slaclab / rogue Goto Github PK

View Code? Open in Web Editor NEW

38.0 27.0 17.0 23.9 MB

SLAC Python Based Hardware Abstraction & Data Acquisition System

Home Page: https://slaclab.github.io/rogue/

License: Other

C++ 58.44% Shell 0.17% Python 32.82% C 0.26% CMake 2.72% Dockerfile 0.09% Jupyter Notebook 5.51%

daq tid-air-es

rogue's People

Contributors

Stargazers

Watchers

Forkers

abunimeh nwhitehorn codacy-badger mengqingwu bhill-slac slacrherbst tubbz-alt pbutti mwittgen yeq71 mcb64 iicc20 mskoufis mfkiwl jjl772 sebv15 konosubakonoakua

rogue's Issues

Add timestamps to logging functions

I am looking throug failure logs for Simons Observatory's SMuRF software, and am finding that it would be really nice to have timestamps on more log messages. For instance, here is a sample of the logs from an old run:

ERROR:pyrogue.Variable.RemoteVariable.AMCc.FpgaTopLevel.AppTop.AppCore.SysgenCryo.Base[7].CryoChannels.etaI[28]:int too big to convert
Traceback (most recent call last):
  File "/usr/local/src/rogue/python/pyrogue/_Variable.py", line 324, in set
    self._block.set(self, value)
  File "/usr/local/src/rogue/python/pyrogue/_Block.py", line 362, in set
    ba = var._base.toBytes(value)
  File "/usr/local/src/rogue/python/pyrogue/_Model.py", line 128, in toBytes
    ba = value.to_bytes(byteCount(self.bitSize), self.endianness, signed=True)
OverflowError: int too big to convert
ERROR:pyrogue.Variable.RemoteVariable.AMCc.FpgaTopLevel.AppTop.AppCore.SysgenCryo.Base[7].CryoChannels.etaI[28]:Error setting value '-50008' to variable 'AMCc.FpgaTopLevel.AppTop.AppCore.SysgenCryo.          Base[7].CryoChannels.etaI[28]' with type Int16. Exception=int too big to convert
1718215269.675174:pyrogue.epicsV3.Value: Error setting value from epics: smurf_server_s7:AMCc:FpgaTopLevel:AppTop:AppCore:SysgenCryo:Base[7]:CryoChannels:etaMagArray

/tmp/fw/rogue_MicrowaveMuxBpEthGen2_v1.1.0.zip/python/CryoDet/DspCoreLib/CryoDetCmbHcd/_SerialEtaScan.py:88: RuntimeWarning: divide by zero encountered in scalar divide
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AppTop.DaqMuxV2[1].DbgInputValid at address 0xb000002c Timeout waiting for register transaction message response'
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AmcCarrierCore.SwRssiServer[1].RetransmitCnt at address 0xa01104c Timeout waiting for register transaction message response'
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AppTop.DaqMuxV2[1].TriggerSwStatus at address 0xb0000004 Timeout waiting for register transaction message response'
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AmcCarrierCore.EM22xx.READ_TEMPERATURE_1 at address 0xd000234 Timeout waiting for register transaction message response'
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AmcCarrierCore.AxiVersion.UpTimeCnt at address 0x000008 Timeout waiting for register transaction message response'
ERROR:pyrogue.PollQueue:'Memory Error for AMCc.FpgaTopLevel.AppTop.AppTopJesd[0].JesdRx.RawData[5] at address 0xc0000154 Timeout waiting for register transaction message response'

There are some timestamps here, but the Timeout waiting for register transaction message response errors have none. Can we add timestamps to all rogue log messages?

Bad access to shared pointer when creating UDP client and server

Constructor of rogue::protocols::udp::Client and rogue::protocols::udp::Server spawns a runThread. Inside this thread a shared_from_this() is called, that returns a shared pointer of the constructed object.
In the case where the object is already constructed, e.g. shared pointer is not weak it works fine.
However, if the spawned thread runThread tries to generate a shared pointer before the object is constructed it throws std::bad_weak_ptr.

Now, it seems that runThread of Client and Server has a usleep(1000); to overcome this problem. However, this is not always sufficient. The runThread shall wait until the object is created, i.e. shared pointer was created out of it.

For example with use of mutex

//! Class creation
rpu::ServerPtr rpu::Server::create(uint16_t port, bool jumbo)
{
   rpu::ServerPtr r = std::make_shared<rpu::Server>(port, jumbo);
   r->constructed();
   return (r);
}

//! Set object is constructed
void rpu::Server::constructed()
{
   mutex_.unlock();
}

//! Run thread
void rpu::Server::runThread()
{
   ris::BufferPtr buff;
   ris::FramePtr frame;
   fd_set fds;
   int32_t res;
   struct timeval tout;
   struct sockaddr_in tmpAddr;
   uint32_t tmpLen;
   uint32_t avail;

   udpLog_->logThreadId();
   std::lock_guard<std::recursive_mutex> lk(mutex_); // wait for construction to end

   // Preallocate frame
   frame = ris::Pool::acceptReq(maxPayload(), false);

   while (threadEn_)
   {

Link to boost tar in /Readme_boost.txt

wget https://sourceforge.net/projects/boost/files/boost/1.64.0/boost_1.64_0.tar.gz

Should be:

wget https://sourceforge.net/projects/boost/files/boost/1.64.0/boost_1_64_0.tar.gz

Missing rogue/hardware/data/DataMap.h

Inside DataMap.cpp
#include <rogue/hardware/data/DataMap.h>

File DataMap.h is missing from the repo.

UDP Server and Client internal thread may cause std::bad_weak_ptr exception to be thrown

shared_from_this() may be called before the object and shared_ptr are fully initialized. This causes std::bad_weak_ptr exception and crashes the application.

Trace Server:
rpu::Server::runThread()
-- ris::Pool::acceptReq(maxPayload(),false)
---- allocBuffer(size,&frSize)
------ shared_from_this()

Trace Client:
rpu::Client::runThread()
-- ris::Pool::acceptReq(maxPayload(),false)
---- allocBuffer(size,&frSize)
------ shared_from_this()

Proposed fix:

auto weakPtr = weak_from_this();
auto sharedPtr = weakPtr.lock();

while (!sharedPtr) {
    std::this_thread::yield();
    sharedPtr = weakPtr.lock();
}

Memory leak in ris::Pool

Buffers that are stored on the buffer queue (dataQ_) are never freed:
https://github.com/slaclab/rogue/blob/main/src/rogue/interfaces/stream/Pool.cpp#L87

Proposed fix:

ris::Pool::~Pool()
{
   while (!dataQ_.empty())
   {
      free(dataQ_.front());
      dataQ_.pop();
   }
}

RSSI Controller std::thread memory leak

In rpr::Controller::start a thread is created using new, but never freed:
https://github.com/slaclab/rogue/blob/main/src/rogue/protocols/rssi/Controller.cpp#L146

Proposed fix:
Add delete thread_ to rpr::Controller::stop or better avoid creating a thread using new.

slaclab / rogue Goto Github PK

rogue's People

Contributors

Stargazers

Watchers

Forkers

rogue's Issues

Add timestamps to logging functions

Bad access to shared pointer when creating UDP client and server

Link to boost tar in /Readme_boost.txt

Missing rogue/hardware/data/DataMap.h

UDP Server and Client internal thread may cause std::bad_weak_ptr exception to be thrown

Memory leak in ris::Pool

RSSI Controller std::thread memory leak

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent