Feed for discussion QNX Momentics Community Support in project Community.

post122707: Re: Disconnection issue : "Invalid target settings" error on QNX Target (VirtualBox)

Albrecht Uhlmann — Mon, 12 Jan 2026 11:32:50 GMT

Hi Intissar,

for me it looks like you are mixing two independent features.

In the IDE, if you chose "QNX Virtual Machine Target", it is supposed to actually create a brand-new local VM based on a template and your settings, and launch it in the background.

However, you already created a full VM by using the installation CD. So basically, what you need to do is:

- Make sure your VM has an IP address assigned.

- Make sure in your VM qconn runs (default port is 8000, but can be changed via cmd line)

- In the IDE, create a target same way as you would for a physical board, and enter the IP address of your already existing and running VM

Hope this helps,

-Albrecht

post122706: Disconnection issue : "Invalid target settings" error on QNX Target (VirtualBox)

Intissar SALBI — Sat, 10 Jan 2026 17:45:45 GMT

I'm currently setting up my QNX development environment and I’m running into a bit of a problem. For context, here are the steps I followed to create my "Target":

On VirtualBox: I created a VM named "Target" using the QNX® Software Development Platform 6.5.0 ISO. I set the OS type to "Linux" and the version to "Ubuntu (64-bit)".

On the IDE: To link everything, I clicked the "ON" button in the launch bar, selected "New Launch Target," and chose "QNX Virtual Machine Target."

Settings: I named the target "Target," set the platform to "vbox," and the architecture to "x86_64." I left the IP address blank for auto-detection.

The Problem:

The connection works initially, but after a short period, the VM suddenly disconnects from the IDE. When I try to relaunch or ping the target from the IDE, I get the following error: "Invalid target settings."

Has anyone encountered this before? Could it be related to the automatic IP detection or some kind of conflict between VirtualBox and the IDE?

Thanks in advance for your help!

post122637: Using vector and list classed from STL in C++ program.

Richard Ellingworth — Wed, 27 Nov 2024 14:25:26 GMT

I am converting some code originally written for QNX 6.3.2 to run on QNX 7. The code makes use of the vector class, but whenever <vector> is included I get the following warnings:

In file included from c:\qnx700\target\qnx7\usr\include\c++\v1\__bit_reference:17:0,

                 from c:\qnx700\target\qnx7\usr\include\c++\v1\vector:266,

                 from ../../common/core/src/SubsystemTranslator.h:62,

                 from ControlsManager.h:47,

                 from TXLXControls.h:51,

                 from TXLXControls.cpp:88:

c:\qnx700\target\qnx7\usr\include\c++\v1\__undef_min_max:16:2: error: #warning : macro min is incompatible with C++. #undefing min [-Werror=cpp]

#warning: macro min is incompatible with C++.  #undefing min

  ^

c:\qnx700\target\qnx7\usr\include\c++\v1\__undef_min_max:26:2: error: #warning : macro max is incompatible with C++. #undefing max [-Werror=cpp]

#warning: macro max is incompatible with C++.  #undefing max

If I comment out the #include <vector> line then the warning goes away. I really want to continue using the vector class, mainly to be consistent with other projects, but I cannot do this as long as this warning keeps appearing. We are treating all warnings as errors.

The odd thing is that almost identical code in another project compiles correctly without so much as a warning.

So is there a way to continue using the vector (and list) classes in a C++ project using the QNX 7 compiler? What do I need to change?

post122619: QConn trace thread problem

Nesredin Mahmud — Wed, 09 Oct 2024 13:50:47 GMT

Hey!

I am unable to collect tracelog from the IDE. The error says

'kernel Trace' has encountered a problem.

QConn trace thread problem.

The ide is successfully connected to the target. I am also able to collect memory related trace events.

Cheers,

/nas

post122522: Conflicts with previous declarations due to #include_next

Umberto Cancedda — Fri, 26 Apr 2024 15:16:52 GMT

Dear community,

I'm working on a QNX 7.1 based project and I'm facing compile errors regarding conflicting declarations. This is somehow related to #include_next statements. From the code I see, that there are a lot of compiler switches controlling the includes but I really cannot figure out which one to set or unset. In my special case it is about the function "long long int std::abs(long long int)" which conflicts with "long long int abs(long long int)". It definitely has to do with stdlib.h and cstdlib.

To be more precise, here is a piece of the build log showing the error:

n file included from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/cstdlib:77,

                 from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/bits/stl_algo.h:59,

                 from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/algorithm:62,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/common_std_inc.hpp:52,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/common_inc.hpp:50,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/xcp_adaptive_service_params.cpp:48:

/home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/bits/std_abs.h:56:15: error: 'long int std::abs(long int)' conflicts with a previous declaration

   abs(long __i) { return __builtin_labs(__i); }

               ^

In file included from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/cstdlib:75,

                 from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/bits/stl_algo.h:59,

                 from /home/developer/qnx710/target/qnx7/usr/include/c++/8.3.0/algorithm:62,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/common_std_inc.hpp:52,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/common_inc.hpp:50,

                 from /home/developer/vrte/project/mcgateway/mcgateway/common/xcp_adaptive_service_params.cpp:48:

/home/developer/qnx710/target/qnx7/usr/include/c++/v1/stdlib.h:111:44: note: previous declaration 'long int abs(long int)'

inline _LIBCPP_INLINE_VISIBILITY long      abs(long __x) _NOEXCEPT {return  labs(__x);}

I'd be grateful for any hint on how to solve these conflicts.

Best regards

Cheers

Umberto

post122316: Re: Error creating a QNX Virtual Machine

Dave Nickerson(deleted) — Wed, 07 Jun 2023 17:30:41 GMT

The missing files are fine.  Just indicates that you have not installed a few optional packages but you don't need them, so it is safe to ignore.

I am guessing that you are using a board that required a special version of our PCI server that does not play well with the VMWare PCI controller.

Reverting that package is possible but then you won't be able to build images for you hardware platform.

Easiest work-around is to create a second (or additional) SDP installation where you choose "Ultraconservative" as the installation type.  Install no updates at that point and see if you can successfully build a VMware image.

If yes, then the issue is as I outlined above.  If you then want to install updates to this new installation, carefully update packages but avoid any PCI server updates or updates with dependancies on the PCI server updates.  The issue is very specific to the board that you are using from the board vendor which requires the updated PCI server and the build files that are being used to create the VMware images.

post122274: Re: Error with creating virtual machine from QNX Momentics IDE

Trevor Dunn(deleted) — Fri, 19 May 2023 19:41:32 GMT

Hello,

Thank you for the fix, everything seems to be working! I appreciate it a lot.

post122273: Re: Error with creating virtual machine from QNX Momentics IDE

Roger Maclean — Fri, 19 May 2023 16:56:36 GMT

This appears to be a VMware bug.

A bit of Googling suggests that people have worked around it by disabling 3D acceleration for the VM. You should be able to do this by configuring the settings for that VM in VMware or alternatively you can probably do this by editing local/vmware_files/vmware.vmx and changing mks.enable3d to "FALSE".

At one point there was a reason that 3D acceleration had to be enabled though hopefully this is no longer the case.

If this works, you can fix it more permanently by changing vmware.vmx in $QNX_HOST/../../common/mkqnximage/vmware.

post122272: Error with creating virtual machine from QNX Momentics IDE

Trevor Dunn(deleted) — Fri, 19 May 2023 16:39:45 GMT

I'm having an issue being able to run my virtual machine using the momentics IDE. The instructions given from my program are the following :

a.Launch the Momentics IDE and create a new Workspace to hold the VM target.

Workspace folder. for example: "C:\Users\john\qnx710\vmimages"

b. To access your target system from the IDE and run programs on it, you have to create a target

connection. Click in the on: box in the launch bar at the top of theIDE, and select New Launch Target.

c. Select QNX Virtual Machine Target in the list of target types. This tells the IDEto set up a target system by

generating a VM that runs QNX Neutrino. Then, click Next.

d. In the New QNX Virtual Machine Target window, configure the settings as needed and click Finish.

Attached are an image of the error I get when launching the VMware, the settings I used when creating my target location in the momentics IDE and my log file.

post122252: Error creating a QNX Virtual Machine

Lorenzo Novara — Thu, 11 May 2023 13:32:46 GMT

I'm trying to create a QNX Virtual Machine with Momentics, but I get some warnings about missing files, then the virtual machine fails to boot.

I'm using QNX SDP 7.1 BuildID 472 and Momentics IDE version 7.1.2.v202203081633

I tried from a Windows and a Linux host, with the same results.

Target type: vmware

Architecture: x86_64

These are the warning:

Generating system partition . . .

Warning: Host file 'usr/sbin/gns' missing.

Warning: Host file 'lib/dll/lsm-qnet.so' missing.

Warning: Host file 'usr/lib/libcurl.so.10' missing.

Warning: Host file 'usr/lib/libcares.so.4' missing.

Generating data partition . . .

While booting, I get all sorts of errors (see attached screenshots), like:

Unable to access /dev/pci

Unable to access /dev/hd0

...

Reference images downloaded from QNX Software Center works fine.

How can I fix this?

post122171: Re: Unbale to create the QNX Virtual Machine

Roger Maclean — Thu, 16 Mar 2023 12:28:50 GMT

When you run the VM VirtualBox should pop up a window that will show console output. If there were errors encountered when booting the VM, they will hopefully appear there.

If this was the first time you've used VirtualBox and you haven't already done so, you will have to configure a network for it (for whatever reason it doesn't have one by default). This is described in the documentation for mkqnximage (the tool Momentics uses to build and run VMs).

post122170: Unbale to create the QNX Virtual Machine

Piraisudan Mahendiran — Thu, 16 Mar 2023 12:08:01 GMT

Hello Team,

I need a QNX virtual machine with the internet access in it.

As mentioned in the below article, I'm trying to create the "New QNX Virtual Machine Target" using QNX Momentics in the windows machine and facing the "Error occurred while finding IP address for vm " error (as attached in the screenshot).

https://www.qnx.com/developers/docs/7.1/#com.qnx.doc.ide.userguide/topic/creating_qnx_vm.html

Can you please help me to resolve this issue?

post122169: Re: Thread with priority 255 running too much time

SHENGFA ZHANG — Thu, 16 Mar 2023 05:13:19 GMT

Is there anyone give me some tips please?

post122165: Thread with priority 255 running too much time

SHENGFA ZHANG — Tue, 14 Mar 2023 04:11:53 GMT

We have a thread with priority 255 loops forever to receive pulse event and post semaphore to notify work thread. The looping pseudo code lists below.

while (true) {

    MsgReceivePulse();

    ret = seg_getvalue(sem, value);

    if (!((0 == ret) && (value > 0))) {

        sem_post(sem);

    }

}

The running time of the above code is abort 10us. But we found the running time was nearly 3ms sometimes, when we used QNX traceevent to catch some events. See the attachment.

This thread's events list below:

696441, 613ms 438us, 0x265209E82, 7, safety Patie#ty_2TimTh, Process and Thread, Running, policy_name FIFO pid 9424945 tid 34 priority 255 policy 1 partition 7 sched_flags 0

696442, 613ms 438us, 0x26520A3C0, 7, safety Patie#ty_2TimTh, Communication, Receive Pulse, scoid 0x10005 pid 9424945 process safety

696443, 613ms 438us, 0x26520A4FA, 7, safety Patie#ty_2TimTh, Kernel Calls, MsgReceivePulsev (64) Exit, rcvid 0x0 rmsg0 0x0 rmsg1 0x0 rmsg2 0x0 info->nd 0 info->srcnd 0 info->pid 0 info->tid 0 info->chid 0 info->scoid 0 info->msglen 0 info->srcmsglen 0 info->dstmsglen 0 info->priority 0 info->flags 0 info->reserved 0

696885, 613ms 762us, 0x265306EF0, 7, safety Patie#ty_2TimTh, System, IPI (64), ipicmd 0x20 CHECK_INTR ipi 0x257f781317 tid 34 pid 9424945 process safety

696903, 613ms 777us, 0x2653128D0, 7, safety Patie#ty_2TimTh, Process and Thread, Running, policy_name FIFO pid 9424945 tid 34 priority 255 policy 1 partition 7 sched_flags 0

698114, 614ms 762us, 0x265612411, 7, safety Patie#ty_2TimTh, System, IPI (64), ipicmd 0x20 CHECK_INTR ipi 0x257f781317 tid 34 pid 9424945 process safety

698134, 614ms 781us, 0x265620D94, 7, safety Patie#ty_2TimTh, Process and Thread, Running, policy_name FIFO pid 9424945 tid 34 priority 255 policy 1 partition 7 sched_flags 0

698691, 615ms 181us, 0x26575868B, 7, safety Patie#ty_2TimTh, Kernel Calls, SyncSemPost (64) Enter, sync_ptr 0x1e3b814a7c count 0 owner 0xfffffffc

698693, 615ms 183us, 0x26575A3DE, 7, safety Patie#ty_2TimTh, Kernel Calls, SyncSemPost Exit, ret_val 0x0

699423, 615ms 763us, 0x26591DC0B, 7, safety Patie#ty_2TimTh, System, IPI (64), ipicmd 0x20 CHECK_INTR ipi 0x257f7810ed tid 34 pid 9424945 process safety

699451, 615ms 786us, 0x26592FE48, 7, safety Patie#ty_2TimTh, Process and Thread, Running, policy_name FIFO pid 9424945 tid 34 priority 255 policy 1 partition 7 sched_flags 0

699846, 616ms 99us, 0x265A23AE1, 7, safety Patie#ty_2TimTh, Kernel Calls, MsgReceivePulsev (64) Enter, chid 0x9 rparts 1

From the attachment, we found this thread is always running and there are many Control events, such as.

696454, 613ms 446us, 0x265210A7F, 1, , Control Events, Buffer, sequence_number 6216609 event_count 0

696455, 613ms 447us, 0x2652112FE, 1, , Control Events, Time, msb 0x174d lsb 0x652112fe

696459, 613ms 451us, 0x265214063, 9, , Control Events, Time, msb 0x174d lsb 0x65214063

Our system has 12 CPU core, and this thread is running on CPU 7/8/9 , the aps partition is 20% (less than 3/12=25%).

So why this thread runs so long? Is it because traceevent? Or do aps partition exceed 20%?

Thanks very much.

post122123: Re: Ctype Module not Found in QNX Virtual Machine with x_86 Architecture

ShAILESH AGRAHARI — Thu, 19 Jan 2023 08:43:40 GMT

even i was able to copy..

but same error still there...

post122122: Re: Ctype Module not Found in QNX Virtual Machine with x_86 Architecture

ShAILESH AGRAHARI — Tue, 17 Jan 2023 09:58:41 GMT

when trying to copy to /system/lib

Getting error "No space Left"

and not getting this location:

local/snippets/system_files.custom

Thanks!

post122121: Re: Ctype Module not Found in QNX Virtual Machine with x_86 Architecture

Roger Maclean — Mon, 16 Jan 2023 14:35:54 GMT

I suspect you are missing libffi.so.6

You can either just copy it to /system/lib on your VM or add:

lib/libffi.so.6=usr/lib/libffi.so.6

to the file local/snippets/system_files.custom so it'll be there next time you build it.

The way to work out what it's missing is to set the environment variable DL_DEBUG to libs when you run python. This will show all the shared objects that are being loaded, or in this case not being loaded.

post122120: Re: Ctype Module not Found in QNX Virtual Machine with x_86 Architecture

ShAILESH AGRAHARI — Mon, 16 Jan 2023 14:06:38 GMT

Anyone please help on this..

post122109: Ctype Module not Found in QNX Virtual Machine with x_86 Architecture

ShAILESH AGRAHARI — Mon, 09 Jan 2023 10:35:11 GMT

Hi All,

Getting the attached error while trying to run some Robot Framework-based script.

I have created the VM Target using QNX Momentics for x_86 Architecture.

Thanks!

post122102: Error occurs while finding IP adress for QNX..

ShAILESH AGRAHARI — Fri, 06 Jan 2023 08:44:55 GMT

Hi All,

While creating a new QNX Qemu VM on my laptop in window environment from Momentics IDE, i am getting below error:

Attached Screen shot

post122019: System Summary view shows little information

Stephen Barton(deleted) — Thu, 08 Dec 2022 09:42:07 GMT

Hi,

When connected to my target J721e evm board I can run tasks and debug but the system information displayed is very limited. e.g. in the attached System Summary view most of the data is 0 or missing.

Is this a limitation of the target? Am I missing some configuration?

Regards

Stephen

post122012: CAN loopback testing

Vic One — Tue, 29 Nov 2022 03:54:35 GMT

Hi, I would like to test the CAN related API in SDP by virtual CAN mode.

But I can't find the CAN driver that support loopback mode in QNX or Internet.

Is there anyone having idea about it?

Thanks.

post122011: Is it possible to do API hook?

Vic One — Tue, 29 Nov 2022 03:52:01 GMT

Hi, I would like to implmenet an application that will monitor the system API events that is calling by other processes.

Does anyone know it is possible to hook the API event in QNX? And any resource for refernce? Thanks.

post121999: Interconnect applications deployed on different VMs (guests) in qnx hypervisor

Ayush Kumar(deleted) — Fri, 18 Nov 2022 12:09:04 GMT

CAUTION - This email is from an external source. Please be cautious with links and attachments. (go/taginfo)

Dear all,

I want to interconnect two applications deployed on two different guests

and also connect with other application deployed on different target board

(rcar s4).

Can anyone help me with the process to do this and the code sample to see

how the application talks with each other?

Thanks & Best Regards

Ayush

--

ayush

post121877: Detecting Input Devices

Bradley Peggram — Tue, 19 Jul 2022 04:33:32 GMT

Hi there,

every so often when I boot up QNX 6.3.0 into Photon, the display comes up with a window displaying: Detecting input devices, please wait...

This pop up will stay up forever and essentially halts the system. What causes this?...as I'm not sure what it is trying to detect...I only have a MS keyboard and MS mouse plugged in. Once this message appears once...it appears every subsequent boot and it's impossible to get rid of so I just re-build the drive. I'm obviously sick of doing this so why is this coming up, how do I prevent it from coming up and if it does come up...how do I fix it?

My input trap is: devi-hid kbd mouse

Thanks.

post121837: No able to use lockf for /dev/shmem/text_lock file

Darshan B(deleted) — Wed, 29 Jun 2022 15:23:38 GMT

I am trying to lock one file which is opened using shm_open function.

Does anyone did this earlier?

I am getting ENOSYS error while using function lockf().

Thanks in advance.

post121807: Re: Getting USB Microphone to work with QNX

T Heo — Tue, 24 May 2022 22:33:45 GMT

Correction, I've been able to get the projects to compile, but NOT able to play an already existing wave file or write out a recorded wave file that seems to play anything from the wave.c and waverec.c example files.

post121806: Getting USB Microphone to work with QNX

T Heo — Tue, 24 May 2022 22:18:22 GMT

Hello all,

Its my first time trying to get something like this working on a QNX system. I have a bunch of USB microphones, listed below, which I am trying to capture audio data from and then play.

1.USB Lavalier Microphone, MAONO 192KHZ/24BIT Plug & Play Omnidirectional Lapel Shirt Collar Clip on Mic for PC, Computer, Mac, Laptop, YouTube, Skype, Recording, Podcasting, Gaming, AU-UL10 from Amazon.com: USB Lavalier Microphone, MAONO 192KHZ/24BIT Plug & Play Omnidirectional Lapel Shirt Collar Clip on Mic for PC, Computer, Mac, Laptop, YouTube, Skype, Recording, Podcasting, Gaming, AU-UL10 : Musical Instruments

2.USB Lavalier Lapel Microphone, ZAFFIRO Clip-on USB Computer Microphone Plug & Play Omnidirectional Mic for PC, Laptop, Mac, PS4. Perfect for Video Recording,Skype, Streaming, Podcasting from Amazon.com: USB Lavalier Lapel Microphone, ZAFFIRO Clip-on USB Computer Microphone Plug & Play Omnidirectional Mic for PC, Laptop, Mac, PS4. Perfect for Video Recording,Skype, Streaming, Podcasting : Electronics

3.USB Lavalier Lapel Microphone, Fifine Clip-on Cardioid Condenser Computer Mic Plug and Play USB Microphone with Sound Card for PC and Mac-K053 from Amazon.com: USB Lavalier Lapel Microphone, Fifine Clip-on Cardioid Condenser Computer Mic Plug and Play USB Microphone with Sound Card for PC and Mac-K053 : Electronics

4.Movo M1 USB Lavalier Lapel Clip-on Omnidirectional Computer Microphone for Laptop, PC and Mac, Perfect Podcasting, Gaming, Streaming and Desktop Mic (20-Foot Cord) from Amazon.com: Movo M1 USB Lavalier Lapel Clip-on Omnidirectional Computer Microphone for Laptop, PC and Mac, Perfect Podcasting, Gaming, Streaming and Desktop Mic (20-Foot Cord) : Musical Instruments

5.Antlion Audio ModMic USB Attachable Noise-Cancelling Microphone with Mute Switch Compatible with Mac, Windows PC, Playstation 4, and More from Amazon.com: Antlion Audio ModMic USB Attachable Noise-Cancelling Microphone with Mute Switch Compatible with Mac, Windows PC, Playstation 4, and More : Electronics

6.Sennheiser XS Lav USB-C Omnidirectional Lavalier Microphone with 2M Cable USB-C Connector from Sennheiser XS Lav USB-C Omnidirectional Lavalier Microphone with 2M Cable USB-C Connector | Sweetwater

I've got the project for wave.c working to the point of being able to call "wave -P 90 FILE.wav" and it will play the wav file just fine. However when I try to use waverec.c to create a file and then play it using the "wave -P FILE.wav" command, i only get silence.

I think it has to do with these USB microphones maybe not being supported or a driver not being properly run but I'm not quite sure. Any help would be appreciated!

post121751: Re: How to make an installer which can be installed on QNX target

Albrecht Uhlmann — Wed, 09 Mar 2022 13:55:29 GMT

http://www.qnx.com/developers/docs/7.1/#com.qnx.doc.neutrino.utilities/topic/m/mkqnx6fsimg.html

The BSP USer's Guides provide some infos about how to install. It is highly board-specific, therefore no generic procedure exists.

Regards,

Albrecht

post121750: Re: How to make an installer which can be installed on QNX target

Darshan B(deleted) — Wed, 09 Mar 2022 12:14:20 GMT

Thank you Albrecht for your reply. but how can I use the mkqnx6fsimg?

Do you have any example code?

Please guide me to create one installer file.

post121749: Re: How to make an installer which can be installed on QNX target

Albrecht Uhlmann — Wed, 09 Mar 2022 08:34:08 GMT

If you want to create an archive that can be installed into an already existing and running target you could use "tar" on the host. You can use the "append" option to build the archive incrementally, and there are limited options to set owner and file permissions.

A more complete and integrated solution is to use "mkqnx6fsimg" which creates an image of a complete filesystem binary image that can be extracted with a binary writer like dd to a disk partition of equivalent size.

Rehards,

Al

post121748: How to make an installer which can be installed on QNX target

Darshan B(deleted) — Wed, 09 Mar 2022 04:53:51 GMT

Hi Team,

I am new to QNX platform. I am working on one project which has many binary files, like A, B C and they have dependency with library X.so Y.so.

I want to pack it to one file, which I can install it on QNX Target at specific location.

I am sure, there will be some way to do it in qnx. Can someone help me to point out the way.

Thanks in advance.

post121721: Re: I2C slave driver solution for Beaglebone Black

Albrecht Uhlmann — Thu, 10 Feb 2022 12:49:50 GMT

Fuzzing around with GPIO tends to never get stable and reliable. I would try to implement a slave mode driver directly for the on-board I2C controller, based on the master driver source in the BSP. Some routines like setting baud rate and so on can probably be reused.

The only thing you need to add entirely new is the communication interface to userland. A small resource manager should do.

You can't re-use the one from master driver because it is buried in libi2c-master.a, and kind of tailored to the master role.

Regards,

Albrecht

post121720: I2C slave driver solution for Beaglebone Black

huseyin goktas(deleted) — Wed, 09 Feb 2022 21:06:10 GMT

Hello,

We are trying to make communication between two beaglebones using I2C. Ethernet and other solutions are not used because of design limitations. I2C slave mode is not supported by qnx bsp. Thus I implemented software i2c solution using GPIO and interrupts. But using the hardware solution would be more reliable and fast. First, I want to ask is it possible to use I2C slave mode by modifying BSP. If possible how should I start?

post121658: access to floating licence via VPN

David Haufe — Mon, 22 Nov 2021 10:30:22 GMT

Hello,

I have a question about which protocol or which port the QDE Momentics Version: 7.0.4.v202004031918 (qnx 6.5 and 6.6 SDKs) are communicate with the licence-server, cause I've got some trouble with my VPN connection in homeoffice.

Connected to our company network over vpn:

I can start the QDE but as soon as I want to build a project I see in the console output that the qcc is called accordingly but nothing else happens.

When I call the qcc via the console with verbose output the behavior is the same, no output occurs.

In the corporate network itself:

After starting the build process, the compiler is called accordingly and the message :

Checkout succeeded: *****/*************

License file: *****@servername.de.company.net

License Server: *****@servername.de.company.net

The corresponding license server is reachable via both connection paths and the connection paths are displayed correctly via tracert, my assumption is that a corresponding port is not released.

Or is that an https connection and i will need a specific cert. ?

Kind regards

post121647: PHAB Widget white Border

surya neelapu(deleted) — Fri, 29 Oct 2021 05:37:05 GMT

After pressing widget button then a while colour border is highlighted after release button while colour border is cleared.

Please suggest us how to remove this highlighted border. I do not require this highlighted border after pressing button.

code:

    PtSetArg( &arg[0], Pt_ARG_FILL_COLOR, QMMI_COLOR_DBLUE1, 0);

    PtSetArg( &arg[1], Pt_ARG_TOP_BORDER_COLOR, QMMI_COLOR_BORDER_BUT_TOP, 0);

    PtSetArg( &arg[2], Pt_ARG_BOT_BORDER_COLOR, QMMI_COLOR_BORDER_BUT_BOT, 0);

    PtSetArg( &arg[3], Pt_ARG_COLOR, QMMI_COLOR_TEXT_NORMAL, 0);

    PtSetArg( &arg[4], Pt_ARG_FLAGS, Pt_SELECTABLE,Pt_SELECTABLE|Pt_TOGGLE|Pt_SET);

    PtSetArg( &arg[5], Pt_ARG_TEXT_STRING,"A c k" ,0);

Please see the attached screen shot for reference.

post121595: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

Albrecht Uhlmann — Tue, 14 Sep 2021 15:14:56 GMT

I am a bit of a loss. Originally you asked a question about having a shared memory area between two VMs. Now you ask why a camera is not present???

As to the latter, camera drivers are often highly board-specific. I can imagine that some dynamic detection algorithm runs at boot time and instantiates the camera drivers only if it finds Intel hardware (or it is looking for a feature in CPU or PCI space that is not present on AMD machine).

Please be more specific about your problem, and create two threads if there are two problems.

-Al

post121594: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Tue, 14 Sep 2021 12:32:09 GMT

@Community team,

Did anyone face any similar process creation issues with AMD processor, that works with Intel machines??

Pls let me know.

Regs,

Mohan

post121591: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Mon, 13 Sep 2021 04:37:01 GMT

Hi Albrecht,

Sorry for the typo in the name, cant find the edit option.

post121590: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Mon, 13 Sep 2021 04:34:47 GMT

Hi Alberich,

As i am new to QNX ,and i am using client's QNX image, i cant comment much on this.

While my team members are trying the same, ones with AMD processor machine cant run the software.

All i found is that a file named "camera" of type: file , is missing in my AMD machine, where as the members with INTEL machine can see it being created in runtime.

and open(camera) function gives the error for me.

Sorry if this info is less. Can u help me with this info ?

Thanks,

S Mohan

post121588: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

Albrecht Uhlmann — Fri, 10 Sep 2021 14:43:05 GMT

I think one of the reasons why VMs are used is to isolate things completely. So Shared Memory is a security issue. I have been searching around a bit, it doesn't seem to be a standard feature.

Mayb the QNX Hypervisor can do it, protected by SMMU.

What are you trying to achieve?

Regards,

Albrecht

post121587: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Thu, 09 Sep 2021 09:34:43 GMT

Any help in this regard would be great

post121586: Re: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Wed, 08 Sep 2021 11:52:00 GMT

Update:

I am able to ping between windows host and the VM's and also between the VM's

post121585: Unable to communicate between two QNX VMWare instance in AMD machines (working in Intel)

mohan S(deleted) — Wed, 08 Sep 2021 08:00:33 GMT

Hi Team,

I am new to QNX development.

I am trying to set up the environment given from my client, where two seperate VMWare instances uses a shared memory , simulating two processors.

When i set up these two machines, I am struck  in a place where the communication is not working between the two instance.

I found that the error is make by open() , where a "file" must be created in runtime, but is it is not being created in runtime.

I have tried various trial and error methods to fix this. as my other team members with INTEL processor can able to run test the simulation fine.

OS: QNX Neutrino 6.3.2

VMWare Workstation player 16

My device:

AMD Ryzen 5 pro

Windows 10  - 64bit

Any help or any more info  in this regard would be great. Let mw know , if more input from my side is required to help me.

Thanks,

S Mohan.

post121410: Compiling PCL example

Madhav Cham(deleted) — Fri, 07 May 2021 05:05:41 GMT

I am trying compile and execute PCL example on QNX having SDP 7.0 evaluation version

please find the source code below

#include <iostream>

#include <pcl/io/pcd_io.h>

#include <pcl/point_types.h>

int main (int argc, char** argv)

{

  pcl::PointCloud<pcl::PointXYZ> cloud;

  // Fill in the cloud data

  cloud.width    = 5;

  cloud.height   = 1;

  cloud.is_dense = false;

  cloud.points.resize (cloud.width * cloud.height);

  for (auto& point: cloud)

  {

    point.x = 1024 * rand () / (RAND_MAX + 1.0f);

    point.y = 1024 * rand () / (RAND_MAX + 1.0f);

    point.z = 1024 * rand () / (RAND_MAX + 1.0f);

  }

  pcl::io::savePCDFileASCII ("test_pcd.pcd", cloud);

  std::cerr << "Saved " << cloud.size () << " data points to test_pcd.pcd." << std::endl;

  for (const auto& point: cloud)

    std::cerr << "    " << point.x << " " << point.y << " " << point.z << std::endl;

  return (0);

}

command to compile

q++ -Vgcc_ntox86_64 -std=c++14 -I /home/madhav/pcl/io/include -I /home/madhav/pcl/common/include -I /home/madhav/boost_1_76_0 -I /usr/include/eigen3 -I /home/madhav/pcl/2d/include  pcl.cpp -o pcl

getting following error

In file included from /home/madhav/boost_1_76_0/boost/interprocess/sync/file_lock.hpp:26:0,

                 from /home/madhav/pcl/io/include/pcl/io/boost.h:66,

                 from /home/madhav/pcl/io/include/pcl/io/file_io.h:43,

                 from /home/madhav/pcl/io/include/pcl/io/pcd_io.h:45,

                 from pcl.cpp:2:

/home/madhav/boost_1_76_0/boost/interprocess/detail/os_thread_functions.hpp:72:7: error: #error "No high resolution steady clock in your system, please provide a patch"

#     error "No high resolution steady clock in your system, please provide a patch"

       ^

/home/madhav/boost_1_76_0/boost/interprocess/detail/os_thread_functions.hpp:319:8: error: #error "Can't obtain system tick value for your system, please provide a patch"

       #error "Can't obtain system tick value for your system, please provide a patch"

        ^

In file included from /home/madhav/pcl/io/include/pcl/io/impl/pcd_io.hpp:50:0,

                 from /home/madhav/pcl/io/include/pcl/io/pcd_io.h:788,

                 from pcl.cpp:2:

/home/madhav/pcl/io/include/pcl/io/low_level_io.h:61:24: fatal error: sys/fcntl.h: No such file or directory

compilation terminated.

cc: /home/madhav/QNX_setup/qnx700_2/host/linux/x86_64/usr/lib/gcc/x86_64-pc-nto-qnx7.0.0/5.4.0/cc1plus error 1

thanks,

Madhav

post121409: Compiling PCL example

Madhav Cham(deleted) — Fri, 07 May 2021 05:04:25 GMT

post121407: Third party library Compilation on QNX

Madhav Cham(deleted) — Fri, 07 May 2021 04:47:41 GMT

Would like to compile and build Flann and execute sample example on target QNX

Using SDP 7.0 evaluation license

Thanks,

Madhav

post121406: Third party library Compilation on QNX

Madhav Cham(deleted) — Fri, 07 May 2021 04:47:04 GMT

Would like to compile and build Flann and execute sample example on target QNX

Using SDP 7.0 evaluation license

Thanks,

Madhav

post121405: Third party library Compilation on QNX

Madhav Cham(deleted) — Fri, 07 May 2021 04:43:42 GMT

Would like to compile and build Flann and execute sample example on target QNX

Using SDP 7.0 evaluation license

Thanks,

Madhav

post121404: Re: Third party library support on QNX - cross compilation of dependent library

Albrecht Uhlmann — Wed, 28 Apr 2021 07:20:57 GMT

Hi Madhav,

1. yes, unless it is provided by other parties (like QNX themselves or others). Knowing the POSIX compliance of QNX, this is often a task that can be done with very moderate effort and has the advantage that you can select exactly the source version you want and do the "configure" part yourself.

2. somtimes, in the QNX Software Center is a section on "Ported Third-Party Software"

3. No, binaries built for Linux do not run on QNX (at least not on 6.x where we actually ran into this issue - maybe possible on 7.x but I really doubt it. Maybe QNX guys can clarify that).

Regards,

Albrecht

post121403: Third party library support on QNX - cross compilation of dependent library

Madhav Cham(deleted) — Wed, 28 Apr 2021 07:08:54 GMT

Hi,

Would like to know how to compile/install third party library like Point cloud

library and its dependant c++ libraries like boost,eigen3,flann etc.

1. Do we have to build/compile each dependant library using qcc compiler

2. Does qnx provide prebuilt shared libraries for boost,eigen3,flann etc which

we we can directly use and link with our application

3. Or Dependant posix complaint libraries build on linux_x86 can be used to

link with an application intended to run on qnx platform

Thanks,

Madhav

post121390: Re: List file descriptors

Sorin Lazareanu(deleted) — Thu, 22 Apr 2021 15:42:45 GMT

Thanks!

post121389: Re: List file descriptors

Veer Beede — Thu, 22 Apr 2021 15:26:11 GMT

How about "pidin fds"?

post121388: List file descriptors

Sorin Lazareanu(deleted) — Thu, 22 Apr 2021 15:23:42 GMT

How do I console list file descriptors in use by a certain process?

post121379: Building Dotnet-Runtime on QNX 7.1

Nico Denicu(deleted) — Wed, 14 Apr 2021 06:51:33 GMT

I would like to know if dotnet-runtime (https://github.com/dotnet/runtime/) could be ported on QNX 7.1.

I managed to build it on QNX with some exceptions. One of the exceptions that it is really important to get a working version, is the use of "sigaltstack" functionality.

For the moment, the sigaltstack is not supported and I would like to know how we can deal with?

It is all about: https://github.com/dotnet/runtime/blob/main/src/coreclr/pal/src/thread/thread.cpp

Thanks & Regards,

Nicu

post121293: QNX 6.6.0 SP1, x86 platform self hosted usb driver disables usb host's port

Janusz Ruszel — Thu, 25 Feb 2021 15:40:34 GMT

Randomly the usb hub port is disabled and the device (usb camera) is not accessible until the system is re-powered. The only message in the system log is:

Jan 21 12:36:50    2    12     0 hub_transfer_done: hub (busno 1, devno 4, vid 0x0424, did 0x2517) - port 1 disabled

Jan 21 12:36:50    2    12     0 hub_transfer_done: hub (busno 1, devno 4, vid 0x0424, did 0x2517) - port 1 disabled

Jan 21 12:36:50    2    12     0 CLASS_PortDisabled: port 1

The usb server is started as:

       io-usb -v -dehci -dxhci

The system usb configuration is attached.

Any idea or root cause why the driver is disabling the hub port ?

Regards,

Janusz

post121201: Re: readlink() error while reading /proc/self/exefile on QNX

Roger Maclean — Tue, 02 Feb 2021 17:29:59 GMT

Well if you're not fussy about line length and not too hung up on using the preprocessor you can do anything in C in one line :-)

Not really sure what you're after (or why you're after it). There's no single function that does everything.

post121200: Re: readlink() error while reading /proc/self/exefile on QNX

Darshan B(deleted) — Tue, 02 Feb 2021 17:13:43 GMT

Thanks for your suggestion Roger.

Reading the file worked for me.

But do you know any one line command to read the path of executable?

Because to read the file I need to open the file and read it, then close it. Bit length process.

Just curious to know if we can do it any other way.

Thanks in advance.

post121198: Re: readlink() error while reading /proc/self/exefile on QNX

Roger Maclean — Tue, 02 Feb 2021 15:41:17 GMT

Read the file itself. It's just a text file, not a symlink.

post121197: readlink() error while reading /proc/self/exefile on QNX

Darshan B(deleted) — Tue, 02 Feb 2021 15:29:37 GMT

I am working on QNX platform, in which I need to get the path of executable which is running.

I have wrote a small peice of code, which is returning always -1:

#include <stdio.h>

#include <stdlib.h>

#include <unistd.h>

#include <errno.h>

#include <string.h>

extern int errno;

int main( int argc, char** argv )

{

    char buf[512] = {0};

    const char mypath[100] = "/proc/self/exefile";

    errno = 0;

    printf("The value readlink:: %d %s\n",readlink(mypath, buf, 512 ), strerror( errno ));

    return( 0 );

}

When I ran above code then I get following output:

The value readlink:: -1 No such file or directory

Am I missing anything? What needs to be done to get my current exe path in QNX?

post120876: How to setup DNS server in QNX7.0

Darshan B(deleted) — Mon, 03 Aug 2020 08:25:14 GMT

I am working on gRPC  for QNX ARM 7.0 platform.

I am able to build the gRPC code for QNX platform. While running these binaries I am facing an issue related to DNS server.

D0101 23:54:27.828777594       1 grpc_ares_ev_driver.cc:313]           (c-ares resolver) request:1036ce90 readable on c-ares fd: 3

D0101 23:54:27.829777489       1 grpc_ares_wrapper.cc:243]             (c-ares resolver) request:1036ce90 on_hostbyname_done_locked host=localhost C-ares status is not ARES_SUCCESS: Could not contact DNS servers

D0101 23:54:27.829777489       1 grpc_ares_ev_driver.cc:98]            (c-ares resolver) request:1036ce90 Ref ev_driver 10361460

D0101 23:54:27.829777489       1 grpc_ares_ev_driver.cc:407]           (c-ares resolver) request:1036ce90 notify read on: c-ares fd: 3

D0101 23:54:27.829777489       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.830777436       1 grpc_ares_ev_driver.cc:313]           (c-ares resolver) request:1036ce90 readable on c-ares fd: 3

D0101 23:54:27.830777436       1 grpc_ares_ev_driver.cc:98]            (c-ares resolver) request:1036ce90 Ref ev_driver 10361460

D0101 23:54:27.830777436       1 grpc_ares_ev_driver.cc:407]           (c-ares resolver) request:1036ce90 notify read on: c-ares fd: 3

D0101 23:54:27.830777436       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.831777175       1 grpc_ares_ev_driver.cc:313]           (c-ares resolver) request:1036ce90 readable on c-ares fd: 3

D0101 23:54:27.831777175       1 grpc_ares_wrapper.cc:243]             (c-ares resolver) request:1036ce90 on_hostbyname_done_locked host=localhost C-ares status is not ARES_SUCCESS: Could not contact DNS servers

I0101 23:54:27.831777175       1 timer_generic.cc:468]                 TIMER 103614a8: CANCEL pending=true

I0101 23:54:27.831777175       1 timer_generic.cc:468]                 TIMER 103614f8: CANCEL pending=true

D0101 23:54:27.831777175       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.832777227       1 grpc_ares_ev_driver.cc:119]           (c-ares resolver) request:1036ce90 delete fd: c-ares fd: 3

D0101 23:54:27.832777227       1 grpc_ares_ev_driver.cc:449]           (c-ares resolver) request:1036ce90 ev driver stop working

D0101 23:54:27.832777227       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.832777227       1 grpc_ares_ev_driver.cc:247]           (c-ares resolver) request:1036ce90 ev_driver=10361460 on_timeout_locked. driver->shutting_down=1. err="Cancelled"

D0101 23:54:27.833776966       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.833776966       1 grpc_ares_ev_driver.cc:276]           (c-ares resolver) request:1036ce90 ev_driver=10361460 on_ares_backup_poll_alarm_locked. driver->shutting_down=1. err="Cancelled"

D0101 23:54:27.833776966       1 grpc_ares_ev_driver.cc:105]           (c-ares resolver) request:1036ce90 Unref ev_driver 10361460

D0101 23:54:27.833776966       1 grpc_ares_ev_driver.cc:108]           (c-ares resolver) request:1036ce90 destroy ev_driver 10361460

D0101 23:54:27.833776966       1 dns_resolver_ares.cc:367]             (c-ares resolver) resolver:10361560 dns resolution failed: {"created":"@86067.831777175","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244,"referenced_errors":[{"created":"@86067.829777489","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244}]}

I0101 23:54:27.834777226       1 resolving_lb_policy.cc:254]           resolving_lb=1036d240: resolver transient failure: {"created":"@86067.834777226","description":"DNS resolution failed","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/dns_resolver_ares.cc","file_line":370,"grpc_status":14,"referenced_errors":[{"created":"@86067.831777175","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244,"referenced_errors":[{"created":"@86067.829777489","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244}]}]}

I0101 23:54:27.834777226       1 client_channel.cc:1332]               chand=103633b8: update: state=TRANSIENT_FAILURE picker=10369730

I0101 23:54:27.835777173       1 connectivity_state.cc:151]            ConnectivityStateTracker client_channel[10363450]: CONNECTING -> TRANSIENT_FAILURE (helper)

I0101 23:54:27.835777173       1 client_channel.cc:3912]               chand=103633b8 calld=103621e0: LB pick returned FAILED (subchannel=0, error={"created":"@86067.834777226","description":"Resolver transient failure","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolving_lb_policy.cc","file_line":262,"referenced_errors":[{"created":"@86067.834777226","description":"DNS resolution failed","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/dns_resolver_ares.cc","file_line":370,"grpc_status":14,"referenced_errors":[{"created":"@86067.831777175","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244,"referenced_errors":[{"created":"@86067.829777489","description":"C-ares status is not ARES_SUCCESS: Could not contact DNS servers","file":"/local/mnt/workspace/gitlab/grpc/src/core/ext/filters/client_channel/resolver/dns/c_ares/grpc_ares_wrapper.cc","file_line":244}]}]}]})

I think this issue is related to DNS server, which is not set in our target. Can you please help me to know how I can setup the DNS server?

I am new to this networking as well as QNX.

Also correct me if I am suspecting wrong thing.

_______________________________________________

General

http://community.qnx.com/sf/go/post120875

To cancel your subscription to this discussion, please e-mail general-networking-unsubscribe@community.qnx.com

post120858: Not able to use lockf, after mapping the memory

Darshan B(deleted) — Wed, 22 Jul 2020 03:43:21 GMT

Hi,

I am trying to use lockf(shm_desc_, F_TLOCK, 0), but I am getting error ENOSYS.

Can anybody help me on this?

My code is simple as below:

```

  std::size_t size = sizeof(int64_t);

  shm_desc_ = shm_open(name.c_str(), O_CREAT | O_RDWR, 0777);

  if (shm_desc_ == -1) {

    LogError("Failed to open shared region");

    return;

  }

  if (ftruncate(shm_desc_, size) == -1) {

    LogError("Failed to access region");

    return;

  }

  value_ = (int64_t *)mmap(0, size, PROT_READ | PROT_WRITE, MAP_SHARED, shm_desc_, 0);

  if (value_ == MAP_FAILED) {

    LogError("Failed to map shared region");

    close(shm_desc_);

    shm_desc_ = -1;

    return;

  }

  if (-1 == lockf(shm_desc_, F_TLOCK, 0)) {

    LogError("trylock failed {}", errno);

    return;

  }

```

post120839: Re: Not able to find libc functions while compiling runtime

Darshan B(deleted) — Tue, 14 Jul 2020 05:31:14 GMT

Is there anyone to guide me on this?

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post120838

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post120838: Re: Not able to find libc functions while compiling runtime

Darshan B(deleted) — Tue, 14 Jul 2020 05:31:13 GMT

Is there anyone to guide me on this?

post120831: Not able to find libc functions while compiling runtime

Darshan B(deleted) — Thu, 09 Jul 2020 12:23:26 GMT

I have cross compiled gRPC (v1.27.3) for QNX 7 (ARM) platform. I am using the protobuf which was already available for

our system (QNX ARM v3.11.4).

After compiling it generated the binaries. I copied these binaries in our code to compile the application for QNX.

We are using cmake to compile QNX runtime code.

While compiling QNX I am getting this issue:

../../lib/libQAic.so: undefined reference to `inet_pton(int, char const*, void*)'

../../lib/libQAic.so: undefined reference to `gai_strerror(int)'

../../lib/libQAic.so: undefined reference to `clock_gettime(int, timespec*)'

../../lib/libQAic.so: undefined reference to `fcntl(int, int, ...)'

