rr_type==1)
{
if (len!=4)
{
throw "unknown data size";
}
os << (int)result[pos] << "." << (int)result[pos+1] << "." << (int)result[pos+2] << "." << (int)result[pos+3];
} else if (rr_type==2)
{
uint16_t pos2=pos;
out_name(os,result,pos2);
} else
{
throw "unimplement type";
}
pos+=len;
}
static void split_rr(const std::vector<uint8_t>& result,
uint16_t& pos,
std::string& name,
uint16_t& rr_type,
uint16_t& rr_class,
uint32_t& TTL,
uint32_t& IP4,
std::string* pname2=nullptr
)
{
if (pos>=result.size())
return;
if (result[pos]==0)
return;
std::stringstream ss;
out_name(ss,result,pos);
name=ss.str();
if (((size_t)pos+10)>=(size_t)result.size())
{
throw "wrong packet size1";
}
rr_type=tools::be2uint16(&result[pos]);
pos+=2;
rr_class=tools::be2uint16(&result[pos]) ;
pos+=2;
TTL=tools::be2uint32(&result[pos]);
pos+=4;
uint16_t len=tools::be2uint16(&result[pos]);
pos+=2;
if (pos+len>result.size())
throw "wrong packet size2";
if (rr_type==1)
{
if (len!=4)
{
throw "unknown data size";
}
IP4=tools::be2uint32(&result[pos]); //*(uint32_t*)&result[pos];
} else if (rr_type==2)
{
uint16_t pos2=pos;
std::stringstream ss;
out_name(ss,result,pos2);
if (pname2)
*pname2=ss.str();
} else
{
throw "unimplement type";
}
pos+=len;
}
static int check_A(const std::vector<uint8_t>& result)
{
if (result.size()<12)
return -1;
uint8_t response=(result[2]>>7);
if (response!=1)
return -2;
uint8_t rcode=result[3]&0xf;
if (rcode!=0)
return -3;
uint8_t opcode=(result[2]>>3)&0xF;
if (opcode!=0)
return -4;
return 0;
}
static int answer_A(const std::vector<uint8_t>& result,
std::string& q_name,
std::vector<uint32_t>& IP4s,
std::vector<uint32_t>* pTTLs=nullptr)
{
int ret=check_A(result);
if (ret<0)
return ret;
std::string name;
uint16_t q_type=0;
uint16_t q_class=0;
uint16_t rr_type=0;
uint16_t rr_class=0;
uint32_t TTL=0;
uint32_t IP4=0;
uint16_t qc=tools::be2uint16(&result[4]);
uint16_t pos=12;
for (int i=0;i<qc;i++)//qc =1 usually
{
std::stringstream ss;
out_name(ss,result,pos);
q_name=ss.str();
q_type=tools::be2uint16(&result[pos]);
pos+=2;
q_class=tools::be2uint16(&result[pos]);
pos+=2;
}
uint16_t ac=tools::be2uint16(&result[6]);
IP4s.resize(ac);
if (pTTLs)
pTTLs->resize(ac);
for (int i=0;i<ac;i++)
{
split_rr(result,pos,name,rr_type,rr_class,TTL,IP4);
if (name!=q_name)
break;
if (rr_type!=q_type)
break;
if (rr_class!=q_class)
break;
IP4s[i]=IP4;
if (pTTLs)
(*pTTLs)[i]=TTL;
}
return ac;
}
static void disp_result(std::ostream&os,const std::vector<uint8_t>& result)
{
if (result.size()<12)
return;
os << std::string((result[2]&0x80) ? "response" : "query") << std::endl;
os << "size=" << result.size() << " bytes" << std::endl;
uint8_t opcode=(result[2]>>3)&0xF;
os << "Opcode:" << (int)opcode;
if (opcode<=2)
os << " " << opcode_msg[opcode];
os << std::endl;
os << ((result[2]&1) ? "recursion asked" : "recursion NOT asked") << std::endl;
if (result[2]&2)
os << "response truncated" << std::endl;
if (result[2]&4)
os << "Authoritative answer" << std::endl;
os << ((result[3]&0x80) ? "recursion available" : "recursion NOT available") << std::endl;
uint8_t rcode=result[3]&0xf;
os << "rcode:" << (int)rcode;
if (rcode<=5)
os << " " << rcode_msg[rcode];
os << std::endl;
uint16_t qc=tools::be2uint16(&result[4]);
os << "Query Count:" << qc << std::endl;
uint16_t ac=tools::be2uint16(&result[6]);
os << "Answer Count:" << ac << std::endl;
uint16_t nc=tools::be2uint16(&result[8]);
os << "Authoritative Name Server Count:" << nc << std::endl;
uint16_t arc=tools::be2uint16(&result[10]);
os << "Additional resource records Count:" << arc << std::endl;
uint16_t pos=12;
os << "_____________" << std::endl;
os << "Query:" << std::endl;
for (int i=0;i<qc;i++)//qc =1 usually
{
out_name(os,result,pos);
os << " (";
os << "type:"; disp_type(os,tools::be2uint16(&result[pos]));
pos+=2;
os << " class:" << tools::be2uint16(&result[pos]) << " ) ";
pos+=2;
os << std::endl;
}
//answer
if (ac)
{
os << "_________________________" << std::endl;
os << "Answer:" << std::endl;
}
for (int i=0;i<ac;i++)
{
disp_rr(os,result,pos);
os << std::endl;
}
if (nc)
{
os << "_________________________" << std::endl;
os << "Authoritative nameservers:" << std::endl;
}
for (int i=0;i<nc;i++)
{
disp_rr(os,result,pos);
os << std::endl;
}
if (arc)
{
os << "_________________________" << std::endl;
os << "Additional resource records:" << std::endl;
}
for (int i=0;i<arc;i++)
{
disp_rr(os,result,pos);
os << std::endl;
}
}
size_t udp_request(const boost::asio::const_buffer& request,
const boost::asio::mutable_buffer& response,
const std::string_view& destination_ip,
const unsigned short port,
boost::system::error_code& ec)
{
using namespace boost;
asio::ip::udp::socket socket(io_context_);
auto remote = asio::ip::udp::endpoint(asio::ip::make_address(destination_ip), port);
socket.open(boost::asio::ip::udp::v4());
size_t sent=socket.send_to(request, remote);
if (request.size()!=sent)
return 0;
return receive_from(socket,response,timeout_,ec);
}
std::size_t receive_from(
boost::asio::ip::udp::socket& sock,
const boost::asio::mutable_buffer& buffer,
boost::asio::chrono::steady_clock::duration timeout,
boost::system::error_code& ec)
{
std::size_t length = 0;
sock.async_receive(boost::asio::buffer(buffer),
[&](const boost::system::error_code& ec1,std::size_t sz){
ec=ec1;
length=sz;
}
);
run(sock,timeout);
return length;
}
void run(boost::asio::ip::udp::socket& sock,boost::asio::chrono::steady_clock::duration timeout)
{
// Restart the io_context, as it may have been left in the "stopped" state
// by a previous operation.
io_context_.restart();
// Block until the asynchronous operation has completed, or timed out. If
// the pending asynchronous operation is a composed operation, the deadline
// applies to the entire operation, rather than individual operations on
// the socket.
io_context_.run_for(timeout);
// If the asynchronous operation completed successfully then the io_context
// would have been stopped due to running out of work. If it was not
// stopped, then the io_context::run_for call must have timed out.
if (!io_context_.stopped())
{
// Cancel the outstanding asynchronous operation.
sock.cancel();
// Run the io_context again until the operation completes.
io_context_.run();
}
}
static std::vector<uint8_t> make_dns_request(const std::string_view& hostname)
{
static uint16_t id=257;
id++;
std::vector<uint8_t> req;
uint16_t epos=12+1+hostname.size();
req.resize(epos+1+4);//plus null plus type (2 bytes) plus class (2 bytes)
*((uint16_t*)&req[0])=id;
req[2]=1;//recursive
req[3]=32;//??, linux has 32, but 0 works too
req[5]=1;//one query
req[epos]=0;//end of string
req[epos+2]=1;//type A - host address
req[epos+4]=1;//class INT
uint16_t cnt=0;
for(int i=hostname.size()-1;i>=0;i--)
{
req[13+i]=(hostname[i]=='.') ? cnt : hostname[i];
cnt=(hostname[i]=='.') ? 0 : cnt+1;
}
req[12]=cnt;
return req;
}
std::string_view resolve(const std::string_view& hostname,std::vector<uint8_t>& result)
{
auto myrequest=make_dns_request(hostname);
return tools::find_first_success(nameservers_,
[&](const std::string_view& nameserver)
{
size_t sz=0;
try
{
boost::system::error_code ec;
sz=udp_request(boost::asio::buffer(myrequest),
boost::asio::mutable_buffer(&result[0],result.size()),
nameserver,53,ec);
if (ec.value()!=boost::system::errc::success)
{
std::cerr << "skipping due to error code not a success : error code=" << ec.value() << " " << ec.message() << std::endl;
return false;
}
if (sz<35)//min packet , 12 header+3 name+4 type/class+ 16 (rr type 1)
{
std::cerr << "skipping because packet size < 35" << std::endl;
return false;
}
if ((myrequest[0]!=result[0])||(myrequest[1]!=result[1]))
{
std::cerr << "skipping because ID respond not matching ID request " << std::endl;
return false;
}
if (check_A(result)<0)
{
std::cerr << "skipping because packet check failed " << std::endl;
return false;
}
}
catch(std::exception& ex)
{
std::cerr << "Skipping because of exception:" << ex.what() << std::endl;
return false;
}
catch(...)
{
std::cerr << "Skipping because of unknown exception:" << std::endl;
return false;
}
result.resize(sz);
return true;
});
}
private:
boost::asio::io_context io_context_;
std::string nameservers_;
boost::asio::chrono::milliseconds timeout_;
};
}
#endif /* DNSRESOLVER_H_ */
/*
* main.cpp
*
* Created on: Jan 12, 2023
* Author: marian
*/
#include <iostream>
#include <vector>
#include "DNSResolver.h"
#define check_offset if (offset==argc) { print_error(); return 1;}
int main (int argc, char **argv) {
auto print_error=[&]()
{
std::cerr << "Usage: " << argv[0] << " [-v] [-t timeout_msecs] [--nottl] [--nons] <query> [csv_ip_nameserver_list] \n";
std::cerr << "Example: " << argv[0] << " bbc.co.uk 127.0.0.53,8.8.8.8" << std::endl;
std::cerr << "Example: " << argv[0] << " -v bbc.co.uk" << std::endl;
};
if (argc < 2) {
print_error();
return 1;
}
int offset=1;
bool verbose= (std::string(argv[offset])=="-v");
if (verbose)
offset++;
check_offset;
int millisecs=3000;
if (std::string(argv[offset])=="-t")
{
offset++;
check_offset;
millisecs=std::stoi(argv[offset++]);
}
check_offset;
bool ttl=true;
if (std::string(argv[offset])=="--nottl")
{
ttl=false;
offset++;
}
check_offset;
bool ns=true;
if (std::string(argv[offset])=="--nons")
{
ns=false;
offset++;
}
check_offset;
const char *dname=argv[offset];
const char *nameservers= (offset+1<argc) ? argv[offset+1] : "127.0.0.53,8.8.8.8";
if (ns)
std::cout << "nameservers:" << nameservers << std::endl;
DNS::resolver resolver{nameservers,boost::asio::chrono::milliseconds{millisecs}};
if (verbose)
{
std::string_view nameserver_used;
resolver.resolve_ostream(dname,nameserver_used,std::cout);
return 0;
}
std::vector<uint32_t> TTLs;
std::string_view nameserver_used;
std::vector<uint32_t> IP4s=resolver.resolve(dname,nameserver_used,&TTLs);
if (ns&&(!nameserver_used.empty()))
std::cout << "nameserver:" << nameserver_used << std::endl;
int n=IP4s.size();
if (!n)
{
std::cerr << "Error: No IPs found" << std::endl;
return 2;
}
for (int i=0;i<n;i++)
{
tools::out_IP4(std::cout,IP4s[i]);
if (ttl)
std::cout << " (TTL:" << TTLs[i] << " sec)" ;
std::cout << std::endl;
}
return 0;
}

ability to run a packaged task with std::bind function parameters in a seperate thread via template function

The question i have is the line where i indicate ERROR below gives error as std::thread constructor works out the function to be invoked and requires
the parameter as needed by the function signature.