../../lib/libQAic.so: undefined reference to `unsetenv(char const*)'

../../lib/libQAic.so: undefined reference to `setenv(char const*, char const*, int)'

../../lib/libQAic.so: undefined reference to `nanosleep(timespec const*, timespec*)'

../../lib/libQAic.so: undefined reference to `stat(char const*, stat*)'

../../lib/libQAic.so: undefined reference to `pthread_cond_timedwait(_sync*, _sync*, timespec const*)'

../../lib/libQAic.so: undefined reference to `read(int, void*, unsigned long)'

../../lib/libQAic.so: undefined reference to `inet_ntop(int, void const*, char*, unsigned int)'

../../lib/libQAic.so: undefined reference to `pipe(int*)'

../../lib/libQAic.so: undefined reference to `write(int, void const*, unsigned long)'

../../lib/libQAic.so: undefined reference to `unlink(char const*)'

../../lib/libQAic.so: undefined reference to `close(int)'

../../lib/libQAic.so: undefined reference to `sysconf(int)'

../../lib/libQAic.so: undefined reference to `getaddrinfo(char const*, char const*, addrinfo const*, addrinfo**)'

../../lib/libQAic.so: undefined reference to `freeaddrinfo(addrinfo*)'

collect2: error: ld returned 1 exit status

test/kmd/CMakeFiles/qaic_json_test.dir/build.make:117: recipe for target 'test/kmd/qaic_json_test' failed

make[2]: *** [test/kmd/qaic_json_test] Error 1

CMakeFiles/Makefile2:1012: recipe for target 'test/kmd/CMakeFiles/qaic_json_test.dir/all' failed

make[1]: *** [test/kmd/CMakeFiles/qaic_json_test.dir/all] Error 2

make[1]: *** Waiting for unfinished jobs....

These all functions are already in libc. But they are not able to find it. The path is already included in environment

variable, where libc is available. But still did not able to find it.

Can anybody help me on this?

post120803: Anybody has compiled gRPC for QNX-7.0

Darshan B(deleted) — Tue, 23 Jun 2020 15:05:56 GMT

I want to cross compile gRPC for QNX-7.0 arm .

I m getting too many error related to POSIX, Thread, Link libraries.

Do we have any step by step guideline for this?

I am referring https://github.com/grpc/grpc/tree/master/src/cpp

cmake compilation method.

post120802: Re: gRPC compilation failed

Darshan B(deleted) — Tue, 23 Jun 2020 10:33:29 GMT

Also I am getting this issue

qnx_tools/qnx7/aarch64le/usr/lib/libc++.so: file not recognized: File format not recognized

post120801: Re: gRPC compilation failed

Albrecht Uhlmann — Tue, 23 Jun 2020 10:25:54 GMT

in usr/lib, I can see libc++.so and libc++.a. Maybe there is a mismatch, e.g. linker interprets the .so as static lib?

libc++ is also a symlink, maybe following links is behaving crazy.

-Al

post120800: Re: gRPC compilation failed

Darshan B(deleted) — Tue, 23 Jun 2020 10:11:28 GMT

Tried deleting pthread in CMakeLists.txt, but still facing this issue

post120799: Re: gRPC compilation failed

Peter Weber — Tue, 23 Jun 2020 09:32:33 GMT

sorry, delete my last comment.

It is aarch64 right?

post120798: Re: gRPC compilation failed

Peter Weber — Tue, 23 Jun 2020 09:30:55 GMT

... and fix the search path for libraries.

You cant link a x86_64 bit project against a aarch64 lib. Different architecture!

post120797: Re: gRPC compilation failed

Peter Weber — Tue, 23 Jun 2020 09:29:50 GMT

delete -lpthread lib

In QNX, posix threads are implemented in libc.

post120796: gRPC compilation failed

Darshan B(deleted) — Tue, 23 Jun 2020 09:26:55 GMT

I am trying to cross compile gRPC for QNX7 platform using cmake.

While compiling I am getting error

qnx_tools/x86_64/usr/bin/i586-pc-nto-qnx7.0.0-ld: cannot find -lpthread

qnx7/aarch64le/usr/lib/libc++.so: file not recognized: File format not recognized

Please guide me how I can solve this issue.

I am using command :

cmake -DCMAKE_TOOLCHAIN_FILE=/tmp/qnxqrm/toolchain.cmake ../..

toolchain.cmake

SET(CMAKE_SYSTEM_NAME QNX)

SET(CMAKE_SYSTEM_PROCESSOR arm)

set(CMAKE_CROSSCOMPILING 1)

set(CMAKE_SYSROOT "$ENV{TOOLCHAIN_PATH}")

set(CMAKE_C_COMPILER "$ENV{TOOLCHAIN_PATH}/usr/bin/aarch64-unknown-nto-qnx7.0.0-gcc-5.4.0")

set(CMAKE_CXX_COMPILER "$ENV{TOOLCHAIN_PATH}/usr/bin/aarch64-unknown-nto-qnx7.0.0-g++-5.4.0")

set(CMAKE_CXX_FLAGS "-Vgcc_ntoaarch64 -O2 -Wc,-Wall -DBUILDENV_qss -g -Os -Wall -march=armv8-a -mcpu=cortex-a57 -mtune=cortex-a57 \

        -fstack-protector-strong -DNDEBUG -DFMT_HEADER_ONLY -EL -DVARIANT_le -std=c++11 -stdlib=libstdc++ -lang-c++ \

        -I . \

        -I$ENV{TOOLCHAIN_PATH}/usr/include \

        -I$ENV{TOOLCHAIN_PATH}/usr/include/WF \

        -I$ENV{TOOLCHAIN_PATH}/usr/include/KHR \

        -I$ENV{INSTALL_ROOT_nto}/usr/include \

        -Wl,-L$ENV{INSTALL_ROOT_nto}/aarch64le/lib \

        -Wl,-L$ENV{INSTALL_ROOT_nto}/aarch64le/usr/lib \

        -Wl,-L$ENV{TOOLCHAIN_PATH}/aarch64le/lib \

        -Wl,-L$ENV{TOOLCHAIN_PATH}/aarch64le/usr/lib")

Also let me know if I have missed any flags in CXX.

post120603: Re: QNX 6.5 SP 1 and USB PCL printers and QNX 6.6.0

Janusz Ruszel — Fri, 08 May 2020 21:21:52 GMT

Similar symptom on x86 (system freezes) based system with  HP LaserJet P2014 printer attached to  USB port. The system log listed below, please advice.

Mar 19 03:41:44    6    12   100 bus::USB

busno::0x01

devno::0x09

vendor_id::0x03f0

product_id::0x3917

configurations::1

configuration::1

serial_number::LW211GS

device_class::0x00

max_packet_size0::64

manufacturer::Hewlett-Packard

product::HP LaserJet P2014

upstream_device_address::0

upstream_host_controller::1

upstream_port::2

upstream_port_speed::High

topology::(0,2)

drivers_matched::0

drivers_running::0

Mar 19 03:41:44    5    12   100 USB-1.9:0: vid=03f0, did=3917: No match found, class=0x00

Mar 19 03:41:44    2    12     0 USB_SelectInterface:  Select iface devno 9, ifc 0, alt 0

Mar 19 03:41:44    2    12     0 process_xfer_complete: No Outstanding Transfer on XRING... where is this event coming from???

Mar 19 03:41:44    2    12     0 process_xfer_complete: evt->w0 = 0xda229000 evt->w1 = 0x0 evt->w2 = 0x1b000000 evt->w3 = 0x9028000

Mar 19 03:41:44    2    12     0 process_xfer_complete: No Outstanding Transfer on XRING... where is this event coming from???

Mar 19 03:41:44    2    12     0 process_xfer_complete: evt->w0 = 0xd6452000 evt->w1 = 0x0 evt->w2 = 0x1b000000 evt->w3 = 0x9038000

Mar 19 03:44:05    2    12     0 process_control_xfer: Transfer Failed ep->devaddr = 0x9 ep->num = 0 ep->dir = 0x0 Completion Code = TRB_COMP_CODE_USB_TRANSACTION_ERROR

Mar 19 03:44:05    2    12     0 process_control_xfer: evt->w0 = 0xd9bd15f0 evt->w1 = 0x0 evt->w2 = 0x4000001 evt->w3 = 0x9018000

Mar 19 03:44:05    2    12     0 xhci_ctrl_transfer:control_xfer_wait() failed  rc = -1, epstate = 0x3

Mar 19 03:44:05    2    12     0 xhci_ctrl_transfer: Details : addr = 0x9 buffer = 0xd7ff6014 len = 1 flags = 0x80000004

post120495: Re: How to print to console

Albrecht Uhlmann — Tue, 21 Apr 2020 18:13:56 GMT

This works only on x86 targets.

You can also try to redirect to /dev/console. This device is managed by slogger2, and all your output will appear in the system log. Can be useful to correlate your own messages with other stuff going on in the system.

Regards,

Albrecht

post120491: Re: How to print to console

Duncan Perrett(deleted) — Tue, 21 Apr 2020 15:27:19 GMT

Ah - I've just succeeded with

> /dev/con1

post120490: Re: How to print to console

Duncan Perrett(deleted) — Tue, 21 Apr 2020 15:25:53 GMT

I can redirect to a file but I want to know how to redirect to the console.

ssh -l root 192.168.50.51 "/tmp/QNX > /tmp/logtxt"

I've tried > /dev/ttyp0 but no output is seen.

post120489: How to print to console

Duncan Perrett(deleted) — Tue, 21 Apr 2020 14:52:26 GMT

I cross-compile and build an Arm7 le QNX x86_64 executable ("QNX") on a Windows host and then scp it onto the pre-built Vmware virtual QNX target I downloaded from the Software Center - QNX_SDP-x86_64.

This works and I can see my executable is listed by typing pidin | grep QNX

I execute it remotely from the Windows host using ssh, eg:

scp ./QNX root@192.168.50.51:/tmp

ssh 192.168.50.51 -l root /tmp/QNX

Currently the QNX executable just prints "Hello World" every 2 seconds to a console.

How can I ensure that the executable prints out to the terminal on the virtual QNX target in VMWare?

I have read about tinit but it doesn't seem to exist in my virtual machine.

post120064: Re: grpc on QNX

Zhilkin Sergey(deleted) — Fri, 25 Oct 2019 08:51:44 GMT

Hi!

Follow instructions in https://github.com/grpc/grpc/tree/master/src/cpp

post120063: Re: Neutrino 7 Packages

Zhilkin Sergey(deleted) — Fri, 25 Oct 2019 08:45:24 GMT

For building 7.0 packages, someone need NC licence for 7, that, I think, doesn't exist anymore.

post120053: Neutrino 7 Packages

Nathan Wright — Thu, 17 Oct 2019 13:13:01 GMT

Hello!

We are migrating from 6.5 to 7. On 6.5 we got many supporting packages from here...

ftp://ftp.netbsd.org/pub/pkgsrc/packages/QNX/i386/6.5.0_head_20110826/

Is there a comparable site to get QNX 7 packages from? We need stuff like wget, python, sqlite, etc, that our apps depend on.

Thanks,

Nathan

post119775: QNX SDP 7.0 BSP for Texas Instruments Jacinto 6 DRA74x EVMc help

youyang cheng — Sat, 29 Jun 2019 09:26:12 GMT

hi all , i am currently using momenttics to build BSP project ,which was downloaded from QNX software center, but failed

here are errors:

Description Resource Path Location Type

fatal error: hw/pci.h: No such file or directory audio_driver.h /ti-j6-dra74x-vayu-evm_650sp1/install/usr/include line 38 C/C++ Problem

Any help would be appreciated.

post119732: Re: Overlay Filesystem

nathan bazzell — Thu, 06 Jun 2019 14:36:32 GMT

I was able to get what I wanted using this command:

mount -t qnx6 -o sync=mandatory,overalloc /dev/hd0t177 /

Files can be overwritten in the writable space and once deleted we see the read-only files back. Works for folders too.

post119720: Re: Overlay Filesystem

nathan bazzell — Fri, 31 May 2019 12:09:59 GMT

I'm using a 16GB SD Card for initial development.  The boot partition is 500M formatted to vfat.  (mkfs.vfat shows Id=83 and "df -Th" shows /dev/sdb1 is type vfat.)

The filesystem I want to overlay as writable space is qnx6 because I need support for symbolic links.  Not sure if its a problem mixing filesystem types for the purpose of doing an overlay.

Here are the commands to create the qnx6 writable partition which I'm want to lay on top of the read-only (vfat) boot partition:

  fdisk /dev/hd0 add -t 177 -p 75

  mount -e /dev/hd0

  mkqnx6fs -q /dev/hd0t177

  mount -t qnx6 -o sync=mandatory /dev/hd0t177 /writable_partition

post119719: Re: Overlay Filesystem

Dennis Kellly(deleted) — Thu, 30 May 2019 20:42:18 GMT

What is the mediaf or the filesystem? mmcsd, ahci, other?

post119718: Overlay Filesystem

nathan bazzell — Thu, 30 May 2019 17:35:23 GMT

Could someone provide an example on how to use QNX's mount to achieve an overlay filesystem using the "mount -o before after" command?

Our QNX image is on a fat32 filesystem containing the boot and image files. This comes up as read only. I'd like to overlay a writable filesystem (qnx6 for now) on top of this to ease development. To the extent that a file created in the writable space would hide the file in the read-only space and be used in place of it. For example, to achieve a similar result as the (typical Linux mount command using the under/over i.e. “mount -t overlay overlay -olowerdir=/lower,upperdir=/upper,workdir=/work /merged”.)

post119698: grpc on QNX

Kok Keong Neo(deleted) — Tue, 07 May 2019 09:25:58 GMT

How to use grpc on QNX?

post119578: Re: BOOST QNX 7 (lib: Archive has no index)

Ian Clough(deleted) — Tue, 12 Mar 2019 19:40:08 GMT

Using Ubuntu to build:

bootstrap.sh --with-icu

./b2 toolset=qcc target-os=qnxnto threadapi=pthread link=static -l240 --layout=system --without-python cxxflags="-Vgcc_ntox86_64 -Y_gpp" linkflags="-Vgcc_ntox86_64 -Y_gpp -lang-c++" -j7 > buildlog.txt 2>&1

However, when I attempt to build the boost test program which relies on libboost_regex I get:

make -j8 all

qcc -lang-c++ -Vgcc_ntox86 -c -Wp,-MMD,build/x86-debug/src/t2.d,-MT,build/x86-debug/src/t2.o -o build/x86-debug/src/t2.o -Wall -fmessage-length=0 -g -O0 -fno-builtin src/t2.cpp

qcc -lang-c++ -Vgcc_ntox86 -o build/x86-debug/t2 -LC:/qnx700/target/qnx7/usr/pkg/lib/boost_1_69_x86_64 build/x86-debug/src/t2.o -lboost_regex

C:/qnx700/target/qnx7/usr/pkg/lib/boost_1_69_x86_64\libboost_regex.a: error adding symbols: Archive has no index; run ranlib to add one

post119474: Re: BOOST QNX 7

Ian Clough(deleted) — Tue, 05 Feb 2019 13:23:05 GMT

"porting of existing third-party code should be done under Linux hosts"

After repeatedly trying to build on a windows host (and failing), I'm pretty sure Linux is the right approach. Thanks Albrecht.

post119471: Re: BOOST QNX 7

Albrecht Uhlmann — Tue, 05 Feb 2019 09:53:10 GMT

We are also very much interested in getting a port of recent version of boost.

From my experience, any porting of existing third-party code should be done under Linux hosts.

Regards,

Albrecht

post119470: Re: BOOST QNX 7

Ian Clough(deleted) — Mon, 04 Feb 2019 19:23:20 GMT

More specifics:

I've tried building Boost on my QNX 7.0 hosted on a windows 10 platform but the issues being that the bootstrap.bat will not accept qcc as a toolset. I tried building in a cygwin shell, but this also failed.

post119466: Boost From Pkg Src 650 on QNX 7

Ian Clough(deleted) — Fri, 01 Feb 2019 19:54:06 GMT

Our 650 codebase is being migrated to QNX 7, we had a pre-built boost (1.53) from pkg src which was compiling, linking and running fine under 650, but when I try to build for 7.0 I get tons of these linker errors. Any thoughts greatly appreciated.

make all

make -j 1 -Cnto-x86-o -fMakefile all

make[1]: Entering directory 'C:/Users/clougia/workspace/SIN_4.0/seg_src/HardwareSpecific/Lanner/MenuingSystem/nto-x86-o'

C:/qnx700/host/win64/x86_64/usr/bin/rm -f C:/Users/clougia/workspace/SIN_4.0/seg_src/HardwareSpecific/Lanner/MenuingSystem/nto-x86-o/menuLcd.exe

C:/qnx700/host/win64/x86_64/usr/bin/qcc -Vgcc_ntox86 -lang-c++ -L"C:\qnx700\target\qnx7\usr\pkg\lib\boost" -Bstatic -o C:/Users/clougia/workspace/SIN_4.0/seg_src/HardwareSpecific/Lanner/MenuingSystem/nto-x86-o/menuLcd.exe mainMenu.o ppsHelper.o -L. -LC:/qnx700//target/qnx7/x86/lib -LC:/qnx700//target/qnx7/x86/usr/lib -Wl,--rpath-link,. -Wl,--rpath-link,C:/qnx700//target/qnx7/x86/lib -Wl,--rpath-link,C:/qnx700//target/qnx7/x86/usr/lib -lboost_thread -lboost_system -lboost_program_options -lboost_filesystem -lsocket

C:\qnx700\host\win64\x86_64\usr\bin\i586-pc-nto-qnx7.0.0-ld: C:\qnx700\target\qnx7\usr\pkg\lib\boost\libboost_thread.a(thread.o): warning: relocation in readonly section `.rodata'.

mainMenu.o: In function `menuSystem::menuSystem(int, char**)':

C:/Users/clougia/workspace/SIN_4.0/seg_src/HardwareSpecific/Lanner/MenuingSystem/mainMenu.cpp:86: undefined reference to `boost::program_options::options_description::options_description(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, unsigned int, unsigned int)'

C:/Users/clougia/workspace/SIN_4.0/seg_src/HardwareSpecific/Lanner/MenuingSystem/mainMenu.cpp:99: undefined reference to `boost::program_options::operator<<(std::__1::basic_ostream<char, std::__1::char_traits<char> >&, boost::program_options::options_description const&)'

mainMenu.o: In function `boost::program_options::typed_value<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, char>::xparse(boost::any&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) const':

c:\qnx700\target\qnx7\usr\include\boost\program_options\detail/value_semantic.hpp:170: undefined reference to `boost::program_options::validate(boost::any&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >*, int)'

mainMenu.o: In function `boost::program_options::validation_error::validation_error(boost::program_options::validation_error::kind_t, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int)':

c:\qnx700\target\qnx7\usr\include\boost\program_options/errors.hpp:372: undefined reference to `boost::program_options::error_with_option_name::error_with_option_name(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int)'

mainMenu.o: In function `boost::program_options::basic_command_line_parser<char>::extra_parser(boost::function1<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>)':

c:\qnx700\target\qnx7\usr\include\boost\program_options\detail/parsers.hpp:77: undefined reference to `boost::program_options::detail::cmdline::set_additional_parser(boost::function1<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&>)'

mainMenu.o: In function `std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > boost::program_options::to_internal<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > >(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)':

c:\qnx700\target\qnx7\usr\include\boost\program_options\detail/convert.hpp:79: undefined reference to `boost::program_options::to_internal(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)'

mainMenu.o: In function `boost::program_options::basic_command_line_parser<char>::basic_command_line_parser(int, char const* const*)':

c:\qnx700\target\qnx7\usr\include\boost\program_options\detail/parsers.hpp:43: undefined reference to `boost::program_options::detail::cmdline::cmdline(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&)'

mainMenu.o: In function `void boost::program_options::validate<int, char>(boost::any&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, int*, long)':

c:\qnx700\target\qnx7\usr\include\boost\program_options\detail/value_semantic.hpp:92: undefined reference to `boost::program_options::invalid_option_value::invalid_option_value(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)'

mainMenu.o:(.data.rel.ro._ZTVN5boost16exception_detail19error_info_injectorINS_15program_options20invalid_option_valueEEE[_ZTVN5boost16exception_detail19error_info_injectorINS_15program_options20invalid_option_valueEEE]+0x18): undefined reference to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options20invalid_option_valueEEEEE[_ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options20invalid_option_valueEEEEE]+0x1c): undefined reference to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost16exception_detail19error_info_injectorINS_15program_options16validation_errorEEE[_ZTVN5boost16exception_detail19error_info_injectorINS_15program_options16validation_errorEEE]+0x18): undefined reference to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options16validation_errorEEEEE[_ZTVN5boost16exception_detail10clone_implINS0_19error_info_injectorINS_15program_options16validation_errorEEEEE]+0x1c): undefined reference to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost15program_options16validation_errorE[_ZTVN5boost15program_options16validation_errorE]+0x18): undefined reference to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost15program_options20invalid_option_valueE[_ZTVN5boost15program_options20invalid_option_valueE]+0x18): more undefined references to `boost::program_options::error_with_option_name::substitute_placeholders(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) const' follow

mainMenu.o:(.data.rel.ro._ZTVN5boost15program_options11typed_valueIicEE[_ZTVN5boost15program_options11typed_valueIicEE]+0x1c): undefined reference to `boost::program_options::value_semantic_codecvt_helper<char>::parse(boost::any&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, bool) const'

mainMenu.o:(.data.rel.ro._ZTVN5boost15program_options11typed_valueINSt3__112basic_stringIcNS2_11char_traitsIcEENS2_9allocatorIcEEEEcEE[_ZTVN5boost15program_options11typed_valueINSt3__112basic_stringIcNS2_11char_traitsIcEENS2_9allocatorIcEEEEcEE]+0x1c): undefined reference to `boost::program_options::value_semantic_codecvt_helper<char>::parse(boost::any&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, bool) const'

C:\qnx700\target\qnx7\usr\pkg\lib\boost\libboost_thread.a(thread.o): In function `T.3093':

thread.cpp:(.text+0xa24): undefined reference to `std::_String_base::_Xlen() const'

C:\qnx700\target\qnx7\usr\pkg\lib\boost\libboost_thread.a(thread.o): In function `std::string::erase(unsigned int, unsigned int)':

thread.cpp:(.text._ZNSs5eraseEjj[_ZNSs5eraseEjj]+0x97): undefined reference to `std::_String_base::_Xran() const'

C:\qnx700\target\qnx7\usr\pkg\lib\boost\libboost_thread.a(thread.o): In function `std::string::append(std::string const&, unsigned int, unsigned int)':

thread.cpp:(.text._ZNSs6appendERKSsjj[_ZNSs6appendERKSsjj]+0x5a): undefined reference to `std::_String_base::_Xran() const'

.

.

.

thanks everyone!

Ian

post119465: BOOST QNX 7

Ian Clough(deleted) — Fri, 01 Feb 2019 19:50:59 GMT

Does anyone have a working version of Boost for QNX 7??

I've built & re-built but the closest I've gotten has been my app failing with

<progname>: Attempting to exec a shared lib

post119464: Re: boost_thread (missing dependency: Threads)

Ian Clough(deleted) — Fri, 01 Feb 2019 19:47:50 GMT

Did you ever find a solution?

post119362: Re: C++ libraries for 6.5SP1

Albrecht Uhlmann — Tue, 18 Dec 2018 13:24:19 GMT

Hi Srini,

can't answer everything, but some of your questions:

- boost library for C++ - How to build? Are all of boost possible to use?

--> Probably need to do a number of modifications to build scripts, and some restrictions may apply due to gcc 4.4 version C++ limitations.

- Latest openssl libraries and the tool itself

--> We managed to build the latest version for x86 without any problems, including the tests.

In any case, I recommend using a self-hosted installation for porting.

Regards,

Albrecht

post119347: C++ libraries for 6.5SP1

Rajagopalan Srinivasan — Sun, 09 Dec 2018 20:48:14 GMT

We are starting a new project targeting 6.5SP1 and considering various options and hoping to get feedback on the following:

boost library for C++ - How to build? Are all of boost possible to use?

Stability etc of the "Dinkumware" libraries - particularly EC++

Latest openssl libraries and the tool itself

Any other template or otherwise libraries for network programming

Any port of python3 available. If not some guidance on porting.

Thanks, srini

post119227: Define custom symbol for custom build configurations in QNX Momentics

Adem Agdas(deleted) — Tue, 23 Oct 2018 10:42:19 GMT

Let's make an example in Eclipse: I have two different source files for a Startup with a main-function, but depending on which build configuration I use, either one of them is used, or the other one.

I do this, by defining a custom symbol with the "-d" option in one of the build configurations, and leaving it out in the other one.

#ifdef CUSTOMSYMBOL

...

Specifically, under Eclipse, I can edit these settings under Project -> Properties -> C/C++ Build -> Settings -> Cygwin C++ Compiler -> Prepocessor imgdesc (I am using Cygwin under Windows for my C++ projects)

This works. Depending on which build configuration I use, only one of the two source files will be built by the compiler and the other one will be ignored.

However, this setting is specific to Cygwin.

I can't figure out, how to do it in QNX Momentics, since I do not have those options under the qcc Compiler.

post119210: boost_thread (missing dependency: Threads)

sangamesh kotagi(deleted) — Fri, 12 Oct 2018 15:12:37 GMT

Hello,

I am trying to compile Vsomeip for QNX7 which needs boost. I get the follwoing error

swat@ubuntu:~/ide-7.0-vsomeip/QNXCMakeVSomeip/build$ cmake ..

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1125 ] _boost_TEST_VERSIONS = 1.68.0;1.68;1.67.0;1.67;1.66.0;1.66;1.65.1;1.65.0;1.65;1.64.0;1.64;1.63.0;1.63;1.62.0;1.62

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1127 ] Boost_USE_MULTITHREADED = OFF

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1129 ] Boost_USE_STATIC_LIBS = ON

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1131 ] Boost_USE_STATIC_RUNTIME = OFF

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1133 ] Boost_ADDITIONAL_VERSIONS =

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1135 ] Boost_NO_SYSTEM_PATHS =

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1203 ] Declared as CMake or Environmental Variables:

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1205 ]   BOOST_ROOT = /home/swat/boost_1_62_0

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1207 ]   BOOST_INCLUDEDIR =

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1209 ]   BOOST_LIBRARYDIR =

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1211 ] _boost_TEST_VERSIONS = 1.68.0;1.68;1.67.0;1.67;1.66.0;1.66;1.65.1;1.65.0;1.65;1.64.0;1.64;1.63.0;1.63;1.62.0;1.62

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1310 ] location of version.hpp: /usr/local/include//boost/version.hpp

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1334 ] version.hpp reveals boost 1.62.0

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1420 ] guessed _boost_COMPILER = -gcc54

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1430 ] _boost_MULTITHREADED =

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1506 ] _boost_RELEASE_ABI_TAG = -

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1508 ] _boost_DEBUG_ABI_TAG = -d

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1571 ] _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH_boost_LIBRARY_SEARCH_DIRS_DEBUG   = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- Could NOT find Threads (missing: Threads_FOUND)

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for SYSTEM_LIBRARY_RELEASE: boost_system-gcc54-1_62;boost_system-gcc54;boost_system-1_62;boost_system;boost_system

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for SYSTEM_LIBRARY_DEBUG: boost_system-gcc54-d-1_62;boost_system-gcc54-d;boost_system-d-1_62;boost_system-d;boost_system;boost_system

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for THREAD_LIBRARY_RELEASE: boost_thread-gcc54-1_62;boost_thread-gcc54;boost_thread-1_62;boost_thread;boost_thread

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for THREAD_LIBRARY_DEBUG: boost_thread-gcc54-d-1_62;boost_thread-gcc54-d;boost_thread-d-1_62;boost_thread-d;boost_thread;boost_thread

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for LOG_LIBRARY_RELEASE: boost_log-gcc54-1_62;boost_log-gcc54;boost_log-1_62;boost_log;boost_log

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for LOG_LIBRARY_DEBUG: boost_log-gcc54-d-1_62;boost_log-gcc54-d;boost_log-d-1_62;boost_log-d;boost_log;boost_log

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for CHRONO_LIBRARY_RELEASE: boost_chrono-gcc54-1_62;boost_chrono-gcc54;boost_chrono-1_62;boost_chrono;boost_chrono

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for CHRONO_LIBRARY_DEBUG: boost_chrono-gcc54-d-1_62;boost_chrono-gcc54-d;boost_chrono-d-1_62;boost_chrono-d;boost_chrono;boost_chrono

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for DATE_TIME_LIBRARY_RELEASE: boost_date_time-gcc54-1_62;boost_date_time-gcc54;boost_date_time-1_62;boost_date_time;boost_date_time

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for DATE_TIME_LIBRARY_DEBUG: boost_date_time-gcc54-d-1_62;boost_date_time-gcc54-d;boost_date_time-d-1_62;boost_date_time-d;boost_date_time;boost_date_time

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for ATOMIC_LIBRARY_RELEASE: boost_atomic-gcc54-1_62;boost_atomic-gcc54;boost_atomic-1_62;boost_atomic;boost_atomic

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for ATOMIC_LIBRARY_DEBUG: boost_atomic-gcc54-d-1_62;boost_atomic-gcc54-d;boost_atomic-d-1_62;boost_atomic-d;boost_atomic;boost_atomic

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for LOG_SETUP_LIBRARY_RELEASE: boost_log_setup-gcc54-1_62;boost_log_setup-gcc54;boost_log_setup-1_62;boost_log_setup;boost_log_setup

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for LOG_SETUP_LIBRARY_DEBUG: boost_log_setup-gcc54-d-1_62;boost_log_setup-gcc54-d;boost_log_setup-d-1_62;boost_log_setup-d;boost_log_setup;boost_log_setup

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for FILESYSTEM_LIBRARY_RELEASE: boost_filesystem-gcc54-1_62;boost_filesystem-gcc54;boost_filesystem-1_62;boost_filesystem;boost_filesystem

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for FILESYSTEM_LIBRARY_DEBUG: boost_filesystem-gcc54-d-1_62;boost_filesystem-gcc54-d;boost_filesystem-d-1_62;boost_filesystem-d;boost_filesystem;boost_filesystem

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1760 ] Searching for REGEX_LIBRARY_RELEASE: boost_regex-gcc54-1_62;boost_regex-gcc54;boost_regex-1_62;boost_regex;boost_regex

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_RELEASE = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_RELEASE = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1813 ] Searching for REGEX_LIBRARY_DEBUG: boost_regex-gcc54-d-1_62;boost_regex-gcc54-d;boost_regex-d-1_62;boost_regex-d;boost_regex;boost_regex

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:400 ]  Boost_LIBRARY_DIR_DEBUG = /usr/local/lib/ _boost_LIBRARY_SEARCH_DIRS_DEBUG = /usr/local/lib/;NO_DEFAULT_PATH;NO_CMAKE_FIND_ROOT_PATH

-- [ /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:1887 ] Boost_FOUND = 1

CMake Error at /usr/local/share/cmake-3.12/Modules/FindBoost.cmake:2048 (message):

  Unable to find the requested Boost libraries.

  Boost version: 1.62.0

  Boost include path: /usr/local/include/

  Could not find the following static Boost libraries:

          boost_thread (missing dependency: Threads)

  Some (but not all) of the required Boost libraries were found.  You may

  need to install these additional Boost libraries.  Alternatively, set

  BOOST_LIBRARYDIR to the directory containing Boost libraries or BOOST_ROOT

  to the location of Boost.

Call Stack (most recent call first):

  CMakeLists.txt:184 (find_package)

-- Boost was not found!

-- Checking for module 'automotive-dlt >= 2.11'

--   No package 'automotive-dlt' found

-- Checking for module 'libsystemd'

--   No package 'libsystemd' found

I have compiled the boost using qcc and the thread library is generated. Please see the attached screenshot.

Can someone help me?

post119204: QNX 6.4.1 on VMware does not show HELP

Cong Pham(deleted) — Fri, 12 Oct 2018 09:02:27 GMT

Hello, I need help with QNX 6.4.1 on VMware

I have QNX 6.4.1 run on VMware.

It works fine. I can run ped, pfm, PhAB normally...

But I can not run helpviewer well. It showes the main window but does not show the contents ( it is transparent ).

The same problem with workspace 0.5C ( a code editor - http://pages.infinit.net/micbel/), It does not load my source-code :(

post119125: Generate the .so library

yang sen(deleted) — Tue, 18 Sep 2018 07:25:28 GMT

Based The QNX C++ Project(Version: 5.0.1),I want to compile and generate the *.so. But,I follow the

QNX Momentics IDE User's Guide,the Key setting place,like this:

http://www.qnx.com/developers/docs/6.6.0_anm11_wf10/index.html#com.qnx.doc.ide.userguide/topic/wizards_Library_Tab.html

So,I can't find the library tab.Help me,thanks!

post119021: Re: CAN Test Between Two Beaglebones

Cristian dos Santos(deleted) — Mon, 30 Jul 2018 17:12:37 GMT

Hi Stefan,

I am trying to play with the Beaglebone Black and CAN Bus using QNX 7.0.

Could you please tell me where I can find the CAN driver?

Thank you in advance.

C.K.

post118879: Re: QNX Code Coverage - HTTP ERROR 500

yang elmar(deleted) — Mon, 18 Jun 2018 06:48:47 GMT

> Hi,

>

> I am able to import all the code coverage files from my target and when I try

> to generate a code coverage report on my QNX IDE (Version: 4.7.0), I get this

> error:

>

> ***********************************************************

> HTTP ERROR 500

> Problem accessing /qnx-coverage/coverage. Reason:

>

> Exception occurred during ctor/Transformation:

> debug is false

> Parsing XSL Stylesheet Document from configuration: /webapp/WEB-INF/xslt/cov-

> html.xslt

> Performing transformation...

>

> javax.xml.transform.TransformerException: java.lang.NullPointerException

>

> ***********************************************************

>

> Any help on this would be appreciated.

>

> Regards

> Sunil Keshava

The first thing you need to know about an "Internal Server Error" is that the error can only be resolved by fixes to the Web server software. It is not a client-side problem meaning that the problem is not with your browser, your computer, or your internet connection. However, there are often ways to quickly get around the problem:

Check the Error Logs

Errors in .htaccess files

Clear your browser cookies and cache

Reload or Refresh the Webpage

Not Yet?

If you've tried to troubleshoot your 500 server error and still can't fix it, then you may need to speak to your web host and see if they are able to find what process or script is causing the error.

http://net-informations.com/q/mis/500.html

post118841: BSP Xilinx Zynq 7000 ZC702 EVM Help

Prentyce Albright(deleted) — Tue, 22 May 2018 17:14:14 GMT

I am currently using an older Xilinx Zynq 7000 ZC702 EVM BSP: http://community.qnx.com/sf/wiki/do/viewPage/projects.bsp/wiki/XilinxZc702?showDetails=true

and other than the fact that it can’t do 4-byte addressing everything else works. I am trying to upgrade to the latest Xilinx Zynq 7000 ZC702 EVM BSP from 2018 which supports QSPI 4-byte addressing. I downloaded the new BSP and I was able to build the project. I also created an image file using a boot.bif:

[bootloader] fsbl.elf

Bitstream.bit

Ipl-xzynq-zc702.elf

[offset=600000, load=0x16000000] QNX-IFS.ub

When I download the image to the board it boots up using the FSBL but never runs the IPL program. I’ve been able to get this to work countless times using the older BSP.

Any help would be appreciated.

post118735: QNX SDP 7 with serial port debugging

kumar bhatia — Mon, 02 Apr 2018 20:44:32 GMT

Hello. Me again :) Now using a dedicated box with QNX SDP 7 loaded and attempting to connect to Momentics over TCP/IP. While this worked well using the VMWare method, it is failing to operate using a dedicated box. Been at this for over 2 days and out of ideas.

Remote box is booting and assigning to wm0 : 192.168.0.102 -> we can ping this box (so we think it is this network card we are pinging) and vice versa with the Windows debug machine. Yet Momentics stays red. As noted, VMWare works in seconds.

Using Port 8000.

Giving up on the above, attempting to use a null modem cable (confirmed working for testing our ports between boxes) + our PCIe serial adapters (we manufacture - confirmed working with the devc-serpci driver).

Commands we are using to launch with serial port mode:

devc-serpci ; to launch our PCIe serial card - we are assigned /dev/ser3

Using the same null modem cable, we are able to chat with the remote box on our PCIe serial card in the other box using Teraterm @ 57600 baud. Also ok @ 115200 baud.

Next, rebooted to get a fresh start and ran devc-serpci again...

devc-pty &

stty </dev/ser3 ; boot default is 57600 - ok, so we will use this value for now

pdebug /dev/ser3,57600 &

Then on the remote box, created a new serial port debug session with our COM13 (confirmed this is the label) @ 57600 and the box sits there without adding this new entry.

Is serial port debugging still available on QNX SDP 7 ?

Any ideas? We have tried I believe all possible permutations known to us.

post118734: Re: QNX SDP 7 on VMWare ESXi ?

kumar bhatia — Mon, 02 Apr 2018 20:30:23 GMT

Thank you. Now working on a dedicated box with QNX SDP 7. Have some more questions so will start a different thread.

post118731: Re: QNX SDP 7 on VMWare ESXi ?

Dave Nickerson(deleted) — Mon, 02 Apr 2018 18:14:41 GMT

VMware environment was not intended for QNX Neutrino driver development. VMWare primarily provides a simulated x86 platform that is capable of running QNX Neutrino RTOS, using the simulated hardware provided by VMware.

VMware is a great platform for QNX Neutrino application development but not really intended for hardware driver development.

For developing QNX Neutrino hardware drivers or for determining if a particular hardware device is supported under QNX Neutrino, you should be setting up a real hardware platform running QNX Neutrino natively on the platform.

Dave

post118730: QNX SDP 7 on VMWare ESXi ?

kumar bhatia — Fri, 30 Mar 2018 18:41:13 GMT

Hello again. Now making some progress in compiling source code with Momentics on VMWare 14 with the supplied trial version of QNX SDP 7 for VMWare image. The compiled code is unable to locate our PCIe adapter.

Using pci-tool -v log, confirmed that the QNX SDP 7 platform is not detecting our adapter but ok with many others in the box.

Reviewing the Device Manager for the same box hosting this VMWare + QNX SDP 7 and our installed adapter (VID: 13a8) -> the PCIe adapter is operating fine. VMWare 14 apparently does not allow for PCI passthrough.

From our knowledge, only VMWare ESXi allow for PCIe passthrough. But is QNX SDP 7 supported on ESXi?

Or is there a method to allow for this add-on PCIe adapter to be seen by the QNX SDP 7 image? Believe that VMWare 14 is blocking this access but may be wrong.

Thank you.

post118728: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 20:37:46 GMT

Proof that the graphics mode is now working :) After following the putty mode access stuff. Next hurdle is to compile the driver code. Thanks again.

post118727: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 20:23:59 GMT

Hello. Was about to start drinking alcoholic beverages (I don't drink) and thought to email QNX. They prompted reply with the following info which we must have missed...courtesy of Dennis Kelly of QNX:

Hi Kumar-

>>> VMWare which is yielding a black screen only upon launch with ./etc/graphics-startup.sh

The VMware image is behaving normally.

You must note the IP address (use ifconfIg) prior to starting graphics script. Then, use putty in Windows to connect to the IP address of the VMware image and login as user “qnxuser” and password “qnxuser”. (Once you start graphics it clears the screen and you cannot use the PC console any longer.) To prove graphics is working, run “gles2-gears”.

>> We will try this asap and post our results. Thanks for everyone for their assistance and hoping this will get us moving. From what we have seen, there is no stand alone ISO image for this version of the OS. QNX 7 is based on the VMWare platform which is fine, once it works for us :)

post118726: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 18:22:43 GMT

I think we will give up on using VMWare with QNX 7.

Is there an ISO or similar install disk we can download and create to install QNX 7 on a blank drive on a dedicated PC?

From the Software Center, can only see the VMWare QNX installation (which for us is not working in graphics mode).

Thanks.

post118725: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 16:49:29 GMT

I think the attached is now correct :) Concerned about the file errors. In the meantime, completely removed from VMWare and also uninstalled this image using the Software Center tool. Then installed a fresh copy again and re-applied with VMWare. Launched and observed that the hard disk led is very busy on first launch but hardly at all if restarted.

For example,

2018-03-29T12:16:48.438-04:00| vmx| I125: ConfigDB: Failed to load C:\Users\lab\AppData\Roaming\VMware\config.ini

2018-03-29T12:16:48.438-04:00| vmx| I125: OBJLIB-LIB: Objlib initialized.

2018-03-29T12:16:48.438-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.438-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware\VMware Workstation): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.439-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.439-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware\VMware Workstation): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.439-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.439-04:00| vmx| I125: FILE: FileCreateDirectoryRetry: Non-retriable error encountered (C:\ProgramData\VMware\VMware Workstation): Cannot create a file when that file already exists (183)

2018-03-29T12:16:48.439-04:00| vmx| I125: DictionaryLoad: Cannot open file "C:\Users\lab\AppData\Roaming\VMware\config.ini": The system cannot find the file specified.

2018-03-29T12:16:48.439-04:00| vmx| I125: [msg.dictionary.load.openFailed] Cannot open file "C:\Users\lab\AppData\Roaming\VMware\config.ini": The system cannot find the file specified.

2018-03-29T12:16:48.439-04:00| vmx| I125: PREF Optional preferences file not found at C:\Users\lab\AppData\Roaming\VMware\config.ini. Using default values.

post118724: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Thu, 29 Mar 2018 16:37:33 GMT

You have to wait until after the ‘waitfor /dev/screen’ before you try running anything that requires graphics, screen won’t have had a chance to finish its initialization.

On 2018-03-29, 11:58 AM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    revised script







    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118722

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118723: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Thu, 29 Mar 2018 16:35:40 GMT

You didn’t include graphics-startup.sh.

You need to run slog2info after starting screen.  The log output you show has nothing from screen and it should, even if it fails.

On 2018-03-29, 11:47 AM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    Updated the graphics-startup.sh script (copied below). Still black after waiting 30+ seconds..log is attached from this session.



    Our interest is to compile the devc-8250 and related PCI driver code to modify and/or debug for future possible designs. Perhaps consider to use a different and dedicated box, without VMWare?







    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118721

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118722: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 16:17:31 GMT

revised script

post118721: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 16:06:29 GMT

Updated the graphics-startup.sh script (copied below). Still black after waiting 30+ seconds..log is attached from this session.

Our interest is to compile the devc-8250 and related PCI driver code to modify and/or debug for future possible designs. Perhaps consider to use a different and dedicated box, without VMWare?

post118720: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Thu, 29 Mar 2018 15:58:04 GMT

Did you try running sw-vsync after running  /etc/graphics-startup.sh?  It’s not clear exactly what you’re doing.  If screen was running properly it would just appear black since it never displays anything on its own accord.

You may have difficulty running additional commands in the console after starting screen since if nothing else, nothing will appear on the screen so you might want to create a script that first runs /etc/graphics-startup.sh and then runs sw-vsync.  Or if you can connect to the target over the network (I don’t know if sshd or telnetd are running) then you could do it from another terminal.  When you run sw-vsync you’ll likely want to ensure the environment variables LD_LIBRARY_PATH and GRAPHICS_ROOT are set as in the /etc/graphics-startup.sh script.

On 2018-03-29, 11:28 AM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    Attached is the log from a session after running ./etc/graphics-startup.sh script. There is a brief text flash display of "starting screen" text and then all black. Using vi editor, confirmed this detail is correct from the text script.



    Should we run again but with the:



    slog2info -b screen



    and then



    ./etc/graphics-startup.sh









    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118719

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118719: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 15:47:29 GMT

Attached is the log from a session after running ./etc/graphics-startup.sh script. There is a brief text flash display of "starting screen" text and then all black. Using vi editor, confirmed this detail is correct from the text script.

Should we run again but with the:

slog2info -b screen

and then

./etc/graphics-startup.sh

post118718: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Thu, 29 Mar 2018 15:34:23 GMT

Screenshot doesn’t show you running /etc/graphics-startup.sh, have you?  The logs show no sign of screen which I would have expected even if it had failed.

On 2018-03-29, 11:06 AM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    Hello. Thanks for the advice.



    Read on the VMWare website that VMWare needs to be cycled to read the updated .vmx file so did that but still no graphics.



    Did check the box for 3D graphics but received a warning that the host does not offer this support.



    I do not think we are running the sw-vsync command correctly. What is the proper way to launch this tool? Output is attached.











    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118717

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118717: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Thu, 29 Mar 2018 15:26:19 GMT

Hello. Thanks for the advice.

Read on the VMWare website that VMWare needs to be cycled to read the updated .vmx file so did that but still no graphics.

Did check the box for 3D graphics but received a warning that the host does not offer this support.

I do not think we are running the sw-vsync command correctly. What is the proper way to launch this tool? Output is attached.

post118716: Re: Unable to start in graphics mode with QNX 7

Michael Van Reenen — Thu, 29 Mar 2018 14:51:19 GMT

The graphics solution has changed from QNX650 (Photon) to SDP 7.0 (screen/composition manager).

There's no desktop application in SDP 7.0 so you may be seeing just a mouse because no graphics applications are running. Do you see anything appear on the screen if you run 'sw-vsync'?

post118714: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Thu, 29 Mar 2018 12:32:50 GMT

You might run

slog2info –b screen

to get logging output from screen.  This might give a clue to you or someone else.  Unfortunately I don’t do much with graphics these days so you’ve exhausted my one trick for getting graphics to work.

I have a vague memory that you had to select “Enable 3D graphics support” in the virtual machine settings.

On 2018-03-28, 6:17 PM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    Thank you Roger for your very prompt post. Attempted a mix of values including 10 for that parameter and still no video (remains black). Attempted down to 7 for that field. Also configured the VMWare to emulate down to older workstation 6.5-7.x



    The mouse pointer does disappear near the middle of the screen so wondering if the issue is related to the colour settings, yet the older QNX 6.5 boots just fine.