Is there any way to solve this ? if i attempt to decode the function name and argument list from packaged_task then i cant use the get_future function over packaged task and need to add my own promise/future code to handle this.
#include<iostream>
#include<string>
#include<thread>
#include<future>
#include<functional>
using namespace std;
int sampleAddFunction(const int& a, const int& b)
{
int sum = a + b;
cout << "sum = " << sum << endl;
return(sum);
}
template<typename T> T asyncExecutor(std::packaged_task<T(T, T)>&& package)
{
std::future<T> result = package.get_future();
std::thread task_td(std::move(package)); // ERROR here as the std::thread identifies the function name from package and requires the params to be passed. How to handle this ?
task_td.join();
return(result.get());
}
int main(int argc, char* argv[])
{
// Executing via directly calling through main.
int testResult1 = sampleAddFunction(100, 200);
cout << "testResult1 = " << testResult1 << endl;
// Attempt to create a std::packaged_task and then run it in another thread.
std::packaged_task<int(int,int)> task(std::bind(sampleAddFunction, 10, 20));
std::future<int> result = task.get_future();
std::thread t(std::move(task), 100, 200); // 100 and 200 are dummy parameters.
t.join();
int testResult2=result.get();
cout << "testResult2 = " << testResult2 << endl;
// Attempt to run this in seperate thread and get results.
std::packaged_task<int(int,int)> task2(std::bind(sampleAddFunction, 15, 27));
int testResult3 = asyncExecutor<int>(std::move(task2), 100, 200);
cout << "testResult3 = " << testResult3 << endl;
}
This should work.
#include<iostream>
#include<string>
#include<thread>
#include<future>
#include<functional>
using namespace std;
int sampleAddFunction(int a, int b)
{
int sum = a + b;
cout << "sum = " << sum << endl;
return(sum);
}
template<typename R, typename F, typename... Ts>
R asyncExecutor(F&& package, Ts... args)
{
std::future<R> result = package.get_future();
std::thread task_td(std::move(package), args...);
task_td.join();
return(result.get());
}
int main(int argc, char* argv[])
{
std::packaged_task<int(int,int)> task2(sampleAddFunction);
int testResult3 = asyncExecutor<int>(std::move(task2), 15, 27);
cout << "testResult3 = " << testResult3 << endl;
}
You are constructing a binary packaged_task (std::packaged_task<int(int,int)>) from a nullary function, the result of your bind (std::function<int()>).