    My .vmx file is attached.



    Will test other ideas in the morning. Welcome other thoughts you may have to resolve.







    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118712

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118712: Re: Unable to start in graphics mode with QNX 7

kumar bhatia — Wed, 28 Mar 2018 22:36:42 GMT

Thank you Roger for your very prompt post. Attempted a mix of values including 10 for that parameter and still no video (remains black). Attempted down to 7 for that field. Also configured the VMWare to emulate down to older workstation 6.5-7.x

The mouse pointer does disappear near the middle of the screen so wondering if the issue is related to the colour settings, yet the older QNX 6.5 boots just fine.

My .vmx file is attached.

Will test other ideas in the morning. Welcome other thoughts you may have to resolve.

post118711: Re: Unable to start in graphics mode with QNX 7

Roger Maclean — Wed, 28 Mar 2018 22:13:59 GMT

There’s a setting called virtualHW.version in the .vmx file for your virtual machine.  You might like to try setting it to 10.  The default value (which I think is 12) doesn’t allow graphics to work at least with SDP 7 (I don’t know about 6.5).

On 2018-03-28, 5:44 PM, "kumar bhatia" <community-noreply@qnx.com> wrote:

    Hello. Hope someone can assist. We have just installed today VMWare 14.1.1 Workstation (64 bit) on a Windows 7 64 bit OS -> installed Ubuntu 16.04 LTS (64 bit); this OS is working fine.



    Next, installed the VMWare version of (time limited for 30 days) QNX 7 using the software center -> browsed to the installation folder and allow for VMWare to run this VMWare image. The image boots in text mode but appears to fail to run if under graphics mode. Repeated the same procedure with the older QNX 6.5 ISO -> runs fine under VMWare with graphics.



    We are attempting to run the graphics version as follows:



    # ./etc/graphics-start.sh <ENTER>



    end up with a black screen with only the mouse pointer. Believe the screen driver is an issue. The VM is not locked and we can close the VM fine.



    Attempted to increase the ram size under VMWare - no difference.



    Attempted to limit the graphics resolution under VMWare to 640x480 - no difference.



    Any suggestions on what we are doing wrong?











    _______________________________________________



    QNX Momentics Community Support

    http://community.qnx.com/sf/go/post118710

    To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post118710: Unable to start in graphics mode with QNX 7

kumar bhatia — Wed, 28 Mar 2018 22:03:33 GMT

Hello. Hope someone can assist. We have just installed today VMWare 14.1.1 Workstation (64 bit) on a Windows 7 64 bit OS -> installed Ubuntu 16.04 LTS (64 bit); this OS is working fine.

Next, installed the VMWare version of (time limited for 30 days) QNX 7 using the software center -> browsed to the installation folder and allow for VMWare to run this VMWare image. The image boots in text mode but appears to fail to run if under graphics mode. Repeated the same procedure with the older QNX 6.5 ISO -> runs fine under VMWare with graphics.

We are attempting to run the graphics version as follows:

# ./etc/graphics-start.sh <ENTER>

end up with a black screen with only the mouse pointer. Believe the screen driver is an issue. The VM is not locked and we can close the VM fine.

Attempted to increase the ram size under VMWare - no difference.

Attempted to limit the graphics resolution under VMWare to 640x480 - no difference.

Any suggestions on what we are doing wrong?

post118652: Instantiation of an std vector instance with an initialization list crashes with a SignalKill, bug with the builder tools?

Xiaofeng Zhao(deleted) — Mon, 05 Mar 2018 15:33:22 GMT

Source code attached

#include <cstdlib>

#include <iostream>

#include <vector>

#include <string>

int main(int argc, char *argv[]) {

    std::vector<int> test0{1, 2, 3, 4};

    std::vector<std::string> test{"a", "b", "c"};

    std::cout << test[0];

    return 0;

}

post118651: Instantiation of an std vector instance with an initialization list crashes with a SignalKill, bug with the builder tools?

Xiaofeng Zhao(deleted) — Mon, 05 Mar 2018 15:24:49 GMT

post118560: Re: Bind all processes to a single core by default.

Alastair Hoyle(deleted) — Mon, 12 Feb 2018 11:50:42 GMT

Hi

Thanks for your suggestions. I have edited the start-up script in my boot image and I have managed to move all the tasks onto a single core, except “procnto-smp-instr”. Am I right in assuming that “procnto-smp-instr” must be free to run on all cores? I assume since this is the microkernel it needs to run threads on all the cores.

Thanks

Alastair

post118527: instantiating std::vector with a initialization list caused the program to crash

Xiaofeng Zhao(deleted) — Thu, 08 Feb 2018 15:45:38 GMT

QNX software Development Platform 6.6

#include <cstdlib>

#include <iostream>

#include <vector>

#include <string>

int main(int argc, char *argv[]) {

    std::vector<int> test0{1, 2, 3, 4};

    std::vector<std::string> test{"a", "b", "c"};

    std::cout << test[0];

    return 0;

}

this program crashes (SignalKill) at the first line of the main function

post118516: Re: Inclusion of assembly file in QNX C project in Momentics IDE

Albrecht Uhlmann — Fri, 02 Feb 2018 12:59:32 GMT

I do not know what you got from NXP so I can't tell. Maybe a Linux/Windows lowe/upper case issue? Are you building under Windows or Linux, and how did NXP build?

I checked the source of the i.MX6 Sabresmart BSP startup program. It contains a lot of assembly source files, all with capital .S extension. In the common.mk/Makefile I found no entry relating to these, so in my opinion this should work out of the box if the Momentics/QNX make build system is used (common.mk, qconfig.mk, qtargets.mk etc.)

Maybe NXP have created an entirely customized configuration.

Why not talk to them?

Regards,

Albrecht

post118515: Re: Inclusion of assembly file in QNX C project in Momentics IDE

Girish Deshmukh(deleted) — Fri, 02 Feb 2018 05:20:48 GMT

Thanks Albrecht Uhlmann. But I am just trying to build a project which I received from NXP, and they have used .s extension for assembly files. Is anything special setting should I do in momentics if I want to build that NXP project? In another project there is no asm files so I build that project successfully but in this project I am facing problem because I think momentics is not recognising .s file.

I tried to add .s file type in momentics (Windows -> Preferences -> C/C++ -> File Types), here I added .s file types but it didn't solve my problem.

post118512: Re: Using dynamic library in QNX7

Albrecht Uhlmann — Thu, 01 Feb 2018 18:10:16 GMT

The problem must be related to the linker, not header files, because if they were not found then you would get a compiler error.

If vsomeip is a library, then surely it must be somewhere in the system, something like "libvsomeip.so.1" or so. You have to tell the linker that you explicitly wish it to be included in the link. Setting the shared library path alone is not sufficient. You can do this in the properties of the project, under the "QNX C/C++ Project" entry, tab "Linker", page "Libraries". Note that the preceding "lib" muat not appear there.

Regards,

Albrecht

post118511: Re: Inclusion of assembly file in QNX C project in Momentics IDE

Albrecht Uhlmann — Thu, 01 Feb 2018 18:00:53 GMT

I believe that the file extension has to be a capital S. Does this work?

Regards,

Al

post118506: Inclusion of assembly file in QNX C project in Momentics IDE

Girish Deshmukh(deleted) — Thu, 01 Feb 2018 13:47:03 GMT

Hello,

I am trying to build a QNX C project which includes .c files and .s files (.s files is an assembly code), but Momentics IDE is unable to build it successfully. It is generating obj file for .c files but it is not generating file for .s files.

Can anyone help?

post118496: Using dynamic library in QNX7

Ahmed Samir Khalil(deleted) — Wed, 31 Jan 2018 08:39:49 GMT

Hello,

I get started with vsomeip library which is built & provided by QNX, & I am using QNX7. A simple application "service-example.cpp" is used as jump-start, it is attached & already available under the following link:

https://github.com/GENIVI/vsomeip/wiki/vsomeip-in-10-minutes

The problem is... When building the application, I get the following error message:

"undefined reference to vsomeip::runtime::get()"

* More details about the build environment:

- I am using QNX executable > C/C++

- Shared library path is already set, via Window > Preferencess > QNX > Shared Libraries

- The header files are copied to: /home/user/qnx700/target/qnx7/usr/include

- Noting that: For QNX C/C++ executable, there is no linker options -l nor -L ?!

- When commenting out the line of creating the service

app = vsomeip::runtime::get()->create_application("Hello");

the program is built fine !

- the namespace of vsomeip is used: using namespace vsomeip;

- The main header file is Included: #include <vsomeip/vsomeip.hpp>

Thank you.

Best regards,

Ahmed

post118148: Re: Memory Errors tab not populating with any output data

Elena Laskavaia — Wed, 25 Oct 2017 15:18:34 GMT

Example you gave is not a leak per se, it depends what you do with str pointer

char* str = new char[5];

str=null;

That would be a leak

post118145: Memory Errors tab not populating with any output data

David Lypka(deleted) — Tue, 24 Oct 2017 14:07:52 GMT

I am a QNX novice so I am trying to learn to use the various Memory diagnostic tools.

I am trying to learn to use the Memory Errors tool first.

I followed the Help ... Welcome .. 10 Steps to developing a QNX program. Quick Start Guide. So I created a new QNX C++ project (NOTE I selected C++ instead of C) which generates a default C++ program.

I connected to Neutrino running in the VMWare sample VM and it connects properly and I can display the System Info.

I created a C/C++ QNX Qconn(IP)s Launch Configuration

In the C++ main I added some code to create a memory leak.

My code to create a memory leak in the my_first_project.cc is:

char* str = new char[5];

str[0] = 65;

Then I followed the steps to activate the Memory Errors Checker.

On the Debug Perspective I selected Debug Configurations ... selected my_configuration .. selected the Tools tab and selected the Memory Errors item.

When I run the application under Debug, the memory errors output tabs never populates with any data.

What could be the problem?

post118082: Re: Bind all processes to a single core by default.

Mike Lorenz — Wed, 27 Sep 2017 08:34:24 GMT

Maybe you can use the command 'on' with your specific runmask ( option '-R' and '-C' ) for starting your processes.

post118040: Re: Bind all processes to a single core by default.

Alastair Hoyle(deleted) — Mon, 18 Sep 2017 08:26:24 GMT

Thank you for your reply. I was experimenting with running our application code on different process cores. I realised I was easily able to lock our application to one or more cores, but this does not allow us exclusive use of those cores. I was wondering if it was possible to have exclusive use of one or more cores. From your reply, it sounds as if this might be possible but not simple. While it would make a nice and clean design if we were able to assign 3 of the 4 cores on our processor for exclusively for our use, in reality it is not causing us an issue at the moment. If this does become a problem, then I will investigate your suggestion in more detail.

Thank you for your time

Alastair

post118036: Re: Bind all processes to a single core by default.

Albrecht Uhlmann — Fri, 15 Sep 2017 09:54:28 GMT

You could try to write a system boot script that launches all your drivers and system tasks. You invoke that script with on -C0 <your core 0 boot script.sh>. To the best of my knowledge, the binding on core 0 will be inherited. As a last point in that script, you could call you application starter script like this: on -C1 -C2 -C3 <your application start script.sh>.

Regarding the interruption you are seeing and trying to resolve, not sure if this binding will solve all your needs, at least not without more core migrations. Maybe using priorities (perhaps try SCHED_SPORADIC) or APS is a better choice. It depends on the timing resolution that you require.

Regards,

Albrecht

post118020: Re: RE: QNX Code Coverage - HTTP ERROR 500

Murugaiyan Perumal(deleted) — Thu, 07 Sep 2017 14:48:22 GMT

I can solve the http error 500 after following the work around solution suggested by QNX.

The problem can be a bug in the IDE that we noticed recently in the IDE 7.0. The code for 6.6 appears to have the same issue. Reference number is jira 2221099.

The issue seems to be that the name of the folder where the coverage project is stored on disk does not match the name of the coverage project in the IDE. The hope is to have this fixed by the next release of Momentics, but the workaround is to rename the coverage project in the IDE to match the name of its containing folder on disk.

Can this be your problem? Please make sure the IDE project name and the name on the disk match.

post117937: Re: RE: QNX Code Coverage - HTTP ERROR 500

Murugaiyan Perumal(deleted) — Wed, 02 Aug 2017 05:59:10 GMT

Hi Sunil,

Are you able to solve the problem. I'm facing same issue while generating the code coverage report.

Please post the solution, if you already found it.

Thanks,

Murugaiyan

post117926: Re: Building QNX C++ Project with linked resources

Sesh — Mon, 31 Jul 2017 07:27:57 GMT

Solved this issue by combining SRCVPATH and SRCS to selectively pick source files.

Using EXTRA_SRCVPATH to build entire directory.

post117918: Re: Building QNX C++ Project with linked resources

Sesh — Wed, 26 Jul 2017 21:12:18 GMT

Thanks Will.

As a first step i was able to selectively pick files and folders with a "C++ project" and linked resources and get this to work. I am still trying to figure out a solution for "QNX C++" projects as they do not seem to recognize linked sources during project builds.

post117917: Re: Building QNX C++ Project with linked resources

Will Miles — Wed, 26 Jul 2017 20:14:17 GMT

Hi Sesh,

I'm not an expert in the IDE, but I think you're right; I don't think the make engine looks at "Linked Resources" to select source files.

Maybe try this:

-> Right-clock on the project in the 'Project Explorer', choose 'Properties'

-> Click "QNX C/C++ Project"

-> Click the 'Compiler' tab

-> Change the 'Category' drop down at the top to "Extra Sources Paths"

-> Choose 'Project' to browse the workspaces; select the folder with your desired sources.

Note it is not possible to add just one source file this way -- only an entire folder.

You can do the same with include files on the 'Extra Include Paths' category.

Good luck,

-Will

post117914: Building QNX C++ Project with linked resources

Sesh — Wed, 26 Jul 2017 07:08:36 GMT

Hello,

We are using QNX 6.6. Currently our projects are built as QNX C++ projects for multiple platform. Due to some internal needs this single large project must be broken into a smaller set of sub-projects.

In the example shown below Project1 may pull in func2.cpp, func22.cpp and some files from /cmn. I used "Linked Resources" and managed to pick the specific files/folders of interest for a given project. This scheme does not build all the sources files in the case of QNX C++ project. On creating a "C++ Project" linked resources scheme works fine. Appreciate if someone can help on solving this for a "QNX C++ project". I am guessing "Linked Resouces" updates .project/.cproject but does nothing to makefiles.

Eg:

/cmn

/src

        /featureA

             func1.cpp, func2.cpp etc..

        /featureB

             func11.cpp, func22.cpp etc..

       etc...

/NewProj

       /Project1

            main.cpp

            .cproject

            .project

            common.mk

            Makefile

       /Project2

             main.cpp

            .cproject

            .project

            common.mk

            Makefile

post117762: Bind all processes to a single core by default.

Alastair Hoyle(deleted) — Wed, 24 May 2017 08:56:20 GMT

Is there any way to force all system processes and threads to use a single processor core?

I know I can use 'on' to set the runmask of a particular process and I can use ThreadCtl( _NTO_TCTL_RUNMASK_GET_AND_SET_INHERIT.. to bind particular threads. But is there any way to bind all process to a single core by default, preferably at boot time.

My aim is to get all the system tasks and drivers to run on one processor core, leaving the other cores free to run our real-time tasks. Note I DO want to use the all the processor cores, just I want to use of all but one of the cores exclusively for our application code. Currently I can see our application code is being interrupted by the system threads.

Thanks

Alastair

--------------------------------------------------------------------------------------------------

This email and any attachments are confidential and are for the use of the

addressee only. If you are not the addressee, you must not use or disclose the

contents to any other person. Please immediately notify the sender and

delete the email. Statements and opinions expressed here may not

represent those of the company. Email correspondence is monitored by

the company. This information may be subject to export control

regulation. You are obliged to comply with such regulations.

Renishaw plc (company number 1106260) and Wotton Travel Limited (company

number 01973158) are companies registered in England and Wales with a registered office

at New Mills, Wotton-under-Edge, Gloucestershire, GL12 8JR,

United Kingdom, Telephone +44 1453 524524.

--------------------------------------------------------------------------------------------------

post117712: QNX 6.5.0 support for both LE and BE on ARM926EJ-S

Heston Holtmann(deleted) — Thu, 04 May 2017 18:34:17 GMT

My company is using QNX 6.5.0 on a TI OMAP-L138 ARM926EJ-S based processor (SoC)

We are running the system in Little-Endian (LE) Mode... simply because that was the default build of QNX provided supporting that chip that was provided in the Momentics build system.

..

But I am wondering.. in hind-sight.. was there also a Big Endian (BE) native build of QNX 6.5.0 for the ARM926EJ-S

Did we have a choice all along to stick with Big Endian to minimizing porting complexity from our QNX Application and

protocol code that came from PowerPC (Big Endian).

Heston

post117568: Re: how can Photon library calls debugged in QNX Momentics IDE for a C/C++ QNX QConn (IP) debug session?

Folkert Meeuw(deleted) — Wed, 29 Mar 2017 09:49:05 GMT

@thaupt: snap4.bmp is the function souce code

post117563: how can Photon library calls debugged in QNX Momentics IDE for a C/C++ QNX QConn (IP) debug session?

Folkert Meeuw(deleted) — Tue, 28 Mar 2017 09:13:17 GMT

the (a)pplication (u)nder (t)est located on the target device starts remote from the C/C++ QNX QConn (IP) configuration inside the QNX Momentics IDE located on the virtual machine on the host. At this point the aut is still running, one thread only (snap1.bmp). then a test script starts (crash.sh). the script runs a loop, open a data input dialog, type input data, close input dialog, send data over bus (MVB). here is the point the aut crashs. the loop runs 33 times. with Thread [19] (Suspended: Signal 'SEGSEGV received. Description: Segmentation fault). the stack to observe is 1 CWidgetRenderer::setSelection() at ../../../TheLibPhLib++/WidgetRenderer.cpp:2054. the host system is QNX qnx630mf 6.3.0 2004/04/29-21:23:19UTC x86pc x86. the target is QNX mft11 6.3.0 2005/08/09-15:04:19EDI x86pc x86.

post117382: malloc overcommit

Guillermo Salas — Fri, 27 Jan 2017 19:09:46 GMT

We have a question regarding the behavior when the OS ran out of memory.

Linux system implements the behavior of "malloc overcommit", so when a process tries to call malloc but the machine doesn't have enough memory the OS can kill another process.

https://www.etalabs.net/overcommit.html

Is this the same in QNX?

If so, is it possible to configure like in Linux by editing /proc/sys/vm/overcommit_memory?

post116727: libxml2.so and libxml2.a doesn't include xmlNewProp symbol

Denis Kimcherenko(deleted) — Wed, 31 Aug 2016 07:48:12 GMT

Hi everyone,

My app uses libxml2. QNX SDP 6.5.0 includes LibXml2 version 2.9.1.

But a linker cannot find xmlNewProp.

I run the nm tool both on shared and static libraries and could not find this symbol.

Here is an output:

dyosick@dyosick-work:~/$ nm -aA /opt/qnx650/target/qnx6/x86/usr/lib/libxml2.so | grep xmlNewProp

/opt/qnx650/target/qnx6/x86/usr/lib/libxml2.so:00040512 t xmlNewPropInternal

dyosick@dyosick-work:~/$ nm -aA /opt/qnx650/target/qnx6/x86/usr/lib/libxml2.a | grep xmlNewProp

/opt/qnx650/target/qnx6/x86/usr/lib/libxml2.a:tree.o:00004360 t xmlNewPropInternal

Does any know what the problem is? Do I need to re-build LibXml2 by my own?

post116662: Re: SDP 650 with GCC 4.8.3 cannot recognize C++11 to_string()

Denis Kimcherenko(deleted) — Wed, 17 Aug 2016 07:54:49 GMT

See solution in http://community.qnx.com/sf/discussion/do/listPosts/projects.toolchain/discussion.core_development_tools.topc26761?_pagenum=4

post116655: SDP 650 with GCC 4.8.3 cannot recognize C++11 to_string()

Denis Kimcherenko(deleted) — Tue, 16 Aug 2016 15:37:44 GMT

Try to compile almost "Hello world" as QNX C++ project

#include <iostream>

#include <string>

int main(int argc, char *argv[]) {

std::cout << std::to_string(100500) << std::endl;

return EXIT_SUCCESS;

}

Got compilation error:

/home/dyosick/ide-4.7-workspace/TestToString/TestToString.cc: In function 'int main(int, char**)':

/home/dyosick/ide-4.7-workspace/TestToString/TestToString.cc:6:15: error: 'to_string' is not a member of 'std'

std::cout << std::to_string(100500) << std::endl;

Does anyone now how to enable the C++11 function to_string()?

I compiled with -Wc,std=c++11

post116528: Beagle Bone Black BSP - rebuilt problem

Christine Jakobs(deleted) — Mon, 11 Jul 2016 07:17:20 GMT

OK after trying hard for a long time I really need help on two problems.

I have a new BBB rev. C and downloaded the BSP for Neutrino 6.6. If I use the prebuilt image, the BBB (doesn't matter if stable or experimental version) boots as expected. If I try to rebuilt the image (out of the IDE or on command line), the image does built and the BBB boots up but it says "unable to start..." for all the services.

I looks like during the make process the binaries don't get the executable flag and therefore are not executable when transferred to the SD Card.

On the other hand I really don't know which build file I have to modify in order to change the image. If I understood right, I should modify the build file in the src/hardware/startup/boards/ti-am335x/beaglebone/build. But if i uncomment something in there and rebuilt the image, the resulting build file in the image folder again has this line commented out.

Please if someone can help me? Thanks in advance.

post116371: code coverage exclude system header files

Mallikarjun N(deleted) — Mon, 30 May 2016 18:40:44 GMT

I have generated gcno and gcda files required for code coverage.

After importing these files into QNX IDE, it shows code coverage for all system included header files also(like <vector>) whose values are very less %.

Is there any option/flags while build or run to remove system header files from code coverage data in QNX IDE.

post116191: Re: Caching of _msg_info.scoid.

Dave Nickerson(deleted) — Fri, 29 Apr 2016 19:39:00 GMT

>

> We assume the following regarding the scoid value:

>  - As connections are attached to, and detached from, the channel, a new

> connection will not have a previously used scoid associated with it.

The SCOID values should be re-used as soon as the client associated with it no longer has any CONNECTIONS.

Depending on flags set during the ChannelCreate(), specifically if _NTO_CHF_DISCONNECT is set, then SCOID are required to be freed by calling ConnectDetach(scoid) once you are done freeing up any lists associated with the scoid, in your case the state information you are keeping about each client.  Once the scoid has been freed, the scoid will be available for re-use.

>  - The scoid is never re-assigned while connections are attached. That is if

> Client1 performs a ConnectAttach(...) to the server, then every time a message

>  is received from Client1 the scoid will always be the same, regardless of how

>  many other clients perform ConnectAttach( ... ), Send( ...)  ConnectDetach( .

> .. ).

This is true.

>

> Please let me know if the above assumptions are valid and what is the

> behaviour of QNX regarding assignment and re-use of scoids.

>

> In addition, suppose that the client is multi-threaded, and each thread

> performs ConnectAttach( ...) then Send (... ). In the server, when the message

>  from each thread is received, will a different scoid be associated with each

> thread?

The scoid uniquely identifies a client attached to a server.  As long as the client has a valid connection, (and thus a scoid), other threads within the client process calling ConnectAttach() will not generate a new scoid.

post116115: Caching of _msg_info.scoid.

Alex Clavel — Fri, 15 Apr 2016 21:31:10 GMT

In our application a server manages stateful connections to several clients. The clients always initiate communication with a send, and the server only replies. In the server, when we receive a message from a client, we identify which client the message originated from by using MsgInfo(...) and obtaining the struct _msg_info.scoid. The scoid is used to keep track of the connection state.

We assume the following regarding the scoid value:

- As connections are attached to, and detached from, the channel, a new connection will not have a previously used scoid associated with it.

- The scoid is never re-assigned while connections are attached. That is if Client1 performs a ConnectAttach(...) to the server, then every time a message is received from Client1 the scoid will always be the same, regardless of how many other clients perform ConnectAttach( ... ), Send( ...) ConnectDetach( ... ).

Please let me know if the above assumptions are valid and what is the behaviour of QNX regarding assignment and re-use of scoids.

In addition, suppose that the client is multi-threaded, and each thread performs ConnectAttach( ...) then Send (... ). In the server, when the message from each thread is received, will a different scoid be associated with each thread?

post115899: Re: Assign the proper index for LAN ports

Sean Boudreau(deleted) — Thu, 03 Mar 2016 14:55:19 GMT

You can try:

# ifconfig wm3 name foo0

On Thu, Mar 03, 2016 at 12:44:49AM -0500, Marcus Liang wrote:

> Dear all,

>

> I'm working on QNX6.6 and I got a device with 4 PCI LANs, 1 is Intel i218 and the rests are Intel I210, so all the NICs use devnp-e1000 driver.

>

> After the "io-pkt--v6-hc -de1000" executes, all 4 ports can be recognized and work.

>

> However the index are not same with the label on the device shell. Here is what I got:

>

> Label on the shell ~ ifconfig shows

> LAN A ~ wm3

> LAN B ~ wm2

> LAN C ~ wm0

> LAN D ~ wm1

>

> Is it possible modified the name/index after driver runs? Or I can assign the proper name/index when the driver run?

> I want it to be like:

> LAN A ~ wm0

> LAN B ~ wm1

> LAN C ~ wm2

> LAN D ~ wm3

>

> Thanks!

>

>

>

> _______________________________________________

>

> QNX Momentics Community Support

> http://community.qnx.com/sf/go/post115895

> To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post115895: Assign the proper index for LAN ports

Marcus Liang(deleted) — Thu, 03 Mar 2016 05:44:49 GMT

Dear all,

I'm working on QNX6.6 and I got a device with 4 PCI LANs, 1 is Intel i218 and the rests are Intel I210, so all the NICs use devnp-e1000 driver.

After the "io-pkt--v6-hc -de1000" executes, all 4 ports can be recognized and work.

However the index are not same with the label on the device shell. Here is what I got:

Label on the shell ~ ifconfig shows

LAN A ~ wm3

LAN B ~ wm2

LAN C ~ wm0

LAN D ~ wm1

Is it possible modified the name/index after driver runs? Or I can assign the proper name/index when the driver run?

I want it to be like:

LAN A ~ wm0

LAN B ~ wm1

LAN C ~ wm2

LAN D ~ wm3

Thanks!

post115875: Re: PPS notification group problem

Roger Maclean — Fri, 26 Feb 2016 20:11:02 GMT

..notify is not for selecting attributes to be notified about, it is as an

alternative for ionotify/poll/select, i.e. as a means of discovering which

of several fds have data available. When you open a pps object using

notify=nnn:mmm  the nnn part is for pps to determine which fd it should

notify and mmm is for the client to know which fd it is being notified

about.  I don¹t know that there is any reason to use .notify any longer, I

think it is there for purely historical reasons.

If you only want to get the Œid' attribute from a pps object, you¹d open:

   /pps/crank/test?f=id

f = filter.  I don¹t recall if you can filter more than one attribute.

On 2016-02-26, 2:48 PM, "Brian Edmond" <community-noreply@qnx.com> wrote:

>I have tried using a PPS notification group and they do not seem to work

>as expected.  I created the following file:

>

>/pps/crank/test

>

>Then use this code:

>

>#include <stdlib.h>

>#include <stdio.h>

>#include <string.h>

>#include <fcntl.h>

>#include <sys/pps.h>

>

>int main()

>{

>    char noid[16], buf[256];

>    int notify_fd, fd1;

>    int nread;

>

>    notify_fd = open("/pps/.notify", O_RDONLY);

>    read(notify_fd, &noid[0], sizeof(noid));

>    noid[strlen(noid)-1] = 0;

>    printf("Open Notify fd: <%s>\n", noid);

>

>    sprintf(buf, "/pps/crank/test?notify=%s:id", noid);

>    fd1 = open(buf, O_RDONLY);

>    printf("Open: %s\n", buf);

>

>    memset(buf, 0, sizeof(buf));

>    while((nread=read(notify_fd, &buf, sizeof(buf))) >= 0) {

>        if (nread) {

>            printf("Notify (%d): %s\n", nread, buf);

>        }

>        memset(buf, 0, sizeof(buf));

>    }

>    printf("Done: %d\n", nread);

>    return 0;

>}

>

>So a notification group is used to monitor the "id" attribute of the test

>object.  Then I write to the object:

>    echo "id::1" >> /pps/crank/test

>Now I see the above "Notify" message as I expect.

>Now write to the object again but with a different attribute:

>  echo "abc::1" >> /pps/crank/test

>Now I still see the above "Notify" message.  However I did not change the

>"id" value so why does read return like I did?

>

>Thanks,

>Brian

>

>

>

>_______________________________________________

>

>QNX Momentics Community Support

>http://community.qnx.com/sf/go/post115874

>To cancel your subscription to this discussion, please e-mail

>momentics-community-unsubscribe@community.qnx.com

post115874: PPS notification group problem

Brian Edmond(deleted) — Fri, 26 Feb 2016 19:48:34 GMT

I have tried using a PPS notification group and they do not seem to work as expected.  I created the following file:

/pps/crank/test

Then use this code:

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

#include <fcntl.h>

#include <sys/pps.h>

int main()

{

    char noid[16], buf[256];

    int notify_fd, fd1;

    int nread;

    notify_fd = open("/pps/.notify", O_RDONLY);

    read(notify_fd, &noid[0], sizeof(noid));

    noid[strlen(noid)-1] = 0;

    printf("Open Notify fd: <%s>\n", noid);

    sprintf(buf, "/pps/crank/test?notify=%s:id", noid);

    fd1 = open(buf, O_RDONLY);

    printf("Open: %s\n", buf);

    memset(buf, 0, sizeof(buf));

    while((nread=read(notify_fd, &buf, sizeof(buf))) >= 0) {

        if (nread) {

            printf("Notify (%d): %s\n", nread, buf);

        }

        memset(buf, 0, sizeof(buf));

    }

    printf("Done: %d\n", nread);

    return 0;

}

So a notification group is used to monitor the "id" attribute of the test object.  Then I write to the object:

    echo "id::1" >> /pps/crank/test

Now I see the above "Notify" message as I expect.

Now write to the object again but with a different attribute:

  echo "abc::1" >> /pps/crank/test

Now I still see the above "Notify" message.  However I did not change the "id" value so why does read return like I did?

Thanks,

Brian

post115750: Support for Skylake

James VanOeffelen(deleted) — Tue, 16 Feb 2016 22:24:30 GMT

Any news on support of the Intel Skylake chipset (6th gen iCore) for QNX 6.5 SP1?

post115614: LAN 9220

Francisco Junior(deleted) — Tue, 26 Jan 2016 14:14:10 GMT

Hi,

I'm looking to use an SMC LAN 9220 with a DM3730 (similar to Mistral board). I searched thorugh hardware database in QNX but I didn't find anything.

There is a driver for this IC for QNX?

Regards,

post115531: Problem with including the standard libs

Ivan Angelov — Fri, 15 Jan 2016 15:21:21 GMT

Hi,

I want to send messages through mqueue, but when I run my project mq_send() gives me errno 89. I found that qnx didn't start mqueue service and when I started it manually this seemed to fix the problem. But... I made another project almost identical to the first one and here's what happened. When I try to include standard libs to my project I get this "Unresolved inclusion", but still my project build and runs without problem, almost. Again mq_send gives me the same errno but with the mqueue service running and I'm thinking that "unresolved inclusion" is the core of the problem( not sure though), because the first project is including the same libs without giving me the unresolved thing and running without problem.

Thank you!

post115249: Need Mouse Pointer Without Connecting Mouse

Asheesh Jain(deleted) — Wed, 09 Dec 2015 09:55:34 GMT

Dear ALL,

Is is possible in QNX to get pointer of mouse without connecting the mouse.

It is possible in widows but how to do the same in QNX i don't know.

Kindly reply.

post115016: BSP for Freescale T1040

hamed davaneh(deleted) — Mon, 16 Nov 2015 06:51:53 GMT

Hi all,

There is a BSP for Freescale P5020 (P5020DS), the T1040 is a processor from same family with same PowerPC architecture (e5500). I wonder how much modification needed to use this available BSP for both processors (assuming both have same board design with same peripherals).

post114588: Re: windows host: qnx command line commands not functioning

Steve Reid(deleted) — Mon, 05 Oct 2015 12:56:14 GMT

You probably need to set up the environment in your shell. Run qnx660-env.bat (which you'll find in the directory where you installed 6.6).

post114567: windows host: qnx command line commands not functioning

claude hasler(deleted) — Fri, 02 Oct 2015 11:22:00 GMT

Hi im trying to build a qnx image with the mkifs utlity provided by the SDP 6.6. However, when I try to run any command from the terminal I receive the follwing error message:

qconfig for instance:

qconfig : The term 'qconfig' is not recognized as the name of a cmdlet, function, script file, or operable program.

Check the spelling of the name, or if a path was included, verify that the path is correct and try again.

What do I need to do to get this going?

Cheers

Claude

post114534: including .h/.cpp pair into Momentics

claude hasler(deleted) — Mon, 28 Sep 2015 08:45:37 GMT

Hi, im trying to use an external library (tinyXML2) and have the .h/.cpp files but i cant figure out how to include these into my qnx momentics project

cheers

Claude

post113866: Re: Installing 6.50 SP1 breaks make (can't find qconfig.mk)

Ekta Sachdev — Wed, 27 May 2015 14:44:26 GMT

If you look through the install logs (/opt/qnx650/installLog.txt), double-check whether you're getting an error: "/opt/qnx650/install/qnxsdp/6.5.0SP1/post_install.sh already exists". I had the same problem, and it all narrowed down to the 'post_install.sh' file already being present, because of previous faulty installation cycles. The installer GUI doesn't show any errors if the 'post_install.sh' already exists and it seems to install perfectly, but the logs are filled with errors.

There are two ways to fix it:

(1) Use the cleanup utilities and remove all files associated with SDP6.5.0. Start from scratch and don't miss out on the sudo -E [if using Ubuntu].

(2) Remove the /opt/qnx650/install/post_install.sh file and /opt/qnx650/install/qnxsdp/6.5.0SP1 directory. If you install the service pack after that, it should install properly.

Hope this helps

post113731: RESOURCE MANAGER frame work development

Mythili R(deleted) — Tue, 21 Apr 2015 06:47:36 GMT

Hi, I am working in the resouce manager(multi threaded) framework development. Here the two client processess accessing the

resource manager buffer(i.e reader and writer clients). when we terminate the client process first and then resource

manager process means thread pool destroyed properly and resource manager is detached successfully. if the resource

manager process terminated unfortunately while the client processes are running , the resource manager cant be detached.

sometimes thread pool destroy itself failed. could you please suggest the solution...

post113662: Re: select does not unblock when iofunc_notify_trigger invoked from a relatively low priority thread

Sean Boudreau(deleted) — Wed, 01 Apr 2015 10:37:03 GMT

Sigwaitinfo is normal.

Sent from my BlackBerry 10 smartphone on the Rogers network.

Original Message

From: Malte Mundt

Sent: Wednesday, April 1, 2015 5:52 AM

To: momentics-community

Reply To: momentics-community@community.qnx.com

Subject: Re: select does not unblock when iofunc_notify_trigger invoked from a relatively low priority thread

Hi,

I'm not sure this question can be easily answered unless you provide a little test case. What makes me wonder though is that you mention that your thread is blocked in the state of SigWaitInfo. Is this really the state a thread goes into when blocking on a select() call?

Regards,

- Malte

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post113660

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113660: Re: select does not unblock when iofunc_notify_trigger invoked from a relatively low priority thread

Malte Mundt — Wed, 01 Apr 2015 09:14:05 GMT

Hi,

I'm not sure this question can be easily answered unless you provide a little test case. What makes me wonder though is that you mention that your thread is blocked in the state of SigWaitInfo. Is this really the state a thread goes into when blocking on a select() call?

Regards,

- Malte

post113554: select does not unblock when iofunc_notify_trigger invoked from a relatively low priority thread

SENTHIL RAJAKKANNU(deleted) — Sun, 15 Mar 2015 19:30:49 GMT

Observed in QNX 6.5.0 running on in house hardware with EP9301 Cirrus logic SoC.

The problem described here involved two threads ( a Server and Client) with threads as mentioned below

A server with two threads

   Thread 1 running resource manger message loop to handle request from client process at priority 11

   Thread 2 runs periodically (every 2 seconds) and populates data in a queue and notifies the client by calling iofunc_notify_trigger()

A client with single thread running at priority 11

   This thread runs a loop to read data from server when data is available. This thread blocks on a select call while waiting for server to produce the data for reading.

Problem:

======

   When the client process is started, the select() returns immediately as long as data is available on server queue and the client is able to read the data.

   When the server queue is empty, the client thread blocks on select().

   At this point, the server Thread 2 add data to server queue and notify the client by calling iofunc_notify_trigger().

   The client thread gets unblocked and it again blocks on select() after reading the available data.



   Now, the next iofunc_notify_trigger() from server Thread 2 does not unblock the client thread from select() and it continue to wait on SigWaitInfo state.

   (NOTE: The call to IOFUNC_NOTIFY_INPUT_CHECK(notifylist,1,0) just before the iofunc_notify_trigger() returned true)

   All subsequent check to IOFUNC_NOTIFY_INPUT_CHECK(notifylist,1,0) returns false.

   At this point the server is waiting for client to arm using IO_NOTIFY client message and the client it waiting for server to notify_trigger.

Observations:

=========

    1) If I change the server Thread 2 priority to 11 or more then the problem does not occur.

        I tried with different priorities for client thread and server Thread 2, it works only when the server Thread 2 priority is greater than or equal to the client thread priority.



    2) If I add timeout to select(), the client gets unblocked after timeout and it is able to detect the data availability on next select call and read the data successfully.

       After reading all available accumulated data it goes to blocked state and the problem happens again.

Question:

======

I could not find any mention about the expected priority for a thread  invoking iofunc_notify_trigger (or MsgDeliverEvent) method in the QNX documentation.

How does the priority of the thread invoking the iofunc_notify_trigger() method cause select to block indefinitely?

If priority is the problem, how does it getting unblocked for the first time?

post113325: Re: RE: How do I tell which version of a shared object .so the linker used?

Kevin Schaffer(deleted) — Wed, 18 Feb 2015 15:36:45 GMT

Ok, I figured out what was happening.

I was (I am pretty sure) linking correctly against the self-contained sqlite3 .so we build (I assume??? the linker "checks out" that one of the listed .so actually contains the functions the main executable will want to use.)

But on the target device, the old QNX-supplied sqlite3.so was installed, so when runtime rolled around, the main executable tried to use that one, which failed because further dependent .so down the line were not on the device.

I'm still a little hazy on what the linker does to verify the .so contains what's needed by the main executable being linked, but it looks like something like that. And at runtime, it has the other .so, crosses its fingers, and makes a go of it, choking gracefully.

post113324: RE: How do I tell which version of a shared object .so the linker used?

Max Feil — Wed, 18 Feb 2015 10:08:38 GMT

It sounds like your LD_LIBRARY_PATH is not set correctly at runtime. To trace exactly where shared libs are being picked up from at runtime, set environment variable DL_DEBUG to "libs", for example:

# DL_DEBUG=libs sqlite3

Max

P.S. You can use the -rpath linker option at build time to override LD_LIBRARY_PATH, but that can make things confusing later. I use "objdump -p" to see the rpath info in an executable or shared object.

-----Original Message-----

From: Ryan Mansfield [mailto:community-noreply@qnx.com]

Sent: Thursday, February 12, 2015 5:37 PM

To: Kevin Schaffer; momentics-community

Subject: Re: How do I tell which version of a shared object .so the linker used?

Readelf -d or objdump -x will ‎list the soname of the shared object linked against.

If at link time, you can specify the  -Wl,-t option to trace the shared objects linked a‎gainst.

Regards,

Ryan Mansfield

  Original Message

From: Kevin Schaffer

Sent: Thursday, February 12, 2015 5:22 PM

To: momentics-community

Reply To: momentics-community@community.qnx.com

Subject: How do I tell which version of a shared object .so the linker used?

I have an executable that's generating an error at runtime:

lddFATAL Could not load library libicui18n.so.49

This is because the linker is linking it to point to the libsqlite3.so that comes with QNX 6.6.  However, we have our own self-contained libsqlite3.so we've been using for just such a reason.

How can I tell which .so the linker chooses to dynamic-link against.  This can either be at linktime or later with, say, the objdump utility.

Note:  "our" libsqlite3.so is built and is in the link path (in fact, it is the very first option on the link line, -L(path to our libsqlite3.so directory) when we get around to linking the executable, which is the very last thing we do.  There is no build/link of the libsqlite3.so after that; it occurs much earlier.

We also use this in the link command:

-Wl,-Bdynamic -lsqlite3

So why is it going into the QNX libraries before going into our -L(custom path)?

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post113307

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post113308

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113322: Re: QNX driver File creation

Dave Nickerson(deleted) — Tue, 17 Feb 2015 19:43:40 GMT

See "Writing a Resource Manager" in our developers product documentation.

The most common method to register a resource name in the dev pathname space is:

resmgr_attach()

Lots of background for that call and what you have to be able to handle after you make that call are documented in the above reference.

post113321: QNX driver File creation

anil sharma(deleted) — Tue, 17 Feb 2015 17:17:57 GMT

Can you please let me know the how device file which are shown in dev dir in file system are created ? I wanted to write driver for SPI and wanted to create the device node in dev dir,

Can you please help me in this case

post113308: Re: How do I tell which version of a shared object .so the linker used?

Ryan Mansfield(deleted) — Thu, 12 Feb 2015 22:37:30 GMT

Readelf -d or objdump -x will ‎list the soname of the shared object linked against.

If at link time, you can specify the  -Wl,-t option to trace the shared objects linked a‎gainst.

Regards,

Ryan Mansfield

  Original Message

From: Kevin Schaffer

Sent: Thursday, February 12, 2015 5:22 PM

To: momentics-community

Reply To: momentics-community@community.qnx.com

Subject: How do I tell which version of a shared object .so the linker used?

I have an executable that's generating an error at runtime:

lddFATAL Could not load library libicui18n.so.49

This is because the linker is linking it to point to the libsqlite3.so that comes with QNX 6.6.  However, we have our own self-contained libsqlite3.so we've been using for just such a reason.

How can I tell which .so the linker chooses to dynamic-link against.  This can either be at linktime or later with, say, the objdump utility.

Note:  "our" libsqlite3.so is built and is in the link path (in fact, it is the very first option on the link line, -L(path to our libsqlite3.so directory) when we get around to linking the executable, which is the very last thing we do.  There is no build/link of the libsqlite3.so after that; it occurs much earlier.

We also use this in the link command:

-Wl,-Bdynamic -lsqlite3

So why is it going into the QNX libraries before going into our -L(custom path)?

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post113307

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113307: How do I tell which version of a shared object .so the linker used?

Kevin Schaffer(deleted) — Thu, 12 Feb 2015 22:22:11 GMT

I have an executable that's generating an error at runtime:

lddFATAL Could not load library libicui18n.so.49

This is because the linker is linking it to point to the libsqlite3.so that comes with QNX 6.6.  However, we have our own self-contained libsqlite3.so we've been using for just such a reason.

How can I tell which .so the linker chooses to dynamic-link against.  This can either be at linktime or later with, say, the objdump utility.

Note:  "our" libsqlite3.so is built and is in the link path (in fact, it is the very first option on the link line, -L(path to our libsqlite3.so directory) when we get around to linking the executable, which is the very last thing we do.  There is no build/link of the libsqlite3.so after that; it occurs much earlier.

We also use this in the link command:

-Wl,-Bdynamic -lsqlite3

So why is it going into the QNX libraries before going into our -L(custom path)?

post113271: QNX 6.5 SP 1 usb utility vendor name miss match

janusz ruszel — Mon, 09 Feb 2015 16:56:04 GMT

More details. The uhci driver seems show it right but ehci does not.

post113270: QNX 6.5 SP 1 usb utility vendor name miss match

janusz ruszel — Mon, 09 Feb 2015 16:31:18 GMT

The 0x413c vendor Id is listed as (Samsun). The company id list from usb.org states 0x413c Dell.

post113267: QNX 6.5 SP 1 and USB PCL printers

janusz ruszel — Mon, 09 Feb 2015 13:56:30 GMT

Listed below printers were tested. All of them are PCL printers and data transmitted over USB was in PCL format. I have also tested each usb driver separately uhci/ohci and data corruption have happened. My testing environment is commercial QNX 6.5 SP1 on x86 platform.

Later on I grab the devu-prn source code which was published with DDK 6.3.2 and recompiled with QNX 6.5 SP1 environment. Repeated all test and data corruption still was noticed.

Next step was to making devu-prn not issue the null packet during prn_io_close() and data corruption “disappeared”.

The same test run under evaluation version QNX 6.6 also corrupts the data.

This is repeatable scenario and I am sure could be repeated by QNX driver engineers.

If you have any suggestions/questions or would like me to run more tests I would be glad to help.

Tested Printers : (over USB interface)

Dell B1260

Samsung M2835DW

Samsung SL-M3320ND

Janusz

post113262: RE: QNX 6.5 SP 1 and USB printers

Max Feil — Mon, 09 Feb 2015 06:15:46 GMT

This problem (or one exactly like it) was fixed in 6.4.0, under PR43795. The problem was not actually in devu-prn, but rather in devu-ehci.

You are right that making devu-prn not issue the null packet during prn_io_close() does seem to fix the problem, but this null packet is needed in certain situations. The real problem was in devu-ehci which was not properly updating the data toggle when processing a zero length packet.