You should either not use bind, or have asyncExecutor accept a nullary packaged_task (std::packaged_task<T()>)
int sampleAddFunction(const int& a, const int& b)
{
int sum = a + b;
cout << "sum = " << sum << endl;
return(sum);
}
template<typename T, typename ... ARGS> T asyncExecutor(std::packaged_task<T(ARGS...)>&& package, ARGS ... args)
{
std::future<T> result = package.get_future();
std::thread task_td(std::move(package), args...);
task_td.join();
return(result.get());
}
int main(int argc, char* argv[])
{
std::packaged_task<int()> task(std::bind(sampleAddFunction, 10, 20));
int testResult = asyncExecutor(std::move(task));
cout << "testResult = " << testResult << endl;
std::packaged_task<int(int,int)> task2(sampleAddFunction);
int testResult2 = asyncExecutor(std::move(task2), 15, 27);
cout << "testResult2 = " << testResult2 << endl;
}

Looping back again to Start

now I'm Having problem in repeating the loop after it finished doing the first and i want to try it again without exiting the program? I've been using while loop to do it but still no joy. so i decided to do the if statement. But the Array only accept 4 strings then it exit. Any one who can help? TIA.
#include "stdafx.h"
#include <iostream>
#include <string>
#include <algorithm>
#include <vector>
using namespace std;
template <typename T>
void GetContents(T& Input);
template <typename T>
void DisplayContents(const T& Input);
int main()
{
int PASS = 0;
// To Display the unsorted and sorted Book Titles
std::vector<std::string> books;
GetContents(books);
std::cout << "\nMy original library (number of books: " << books.size() << "):\n\n";
DisplayContents(books);
std::sort(books.begin(), books.end());
std::cout << "\nMy sorted library (number of books: " << books.size() << "):\n\n";
DisplayContents(books);
std::cout << "Press 1 to try again, else to quit: ";
std::cin >> PASS;
std::cout << "\n";
if (PASS == 1)
{
GetContents(books);
}
else
{
return 0;
}
// to input All book titles
template <typename T>
void GetContents(T& Input)
{
const int MAX = 5;
string bookName;
std::cout << "Enter a Book Titles:\n> ";
for (int i = 0; i < MAX; i++)
{
std::getline(std::cin, bookName);
Input.push_back(bookName);
std::cout <<">";
}
}
//Display All input book titles
template <typename T>
void DisplayContents(const T& Input)
{
for (auto iElement : Input)
{
std::cout << iElement << '\n';
}
std::cout << '\n';
system("pause");
}

Why this small c++11 multi threaded program giving segmentation fault

why this program giving seg fault. I tried figuring out the issue using gdb, but no luck.
#include <iostream>
#include <condition_variable>
#include <thread>
#include <chrono>
using namespace std;
condition_variable cv;
mutex cv_m;
mutex m;
int count = 0;
#define COUNT_DONE 10
#define COUNT_HALT1 3
#define COUNT_HALT2 6
void functionCount1()
{
for(;;)
{
m.lock();
count++;
cout << "Counter value functioncount1: " << count << endl;
m.unlock();
if(count >= COUNT_DONE)
return;
}
}
void functionCount2()
{
for(;;)
{
m.lock();
count++;
cout << "Counter value functionCount2: " << count << endl;
m.unlock();
if(count >= COUNT_DONE) return;
}
}
int main()
{
thread t1(functionCount1), t2(functionCount2);
t1.join();
t2.join();
return 0;
}
Your program has undefined behavior: the accesses to count outside the mutex in functionCount1 and functionCount2 are data races. With the UB corrected, it seems fine:
#include <iostream>
#include <mutex>
#include <thread>
using namespace std;
mutex m;
int count = 0;
#define COUNT_DONE 10
void functionCount(const char* name)
{
for(;;)
{
m.lock();
auto c = ++count;
m.unlock();
cout << "Counter value " << name << ": " << c << endl;
if(c >= COUNT_DONE)
return;
}
}
int main()
{
thread t1(functionCount, "functionCount1"), t2(functionCount, "functionCount2");
t1.join();
t2.join();
}
or if you want to be "clever" and confuse your code reviewers:
void functionCount(const char* name)
{
for(;;)
{
auto c = (std::lock_guard<std::mutex>(m), count++);
cout << "Counter value " << name << ": " << c << endl;
if(c >= count_done)
break;
}
}

Resources