If you need more details, let me know,

Max

P.S. I'm a bit confused about releases. You mention the 6.3.2 ddk and also 6.5 SP1. This problem should be long fixed in 6.5, but it will be present in 6.3.2. You can try narrowing things down more by using uhci/ohci instead of ehci and seeing if the problem goes away. There may have been other problems fixed in devu-prn that I'm not aware of. See also PR58292 and the more recent comments from 2011.

-----Original Message-----

From: janusz ruszel [mailto:community-noreply@qnx.com]

Sent: Sunday, February 08, 2015 4:59 PM

To: momentics-community

Subject: Re: QNX 6.5 SP 1 and USB printers

Issue located.

the devu-prn driver terminates the printer transfer with a null Bulk Out packet (prn_io_close function) which causes some printers to corrupt next coming data. Linux generic usb printer drivers do not issue additional null data transfer.

The source code i am looking at is from DDK 6.3.2. Is the source code for QNX 6.6 different?

Janusz

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post113261

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113261: Re: QNX 6.5 SP 1 and USB printers

janusz ruszel — Sun, 08 Feb 2015 21:59:09 GMT

Issue located.

the devu-prn driver terminates the printer transfer with a null Bulk Out packet (prn_io_close function) which causes some printers to corrupt next coming data. Linux generic usb printer drivers do not issue additional null data transfer.

The source code i am looking at is from DDK 6.3.2. Is the source code for QNX 6.6 different?

Janusz

post113256: QNX 6.5 SP 1 x86 platform and USB printers

janusz ruszel — Fri, 06 Feb 2015 23:24:44 GMT

I am having data transfer issue over usb interface to DELL and Samsung

printers. The transferred data seems to be corrupted or missing. Those two

printers also have Ethernet ports and transferring the same date over Ethernet

  interface does not cause data corruption. The format of transferred data is a

  pcl language and both printers support it.

I have used lpd to print to the printer as well cp (copy ) command and in both

  cases the data seems to be randomly corrupted. The first transfer after

printer is powered on is always successful.

The model of printer is Samsung M3320ND.

My guess is that devu-prn driver may be a problem. Any input or possible work

around would be very appreciated.

Best

Janusz

post113255: QNX 6.5 SP 1 and USB printers

janusz ruszel — Fri, 06 Feb 2015 23:21:41 GMT

I am having data transfer issue over usb interface to DELL and Samsung printers. The transferred data seems to be corrupted or missing. Those two printers also have Ethernet ports and transferring the same date over Ethernet interface does not cause data corruption. The format of transferred data is a pcl language and both printers support it.

I have used lpd to print to the printer as well cp (copy ) command and in both cases the data seems to be randomly corrupted. The first transfer after printer is powered on is always successful.

The model of printer is Samsung M3320ND.

My guess is that devu-prn driver may be a problem. Any input or possible work around would be very appreciated.

Best

Janusz

post113162: io-usb-dcd server (QNX6, Sabrelite platform)

Michael Dinkelman(deleted) — Fri, 30 Jan 2015 21:57:55 GMT

When I try to start the io-usb-dcd server, it looks good:

# io-usb-dcd -dusbser-mx6sabrelite-ci ioport=0x02184100,irq=73

until I look at the system log where irq is logged as an unknown option

# sloginfo

Time             Sev Major Minor Args

Jan 01 00:00:08    2    12     0 mx6sabrelite_extra_process_args_callout : unknown option - irq=73

Jan 01 00:00:08    2    12     0 io-usb : Error Initializing Host Controller

However, if I run the server command without the irq option:

# io-usb-dcd -dusbser-mx6sabrelite-ci ioport=0x02184500

It knows something is missing as I get:

Must specify irq with ioport option

I feel like I am chasing my tail. Any clues?

thanks

post113120: How to write Applications for QNX Connected Car Hardware

shiva aetherpal(deleted) — Wed, 28 Jan 2015 20:37:53 GMT

I am planning to write applications to run on QNX Connected car. I already setup QNX SDP and

QNX target on my machine. But I can not still get what else I need to proceed further.

1. which IDE should I use for this? QNX Momentics or Eclipse or something else?

2. Do I need to buy a QNX hardware board for connected car development, or Target created on VMware would be sufficient?

3. Do I need to purchase any license for connected car development?

Thanks

Shiva

post113119: How to write Applications for QNX Connected Car Hardware

shiva aetherpal(deleted) — Wed, 28 Jan 2015 19:01:04 GMT

I am planning to write applications to run on QNX system, which IDE should I use for this? I already setup QNX SDP and QNX target on my machine.

Thanks

Shiva

post113074: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

Nicola Vulpe(deleted) — Fri, 23 Jan 2015 22:27:57 GMT

Hi Shiva,

Glad to hear you got your system going. You are most welcome. Have fun!

N

On 2015-01-23 17:17 , "shiva aetherpal" <community-noreply@qnx.com> wrote:

>Now I am able to create a Target successfully. I do not know what is the

>issue, all the time I was following the correct procedure, but some thing

>was going wrong. After several trials I could install it properly and run

>my test applications.

>

>Thanks for your replies Nicola, they was helpful.

>

>

>

>

>_______________________________________________

>

>QNX Momentics Community Support

>http://community.qnx.com/sf/go/post113073

>To cancel your subscription to this discussion, please e-mail

>momentics-community-unsubscribe@community.qnx.com

post113073: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Fri, 23 Jan 2015 22:17:38 GMT

Now I am able to create a Target successfully. I do not know what is the issue, all the time I was following the correct procedure, but some thing was going wrong. After several trials I could install it properly and run my test applications.

Thanks for your replies Nicola, they was helpful.

post113068: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Fri, 23 Jan 2015 17:57:29 GMT

Hi Nicola,

I can not find any thing with related to "Working with Target Images" on

Foundry27. I am on evaluation 30-day trial version, is that the reason I am

not able access all the documents or I can not configure a Target?

This is the only link I found to install a Target for development. http://www.qnx.com/developers/articles/inst_5856_3.html

Thx

post113066: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Fri, 23 Jan 2015 17:29:32 GMT

post113065: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Fri, 23 Jan 2015 16:14:00 GMT

Thank you, I did exactly that!

post113064: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

Nicola Vulpe(deleted) — Fri, 23 Jan 2015 14:59:50 GMT

Hi Shiva,

Have you tried going to the Apps and Media 1.1 on Foundry 27? Look for the

document "Working with Target Images". Chapter one provides step-by-step

instructions for copying a reference image onto a target, including for

VMWare.

Have fun.

N

On 2015-01-22 16:45 , "shiva aetherpal" <community-noreply@qnx.com> wrote:

>Hi Nicola,

>

>yeah I have seen those documents. Here is what I am trying to do.

>

>I am planning to do QNX app development for car infotainment systems who

>are QNX enabled.

>

>From my basic research, I have to install QNX Software Development

>Program and QNX SDK for Apps and Media. and I have successfully done till

>this. I can also launch QNX Momentics IDE 5.0 and write code.

>

>Next part is to create a QNX target. I hope there are two ways to do that.

>1. Buy QNX board or device and connect to PC and start working as

>provided steps.

>2. Create VMware and install QNX RTOS on that slice, which will work as

>Target for deployment. (just like emulator).

>

>I want to create target as mentioned in the 2nd step. I tried to follow

>the steps mentioned in section "3. Installing the QNX Neutrino RTOS on

>the target system" @

>http://www.qnx.com/developers/docs/6.4.1/momentics/quickstart/about.html.

>But when I do that I get following error. Please find the attachment.

>

>Or please suggest me if there is any other way of setting up a QNX Target

>for installing apps.

>

>

>Thanks

>

>

>

>

>_______________________________________________

>

>QNX Momentics Community Support

>http://community.qnx.com/sf/go/post113041

>To cancel your subscription to this discussion, please e-mail

>momentics-community-unsubscribe@community.qnx.com

post113063: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Elena Laskavaia — Fri, 23 Jan 2015 14:06:20 GMT

You have to add path by adding -L<path> option to linker or adding

library to directory which is already on the path (pre-existing -L's or

one standard dirs in $QNX_TARGET)

On 15-01-09 10:21 AM, Kevin Schaffer wrote:

> This is a new .so from a 3rd party, so I would like to put it in a new folder. Then I would have to add this folder's path to the (development machine) search path, and then I could use -lsecondary?

>

>

>

> _______________________________________________

>

> QNX Momentics Community Support

> http://community.qnx.com/sf/go/post112894

> To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113062: Re: Can not add antyhing to my software

Elena Laskavaia — Fri, 23 Jan 2015 14:03:40 GMT

Well you have to debug your code... Put breakpoints where control

command are and attach the debugger. Or use print statements..

On 15-01-12 12:25 PM, Fabien Expert wrote:

> Hello to you all,

>

> First I want to apologize because I know that I will give a poor description of my problem but any idea to solve it (or things to check) would be greatly appreciated.

> I am currently developping a software in QNX 6.6.0 in C++ to move several motors in real time using Ethercat and it was going great until recently.

>

> Since a few days, each time I add anything to my soft, it compiles perfectly fine and does not show any problem during execution (the log file is perfect indicating that all the loops occur at the correct moment and all the threads are executing) but the motors do not move anymore. It seems to indicate that the part of the code where I send the data using Ethercat is not really executed but if I add a printf, it shows that the function is executed.

>

> Which is even weirder is that the same thing happen regardless of what I add to my code and where.

> I tried to use the tools to detect memory leak and trace the CPU usage and the threads but everything seems normal.

>

> I am not sure my problem is very clear but I would appreciate any insight that could give me ideas to test because I am currently stuck.

>

> Best regards,

> Fabien

>

>

>

> _______________________________________________

>

> QNX Momentics Community Support

> http://community.qnx.com/sf/go/post112909

> To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post113041: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Thu, 22 Jan 2015 21:45:58 GMT

Hi Nicola,

yeah I have seen those documents. Here is what I am trying to do.

I am planning to do QNX app development for car infotainment systems who are QNX enabled.

From my basic research, I have to install QNX Software Development Program and QNX SDK for Apps and Media. and I have successfully done till this. I can also launch QNX Momentics IDE 5.0 and write code.

Next part is to create a QNX target. I hope there are two ways to do that.

1. Buy QNX board or device and connect to PC and start working as provided steps.

2. Create VMware and install QNX RTOS on that slice, which will work as Target for deployment. (just like emulator).

I want to create target as mentioned in the 2nd step. I tried to follow the steps mentioned in section "3. Installing the QNX Neutrino RTOS on the target system" @ http://www.qnx.com/developers/docs/6.4.1/momentics/quickstart/about.html. But when I do that I get following error. Please find the attachment.

Or please suggest me if there is any other way of setting up a QNX Target for installing apps.

Thanks

post113030: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

Nicola Vulpe(deleted) — Thu, 22 Jan 2015 15:58:32 GMT

Hello Shiva,

QNX SDP is the QNX Software Development Platform. The QNX Neutrino

Realtime Operating System is the OS.

Perhaps you could start by reviewing some of the documentation at:

http://www.qnx.com/developers/docs/660/index.jsp

Look under QNX Software Development Platform, where you'll find OS Core

Components > System Architecture. This should give you some useful

information about how things fit together, and how you can get started.

Good luck,

Nicola Vulpe

On 2015-01-22 10:49 , "shiva aetherpal" <community-noreply@qnx.com> wrote:

>Can I install QNX SDP and QNX RTOS on the same machine?

>

>

>

>_______________________________________________

>

>QNX Momentics Community Support

>http://community.qnx.com/sf/go/post113029

>To cancel your subscription to this discussion, please e-mail

>momentics-community-unsubscribe@community.qnx.com

post113029: Re: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Thu, 22 Jan 2015 15:49:51 GMT

Can I install QNX SDP and QNX RTOS on the same machine?

post113016: Unable to create QNX RTOS with VM installation DVD using latest VMware Player

shiva aetherpal(deleted) — Wed, 21 Jan 2015 23:01:10 GMT

Hi I am trying to install QNX RTOS with VM installation DVD which I downloaded from "http://www.qnx.com/download/feature.html?programid=26217" using latest VMware Player 7. Please help me. I want to develop apps for QNX. But I am unable to create a target using VMware on Windows 7. Any clues greatly appreciated. Thanks.

post112909: Can not add antyhing to my software

Fabien Expert(deleted) — Mon, 12 Jan 2015 17:25:51 GMT

Hello to you all,

First I want to apologize because I know that I will give a poor description of my problem but any idea to solve it (or things to check) would be greatly appreciated.

I am currently developping a software in QNX 6.6.0 in C++ to move several motors in real time using Ethercat and it was going great until recently.

Since a few days, each time I add anything to my soft, it compiles perfectly fine and does not show any problem during execution (the log file is perfect indicating that all the loops occur at the correct moment and all the threads are executing) but the motors do not move anymore. It seems to indicate that the part of the code where I send the data using Ethercat is not really executed but if I add a printf, it shows that the function is executed.

Which is even weirder is that the same thing happen regardless of what I add to my code and where.

I tried to use the tools to detect memory leak and trace the CPU usage and the threads but everything seems normal.

I am not sure my problem is very clear but I would appreciate any insight that could give me ideas to test because I am currently stuck.

Best regards,

Fabien

post112894: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Fri, 09 Jan 2015 15:21:55 GMT

This is a new .so from a 3rd party, so I would like to put it in a new folder. Then I would have to add this folder's path to the (development machine) search path, and then I could use -lsecondary?

post112893: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Fri, 09 Jan 2015 15:20:35 GMT

Ok, I added a path to it after the -o (primary link file.o) ../whatever/secondary.so, and it worked fine.

This is a one-off for testing, but it may need to get integrated in the future, so setting it up as a library in a (new) library path is the eventual goal.

post112892: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Ryan Mansfield(deleted) — Fri, 09 Jan 2015 14:08:16 GMT

On 15-01-08 05:15 PM, Kevin Schaffer wrote:

> Ok, I tried

>       -Wl,-Bdynamic -lsecondaryLibrary

>

> and

>       -rpath-link=(pathWithoutSpacesToDevelopmentMachineLocationOfThatSO)

>

> and...cannot find -lsecondaryLibrary

>

>

> How can it not find it?  I gave it the path via -rpath-link.

If it's a dependency of your library and not your original executable,

then you shouldn't be explicitly linking against it by specifying

-lsecondaryLibrary.

-lsecondaryLibrary should be specified when you link libfirstlibrary,

but not a.out. If you specify -lsecondaryLibrary, ld will be looking for

it in the library search path (i.e directories specified by -L).

Regards,

Ryan Mansfield

post112891: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Jeff Baker — Fri, 09 Jan 2015 14:03:01 GMT

As an experiment, if you add the secondary .so to the link line directly as an input file, does that resolve the issue?

post112885: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Thu, 08 Jan 2015 22:40:57 GMT

Ok, this is just precioius.

I copied secondary.so into the same directory as primary.so, and this is the only spot in the entire build tree where either exist, AND IT STILL CANNOT FIND secondary.so.

I've dragged ld, kicking and screaming, into a single room, sawn open its eyeballs, sunk a mechanical claw into the back of its skull, and used a 30-ton robot arm to twist it's head to point its horrified face directly at secondary.so, and it still lies to me that it cannot see it.

post112884: Re: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Thu, 08 Jan 2015 22:15:09 GMT

Ok, I tried

     -Wl,-Bdynamic -lsecondaryLibrary

and

     -rpath-link=(pathWithoutSpacesToDevelopmentMachineLocationOfThatSO)

and...cannot find -lsecondaryLibrary

How can it not find it?  I gave it the path via -rpath-link.

From the -rpath-link description:

-rpath-link=dir

When using ELF or SunOS, one shared library may require another. This happens when an "ld -shared" link includes a shared library as one of the input files.

When the linker encounters such a dependency when doing a non-shared, non-relocatable link, it will automatically try to locate the required shared library and include it in the link, if it is not included explicitly.   <--------------------------------

So how could I "include it explicitely"?

Assume:

Executable

primary.so

secondary.so

where secondary.so contains function x(), which primary.so needs, but Executable does not (directly.)

Assume primary.so is in a path the linker finds when you specify Bdynamic -lprimary.

Assume secondary.so is in some funky directory, say, c:\a\b\c

Would I want to use -lsecondary?  Or could I just provide a path to secondary.so, including the name "secondary.so"?

(Cries) alls I wants to do is tell the linker here's the danged .so, here's where it is, and you will need this because it contains x(), which primary.so needs.

post112882: rpath or rpath-link doesn't resolve downstream .so needs?

Kevin Schaffer(deleted) — Thu, 08 Jan 2015 21:44:47 GMT

I'm linking an executable that links against a .so.  The .so has a function in it that's in another .so.

The link for the original executable chokes at this point, saying that it can't find the secondary .so, needed by the primary one, needed by the executable.

It then immediately chokes, saying "undefined reference to xxx", where xxx is the function in the secondary .so, needed by the primary .so, needed by the executable.  (This function is not, by the way, needed by the executable, just by the .so it needs.)

I tried adding -rpath-link to it, and put the full path from C root all the way to the secondary .so's folder (being careful to swap \ for / .)

I still get the exact same error, including not being able to find the secondary .so.

Objdump shows the primary .so knows it needs the secondary .so, but (does not know where to look at runtime???  How would I check that?  I use -x for objdump and see the "NEEDED" line for it, no runtime path info, so I assumed it was just in the shared library path.)  Or is it embedded, and I need a different objdump argument to add or to look in a different part of the output of objdump?  Is this runtime path, if it exists, where the linker is trying to find the secondary .so, so it ignores rpath-link?

So:

1. Does -rpath-link not apply to "downstream" .so's?  What would I use in its place?

Assume for the sake of argument the (linktime path on the development machine) is not the same as the runtime path on the target device, which it is not.

The linker must find the .so at link time to verify the function it needs is, in fact, inside the .so, right?  So why does it ignore -rpath-link=c:/(somePathWithoutSpaces)/mySoLibDir ?

post112481: Re: The application profiler doesn't show call count information

Elena Laskavaia — Fri, 28 Nov 2014 16:36:37 GMT

I tested different options and it all works from me - running from IDE,

attaching to a process, or doing postmortem.

Make sure that pass -p option to both compiler and linker.

If it still does not work use postmortem import (set var PRODIR instead,

then import gmon.out.* file in IDE).

If you still don't see call graph if doing postmortem - it means

something really went wrong with instrumentation.

On 14-10-21 04:22 PM, Andrey Dobrovolskiy wrote:

> Hi Shawn,

>

> I want to profile by attaching to the running process. I've created a profiler configuration for Sampling and Call Counting, provided path th the shared libraries and source code files.

>

> What happens after I start profiling session is that I do see a live profiling process, with all the function names and their location into the source files. After compiling with -gdwarf-2 flag I'm even able to switch to the particular source and see CPU time distribution over the code lines (though to me it looks to be improper).

>

> What's missing is the call graph and hit count information. The latter is present but only for the node named <filtered> so I would assume it's rather some code from uninstrumented library or a kernel one. Concerning call count I must add that among other functions listed in profiling results there ones called _mcount_ and mcount, which I interpret as a sign that the code does have instrumentation.

>

> For me it's completely a mistery what's going on behind the IDE. As far as I understand after setting a QCONN_PROFILER variable to /de/profiler I should have redirected profiling statistics to the IDE, but how can I check that this is really the case and I'm not still dealing with a plain Sampling profiler approach?

>

> Thank you & regards,

> Andrey

>

>

>

> _______________________________________________

>

> QNX Momentics Community Support

> http://community.qnx.com/sf/go/post112090

> To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post112407: JNI Interactions on QNX with SWIG generated source code

Andrew Lotito(deleted) — Fri, 21 Nov 2014 15:29:45 GMT

Good day,

I'm currently experiencing an issue finding native methods with my JNI jar file and QNX compiled libraries where the related native methods were generated by SWIG. When I run the JNI jar and make a native method call to the library, I get an error claiming it couldn't find the method. I know I've loaded the library without error before making the call, and that the method names for both the Java native methods and the C library methods are correct. I have tested the libraries and the JNI on Linux and Windows, and had no issues finding the method. I have also created another example QNX compiled Library and JNI jar that used JNIEXPORT instead of SWIGEXPORT and had no issues there. I was curious if anyone else has had any issues between SWIG, JNI, and QNX? I'm using Java 5 and QNX 6.4.1. Thank you for your time.

Best Regards,

Andrew

post112259: Re: The application profiler doesn't show call count information

Elena Laskavaia — Tue, 04 Nov 2014 21:13:12 GMT

Did you link with -p as well?

IDE Should have a sample project that has makefile with options for call count instrumentation, does this work?

In general samping works only on 1st core, this is unrelated to your problem, but to have correct info, you have to bind all your threads

to cpu0

post112167: Re: The application profiler doesn't show call count information

Andrey Dobrovolskiy(deleted) — Mon, 27 Oct 2014 09:42:58 GMT

Anyone?

post112091: Re: The application profiler doesn't show call count information

Andrey Dobrovolskiy(deleted) — Tue, 21 Oct 2014 20:29:43 GMT

/de/profiler is of course a /dev/profiler

P.S. My application has multiple threads and runs on a dual core processor (shown as CPU 0 & 1 in kernel traces). Probably this could lead to some side effecst? If I'm not wrong I saw a "know issue" with running sampling and call count profiling on multiprocessor systems. Though even if it's the case the worst thing I can expect is that I won't see profiling info for the code, which was executed on the second CPU core...

post112090: Re: The application profiler doesn't show call count information

Andrey Dobrovolskiy(deleted) — Tue, 21 Oct 2014 20:22:54 GMT

Hi Shawn,

I want to profile by attaching to the running process. I've created a profiler configuration for Sampling and Call Counting, provided path th the shared libraries and source code files.

What happens after I start profiling session is that I do see a live profiling process, with all the function names and their location into the source files. After compiling with -gdwarf-2 flag I'm even able to switch to the particular source and see CPU time distribution over the code lines (though to me it looks to be improper).

What's missing is the call graph and hit count information. The latter is present but only for the node named <filtered> so I would assume it's rather some code from uninstrumented library or a kernel one. Concerning call count I must add that among other functions listed in profiling results there ones called _mcount_ and mcount, which I interpret as a sign that the code does have instrumentation.

For me it's completely a mistery what's going on behind the IDE. As far as I understand after setting a QCONN_PROFILER variable to /de/profiler I should have redirected profiling statistics to the IDE, but how can I check that this is really the case and I'm not still dealing with a plain Sampling profiler approach?

Thank you & regards,

Andrey

post112089: Re: The application profiler doesn't show call count information

Shawn Unger(deleted) — Tue, 21 Oct 2014 20:00:15 GMT

One quick edit (pulled the trigger a little fast):

Change:

From the IDE, right click on the debug binary and select "C/C++ QNX Application Dialog..." (not strictly required, but it saves a couple steps).

To:

From the IDE, right click on the debug binary and select "C/C++ QNX Application Dialog..." from the "Profile As..." menu option.

post112088: Re: The application profiler doesn't show call count information

Shawn Unger(deleted) — Tue, 21 Oct 2014 19:56:42 GMT

Hi Andrey,

There's a few steps that are missing from your description that make things a little unclear (I also don't see any attachments). I will assume that you want to profile an executable from launch and not attaching to an already running executable. Let me know if this is an incorrect assumption.

Based on your description, it appears that compiling and linking worked fine.

Here's a verbose step-by-step of how to get things up and running from the IDE:

From the IDE, right click on the debug binary and select "C/C++ QNX Application Dialog..." (not strictly required, but it saves a couple steps).

In the Profile Configurations, the executable should be listed under "C/C++ QNX QConn (IP)." Select the "Tools" tab and then select the "Add/Delete Tool..." button. Check the checkbox for Application Profiler and press "Okay."

This should give you the Application Profiler tool. Activate the "Sampling and Call Count Instrumentation" radio button.

Hit the "Profile" button and you should be able to switch to the Application Profiling perspective and see your results "live."

Take care,

Shawn Unger

post112078: The application profiler doesn't show call count information

Andrey Dobrovolskiy(deleted) — Tue, 21 Oct 2014 12:44:50 GMT

Hi all,

The application profiler doesn't show call count information in the Execution Time view inside Momentics (see attached

screenshot).  Here are the steps I've taken, can you please tell me where I've gone wrong, or what I've missed?

I'm compiling using a Makefile using the -g -p switches.  I refresh the IDE so it knows about the new binary files.  I

right-click on the debug binary and select Profile As-->Profile Configurations.  I create a new profile under C/C++ QNX

Attach to Remote Process via QCONN (IP).  I'm connecting to my target on port 8000.  Inside the tools menu, I've added

the Application Profiler and selected Sampling and Call Count Instrumentation.

I've tried setting the QCONN_PROFILER to /dev/profiler as suggested here:

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=%2Fcom.qnx.doc.ide.userguide%2Ftopic%

2Fprofiler_profile_already_running_.html

I'm using Momentics IDE version 4.7.0, running QNX version 6.5.0SP1.

Any suggestions or tips would be greatly appreciated.

Thanks,

Andrey

post112046: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Kevin Schaffer(deleted) — Wed, 15 Oct 2014 14:00:01 GMT

Ok, if I enable "Links" toolbar, then this functions for both putting a text document and the QNX Momentics starter, just as it used to in Windows XP.

[.gif of George Baily tearing through Bedford Falls screaming "Yaaaaaaaaaaaay! Yaaaaaaaaaaaaaaaaay!"]

post112042: Re: Makefile question

Kevin Schaffer(deleted) — Tue, 14 Oct 2014 18:32:03 GMT

Lemme ask a slightly different question. Is this a valid dependency line?

my_make_target: SOME_VAR := $(SOME_OTHER_VAR)

Or is the setting of make variables on a dependency line invalid?

post112039: Makefile question

Kevin Schaffer(deleted) — Mon, 13 Oct 2014 16:05:36 GMT

Ok, I'm having trouble forcing a makefile variable to be something for a particular build.  I've stripped it out into a toy file that exhibits the same problem

Makefile

---------------

$(info 1 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

ifeq ($(MYPLATFORM),doggy)

$(info 2 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

SOMEOTHERVAR := helloworld

$(info 3 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

endif

$(info 4 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

$(info 5 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

mymaketarget: MYQCCCALL := $(SOMEOTHERVAR)

$(info 6 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

mymaketarget: $(info 7 MYPLATFORM $(MYPLATFORM) SOMEOTHERVAR $(SOMEOTHERVAR) MYQCCCALL $(MYQCCCALL))

-----------------------------------

Call on command line with complete output:

C:\xyz>make MYPLATFORM=doggy mymaketarget

1 MYPLATFORM doggy SOMEOTHERVAR  MYQCCCALL

2 MYPLATFORM doggy SOMEOTHERVAR  MYQCCCALL

3 MYPLATFORM doggy SOMEOTHERVAR helloworld MYQCCCALL

4 MYPLATFORM doggy SOMEOTHERVAR helloworld MYQCCCALL

5 MYPLATFORM doggy SOMEOTHERVAR helloworld MYQCCCALL

6 MYPLATFORM doggy SOMEOTHERVAR helloworld MYQCCCALL

7 MYPLATFORM doggy SOMEOTHERVAR helloworld MYQCCCALL

make: Nothing to be done for `mymaketarget'.

I don't believe the "nothing to be done" has anything to do with it, because the "real" file does not show that.  The infos show that, in spite of specifying mymaketarget (and several maketarget: lines get executed) the one that assigns a value to MQCCCALL does not get executed.

Why?  How can I force it to be set when specifying mymaketarget?

Is MYQCCCALL being set to some phantom, un-set version of SOMEOTHERVAR?  Or is that line not being executed at all, and why not?  := vs. = does not seem to matter.

post112025: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Jeff Baker — Fri, 10 Oct 2014 15:19:58 GMT

A common complaint in earlier releases was that the product required Administrator privilege to install.  This privilege was required solely for the ability to update the system environment.  The solution was to move to the more common industry practice of leaving the system environment as-is and use a shortcut that sets the appropriate variables locally.  The VBS script is there only as a quick-trick to hide the ugly DOS box that you would normally get when running a batch script.

To save yourself the time of digging through the start menu, there are two solutions.

1)  Pin the VBS script, as you have already done.  Once you have it pinned to your quick launch you can right click on it and choose the run-qde script.  I know the icon is hideous but you'll have to complain to MS about that.

2)  Add the variables defined in the run-qde.bat script to your default system environment and create a normal shortcut to launch qde.exe directly.  Then pin that shortcut.

post112023: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Kevin Schaffer(deleted) — Fri, 10 Oct 2014 14:33:57 GMT

Hey, QNX employees! If anyone reads this, here's a great way to get a bonus this year!

I notice it takes me about 7 extra seconds to go into the Start Menu, then to QNX Momentics, then start Momentics, than it does to click it on the task bar (which, as mentioned, I cannot do currently.)

So at 7 seconds per day, times 260 days a year, times thousands of people who use QNX, you're looking at saving corporate customers tens of thousands of dollars a year by making your release installs' startup link draggable onto the task bar.

So do that for the next release, and claim you saved customers $30,000.

post111991: Re: compiling boost (1.55) cpp libraries for qnx

Niklas Angare — Wed, 08 Oct 2014 19:56:13 GMT

I'm a little late to the discussion but I thought I'd contribute in case it can help others.

I can cross-compile most of Boost 1.55 for Arm from Windows using the following command line:

.\b2 toolset=qcc target-os=qnxnto threadapi=pthread link=static -l240 --layout=system --without-python cxxflags="-Vgcc_ntoarmv7le -Y_gpp -Wc,-std=gnu++0x -D_LITTLE_ENDIAN" linkflags="-Vgcc_ntoarmv7le -Y_gpp -lang-c++" -j7 > buildlog.txt 2>&1

This should work on Linux and QNX as well. -j7 uses seven threads for compilation.

There will be several megabytes of compiler errors.

Remove -Wc,-std=gnu++0x if you don't want C++0x mode.

Replace -Vgcc_ntoarmv7le with -Vgcc_ntox86 to compile for x86.

A slightly better way to do it is to create a user-config.jam in your user directory and put the compiler options in it like this:

using qcc : arm : : <cxxflags>"-Vgcc_ntoarmv7le -Y_gpp -Wc,-std=gnu++0x -D_LITTLE_ENDIAN" <linkflags>"-Vgcc_ntoarmv7le -Y_gpp -lang-c++" ;

"arm" is just a name for the configuration.

The command line to use:

.\b2 toolset=qcc-arm target-os=qnxnto threadapi=pthread link=static -l240 --layout=system --without-python -j7 > buildlog.txt 2>&1

post111975: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Kevin Schaffer(deleted) — Tue, 07 Oct 2014 14:27:00 GMT

I just wanted to drag-and-drop the icon from the start menu to the task bar, like I do for every other single program in existence, and have it work logically and intuitively.

Windows also appears to have difficulties promoting it to the "commonly-used" area at the top of the start menu.

I'll leave it as a Windows bug rather than a QNX Momentics one. I was just hoping someone had realized this was an issue and solved it already.

post111955: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Christoph Nemmaier — Thu, 02 Oct 2014 15:06:01 GMT

Hi,

a short look on the startup link properties shows you why. (Well, I wonder how you managed to pin QDE 5.0 to the taskbar. On my PC, Windows doesn't even show the context menu entries.) It is possible however with QDE 4.7 for the simple reason that the startup menu links to qde.exe directly, while the QDE 5.0 link goes to run-qde.vbs, which in turn starts run-qde.bat, which sets a bunch of environment variables before starting qde.exe. This is in fact great for users like me who run different environments and Eclipses on their PCs. Before 5.0, I used an own startup wrapper for years.

Of course, you could have found all this in the documentation. Didn't you ever stumble upon the quickstart guide? (Chapter 5, "Creating a program project", <http://www.qnx.com/download/download/26166/A_Quickstart_Guide.pdf>)

HTH

Christoph

post111954: Re: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Kevin Schaffer(deleted) — Thu, 02 Oct 2014 14:27:04 GMT

This is a mailing list, not a BBS?

post111953: Pin QNX Momentics IDE 5.0 to (Windows 7) taskbar

Kevin Schaffer(deleted) — Thu, 02 Oct 2014 14:18:58 GMT

Normally I pin common programs to the task bar by going into Start Menu -> Programs -> whatever, and dragging the program to the taskbar.  A little tooltip pops up saying I am about to "Pin to taskbar".

However, when I do this with the QNX Momentics 5.0, it asks me if I want to pin to wscript.exe.  I click yes, and then something gets pinned there.

When clicking it, "Windows Script Host Settings" starts.  No doubt this is wscript.exe.  Whatever it is, it's not QNX Momentics 5.0.

No doubt it's a clever (I guess...) way to start Momentics, but Windows is too stupid to just drag-link it.  I've tried other things (right-click under Start menu provides no "Pin it" options, unlike everything else, and if it did, I'll bet it would do the wrong thing anyway).  Making a link on the desktop, when dragged onto the taskbar, yields the original poor results.

Any ideas?

post111915: Re: Math in a Makefile

Kevin Schaffer(deleted) — Mon, 29 Sep 2014 14:50:24 GMT

If the result of a logical operation such as > is 0 or 1, then that result cannot be passed in to subsequent AND or OR operations as they are based on empty string-as-false, not 0-as-false.

Are there any examples of more complex logic being done in a makefile? Cloning that is probably a lot faster than reverse-engineering how to do this.

post111897: Re: Math in a Makefile

Will Miles — Fri, 26 Sep 2014 17:55:59 GMT

Hi Kevin,

The make supplied by QNX is GNU make. The documentation here: http://www.gnu.org/software/make/manual/html_node/Conditional-Functions.html describes the behaviour of $(if), $(and), and $(or) - note that the predicate is whether or not the argument is empty. Another aspect to understand is that "$$" is just an escaped "$" character. As far as make is concerned, "var = $$((a))" is you asking it to set the variable "var" to the string "$((a))". However when you put that on a line to be interpreted by the shell - such as an 'echo' line - the shell will get the text "$(())" which the shell interprets as a request for arithmetic expansion. The key point is that it is the shell doing the arithmetic, not the make engine.

You might try using the $(shell) function to invoke the shell to have the expansion performed earlier. For example, in your test makefile, try:

M3 = $(shell echo $$(( $(QCCMAJORVER) > 2 )))

..to run the shell, compute the expansion, and save the result as '1' or '0', rather than just storing the text "$(( 6 > 2))".

Hope this helps,

-Will

post111894: Re: Math in a Makefile

Kevin Schaffer(deleted) — Fri, 26 Sep 2014 15:53:54 GMT

In lieu of that, trying the daisy-chain route :(

QCCMAJORVER := 6

M3 = $$(( $(QCCMAJORVER) > 2 ))

...

(stuff)

echo "Hello, world!"

echo $(QCCMAJORVER)

echo "$(M3)"

echo $(M3)

ifeq ($(M3),1)

echo "New"

else

echo "Old"

endif

echo "End"

Gives

echo "Hello, world!"

Hello, world!

echo 6

6

echo "$(( 6 > 2 ))"

1

echo $(( 6 > 2 ))

1

echo "Old"

Old

echo "End"

End

Why does 1 not equal 1, causing it to print Old instead of New? Is it comparing something like a string of "$(( 6 > 2 ))" to 1 instead of evaluating it? What forces it to evaluate in an ifeq lline?

post111892: Re: Math in a Makefile

Kevin Schaffer(deleted) — Fri, 26 Sep 2014 14:38:47 GMT

Well, the permutations attempting to reverse-engineer a not-intuitive system approaches infinity.

Tried this:

echo "Logic test follows"

echo $(or 1,0)

echo $(or 0,0)

echo "Or 0 1"

echo $(or 0,1)

echo "Or 1 1"

echo $(or 1,1)

echo $(and 0,0)

echo $(and 1,0)

echo $(and 0,1)

echo $(and 1,1)

echo $(xor 0,0)

echo $(xor 0,1)

echo $(xor 1,0)

echo $(xor 1,1)

Get this:

echo "Logic test follows"

Logic test follows

echo 1

1

echo 0

0

echo "Or 0 1"

Or 0 1

echo 0

0

echo "Or 1 1"

Or 1 1

echo 1

1

echo 0

0

echo 0

0

echo 1

1

echo 1

1

echo

echo

echo

echo

Note:

Or 0 1

echo 0

0

echo "Or 1 1"

Or 1 1

echo 1

1

So whatever or does, it isn't or.

xor doesn't do anything other than nullify input apparently.

Tried or without comma:

echo $(or 1 1)

-->

echo 1 1

1 1

These aren't errors, nor are they boolean operations.

post111891: How do I edit or delete a post (reply)?

Kevin Schaffer(deleted) — Fri, 26 Sep 2014 14:22:55 GMT

Who would I contact? Plowing down from top level at QNX.com from the Contact page seems fraught with more difficult, tedious slogging than Bilbo in movie after movie after movie.

post111890: Re: Math in a Makefile

Kevin Schaffer(deleted) — Fri, 26 Sep 2014 13:48:38 GMT

Ok, I tried it, got:

[code]echo $(4<2 && 5<6)

C:/Users/wztjhp/AppData/Local/Temp/make9816-1.sh: line 1: 2: No such file or directory

[/code]

Is the && allowed (or some other symbol?)

What I'm really looking for is a quick Apache version # >= test. I have the logic, just not the simplest form for a Makefile.

post111885: Re: Math in a Makefile

Michael Kurt — Fri, 26 Sep 2014 06:33:06 GMT

forgot an "and" between "away" and "try" - sorry.

post111884: Re: Math in a Makefile

Michael Kurt — Fri, 26 Sep 2014 06:32:17 GMT

> echo $$(( 4 < 2 ) && ( 5 < 6 ))

For this, You do not really need the inner parentheses, so throw them away try: echo $$(4<2 && 5<6)

HTH,

Michael.

post111883: Math in a Makefile

Kevin Schaffer(deleted) — Thu, 25 Sep 2014 19:09:43 GMT

The fate of Earth depends on this.

In my Makefile, QCC 6.6 on Windows (7)

     echo $$(( 4 * 2 ))

-> 8, correct

     echo $$(( 4 > 2 ))

-> 1, correct

     echo $$(( 4 < 2 ))

-> 0, correct

but

     echo $$(( 4 < 2 ) && ( 5 < 6 ))

[stuff]Temp/make11500-1.sh: line 1: 4 < 2 ) && ( 5 < 6 : syntax error in expression (error token is ") && ( 5 < 6 ")

The question to save Earth (hurry!) is:  What is the magical token string with which to replace && in the expression above?  I don't want to do a succession of clumsy ifeq 1 type stuff to create a chain of ands (or ors).

post111871: Re: build error

Ryan Mansfield(deleted) — Wed, 24 Sep 2014 23:39:01 GMT

On 14-09-24 06:08 PM, manoj thayilthekkathil wrote:

> Hi,

>

> I am trying to setup development environment for a atmel evk using QNX® Momentics on windows

> I downloaded QNX development platform 6.6 and bsp-atmel-atsama5d3x-ek.zip from QNX.

>

> Here is the error that I observe at the end of build.

>

> C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `wfene'

> C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `strexeq r0,r3,[r2]'

> C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `teqeq r0,#0'

> make[4]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware/devc/serdebug'

> make[3]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware/devc'

> make[2]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware'

> cc: C:/qnx660/host/win32/x86/usr/bin/arm-unknown-nto-qnx6.6.0eabi-as caught signal 1

> make[1]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src'

The BSP is for QNX 6.5.0SP1. It needs to be updated for QNX 6.6.

Specifically, the definition of __inline_InterruptLock in the

arm/neutrino.h packaged with the BSP is missing a "iteet ne" instruction

before the wfene instruction (this has been update in the official QNX

6.6 arm/neutrino.h. While modifying the bsp header, or adding -marm when

compiling would fix the build issue, I can't comment on whether there

would be any runtime issues as I don't believe it has been tested with 6.6.

Regards,

Ryan Mansfield

post111869: build error

manoj thayilthekkathil(deleted) — Wed, 24 Sep 2014 22:08:36 GMT

Hi,

I am trying to setup development environment for a atmel evk using QNX® Momentics on windows

I downloaded QNX development platform 6.6 and bsp-atmel-atsama5d3x-ek.zip from QNX.

Here is the error that I observe at the end of build.

C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `wfene'

C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `strexeq r0,r3,[r2]'

C:\Users\MALEXA~1\AppData\Local\Temp\2qccYJU9L9\intr.s:102: Error: thumb conditional instruction should be in IT block -- `teqeq r0,#0'

make[4]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware/devc/serdebug'

make[3]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware/devc'

make[2]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src/hardware'

cc: C:/qnx660/host/win32/x86/usr/bin/arm-unknown-nto-qnx6.6.0eabi-as caught signal 1

make[1]: Leaving directory `C:/Users/malexander/ide-5.0-workspace/bsp-atmel-atsama5d3x-ek/src'

make[6]: *** [intr.o] Error 1

make[5]: *** [install] Error 2

make[4]: *** [install] Error 2

make[3]: *** [install] Error 2

make[2]: *** [install] Error 2

make[1]: *** [install] Error 2

make: *** [install] Error 2

Any help on this topic is appreciated.

Thanks,

Manoj

post111665: Re: RE: Process Killing

Dave Nickerson(deleted) — Fri, 05 Sep 2014 18:05:58 GMT

Not sure what you are attempting to do.  Have a look at kill or slay.

From a shell, you can terminate a process using the slay command and the name of the process you are terminating.

You could also kill a process with a SIGTERM or SIGKILL signal.

If you are trying to terminate a process and restart it again within a shell using a single command line, then:

# slay process_name ; process_name

But I may not understand what you are trying to accomplish.  Hopefully this is enough to allow you look in the right location in the documentation.

Dave

post111526: Re: OpenGL ES 2.0 sample under QNX 6.5

Joel Pilon(deleted) — Fri, 22 Aug 2014 23:51:49 GMT

‎Sorry for the delay, these kind of questions are likely better to be posted under a graphics forum.

Yes, the screen framework can run under QNX 650 along with GLES2 and GLES3.x for that matter. However, a graphics driver supporting those API's which is based on the screen framework needs to be built against and be functional in a 650 environment. Patches for both a 650 screen and a graphics driver for your platform may already exist or even a complete 650 screen based BSP. If neither does, it likely wouldn't be much effort to produce either.

I'd suggest contacting QNX support/services with your request if can find anything posted on the foundry site.

-Joel

  Original Message

From: Jason Anderson

Sent: Monday, August 18, 2014 5:38 PM

To: momentics-community

Reply To: momentics-community@community.qnx.com

Subject: OpenGL ES 2.0 sample under QNX 6.5

Hi everyone,

I have created a simple OpenGL ES 2.0 project under QNX SDP 6.6, using the Freescale i.MX6 Lite board.  I used the apps media player and gles2-gears as an example to figure this out.

Now I want to port this backwards to QNX SDP 6.5, but I have a few issues:

  1)  I do not know if SDP 6.5 even supports OpenGL ES 2.0, since the Advanced Graphics documentation only mentions ES 1.0.

  2) I cannot find a graphics patch for 6.5 that supports the Vivante graphics chipset.  It seems the 6.5 BSP for the board only comes with software rendering drivers for io-display.

  3) SDP 6.5 does not appear to use the Screen API, so it looks like I have to port my code to us the GF API instead.  I have found some documentation on how to do this, but it would be great if there were some example source to a simple app like gles2-gears that shows how to do this.

Thanks for your help!

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post111480

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post111522: Re: RE: Process Killing

Ajay Kumar Gupta(deleted) — Fri, 22 Aug 2014 05:26:45 GMT

I really appreciate your suggestion but I was trying to solve all these problems from last 15 days but somehow I couldn't solve that's why I am asking all these problems.

post111517: RE: Process Killing

Mario Charest — Thu, 21 Aug 2014 11:02:47 GMT

What about reading the manual.

________________________________________

From: Ajay Kumar Gupta [community-noreply@qnx.com]

Sent: Thursday, August 21, 2014 6:20 AM

To: momentics-community

Subject: Process Killing

I am using QNX Momentics IDE 6.5.0 on Windows platform. I want to know how killing and re-invoking of a process is done?

_______________________________________________

QNX Momentics Community Support

http://community.qnx.com/sf/go/post111516

To cancel your subscription to this discussion, please e-mail momentics-community-unsubscribe@community.qnx.com

post111516: Process Killing

Ajay Kumar Gupta(deleted) — Thu, 21 Aug 2014 10:20:42 GMT

I am using QNX Momentics IDE 6.5.0 on Windows platform. I want to know how killing and re-invoking of a process is done?

post111514: Re: Linux Java Version

Abdur Rahman(deleted) — Thu, 21 Aug 2014 09:02:57 GMT

I am also getting the same issue.

Please help us to solve this issue.

post111513: Re: Port mapped I/O instructions

Ajay Kumar Gupta(deleted) — Thu, 21 Aug 2014 09:00:32 GMT

Thanks

post111505: Photon Application Builder

Ajay Kumar Gupta(deleted) — Wed, 20 Aug 2014 10:54:01 GMT

I am using QNX Momentics IDE 6.5.0 on Windows platform. I have to develop GUI using Photon Application Builder. In that GUI I am trying to implement a widget that shuld blink showing red colour. I don't know how to show any widget blinking. Please tell me if any of you about it.

post111502: Re: Port mapped I/O instructions

Michael Kurt — Wed, 20 Aug 2014 05:47:57 GMT

What you are looking for are the functions mmap_device_io(), munmap_device_io() declared in <sys/mman.h> as well as the in8(), out8(), in16(), out16(), etc - functions from <hw/inout.h>.

post111486: Re: Port mapped I/O instructions

Ajay Kumar Gupta(deleted) — Tue, 19 Aug 2014 08:16:45 GMT

> I am using QNX Momentics 6.5.0 on Windows OS. I want to know the instructions

> available in QNX C forthe port mapped I/O.