Feed for discussion General in project Community.

post122678: QNX 7.0 + qcc 5.4.0: Binary fails to execute when compiled with -Wl,-z,noexecstack

Seunghyeon Lee — Mon, 25 Aug 2025 10:09:45 GMT

Hi everyone,

I'm currently working with QNX 7.0 and using qcc version 5.4.0 as the compiler. I encountered an issue where a binary compiled with the following linker flag:

-Wl,-z,noexecstack

fails to execute on the target system. When I remove this flag, the binary runs as expected.

I’ve verified that the build completes without errors, and the resulting binary appears valid (e.g., correct architecture, permissions, etc.). However, execution simply fails without any meaningful runtime error message.

Has anyone experienced similar behavior with -z noexecstack on QNX 7.0?

Is this a known limitation or incompatibility with the QNX toolchain or runtime?

Any insights, workarounds, or documentation references would be greatly appreciated.

Thanks in advance!

post122668: CPU load

HECTOR GASTAMINZA — Mon, 30 Jun 2025 15:01:46 GMT

Can the CPU load be obtained by apps (or process) in QNX?

post122650: Re: Reusing SLM modules and components.

kim seongmin — Mon, 03 Feb 2025 03:06:49 GMT

I am using the QNX 7.0.0 version.

post122649: Reusing SLM modules and components.

kim seongmin — Mon, 03 Feb 2025 02:48:16 GMT

Hi experts,

I am developing a monitor app that manages apps using SLM.

I want to use the 'Reusing SLM modules' feature in SLM(https://www.qnx.com/developers/docs/7.0.0/index.html#com.qnx.doc.neutrino.utilities/topic/s/slm.html). However, even though I followed the guide, I couldn't reuse the SLM module.

I created my_reusable_modules.xml and executed it in the main XML file using the following code, but an error occurred.

I am sharing the XML files I used.

1. main.xml

<?xml version="1.0" encoding="utf-8"?>

<!DOCTYPE SLM_system [

    <!ENTITY test_A SYSTEM 'A_config2.xml'>

]>

<SLM:system>

    &test_A;

</SLM:system>

2. reusable module

<?xml version="1.0" encoding="utf-8"?>

<SLM:system>

    <SLM:component name="testA">

        <SLM:command>/ti_fs/testA/SLM_testA</SLM:command>

        <SLM:repair>restart</SLM:repair>

    </SLM:component>

</SLM:system>

3. execution command

slm -v /ti_fs/slm-config2.xml

4. error string

slm: error parsing config file: line 6: error in processing external entity reference

Please provide advice and guidance.

post122644: SSH could not access any environment variables written in the /etc/profile

liu zhikang — Wed, 25 Dec 2024 10:25:01 GMT

Hi all the QNX contributors:

I use ssh to connect to remote QNX machine and execute a script to run system test cases automatically, but i found SSH could not find any environment variables written in the /etc/profile, even if i use command like: ssh user@1.1.1.1 'sh /etc/profile; echo $MY_ENV'.

I will appreciate it if you give me some advise help to solve this problem. Thank you!

Version Info:

QNX SDP 7.1

Regards,

Lewis

post122628: qnx vdev demo

wang tao(deleted) — Wed, 23 Oct 2024 01:54:17 GMT

I looked at the official documentation on how to write vdev. But I have been unable to find the relevant source code, so I don't know how to start. Please tell me, where can I get the source code?

post122608: asynchronous I/O method on QNX

Jason Lee — Thu, 19 Sep 2024 04:59:59 GMT

Hello all,

Recently, I'm testing read/write performance of flash memory on my QNX device,

but I'm getting poor performance with POSIX support method `aio_read()` and `aio_write()`,

after searching , I found out It might because it was using multi-thread to perform async I/O , causing heavily context switch/thread creation/thread deletion.

so, does there exist any other I/O method which supports asynchronous I/O besides `aio_read()` and `aio_write()`,

or there's better way to perform test of read/write.

post122597: tcpdump timestamp issue

Lin Xiang — Fri, 16 Aug 2024 12:33:41 GMT

Hi QNXers:

I enter a wired issue about tcpdump timestamp:  tcpdump timestamp is almost always larger than application timestamp when receive the package.

Using below steps can reproduce the issue on QNX qemu virtual machine:

QNX: start a netcat server with shell: ./netcat -lp 8888 -e /bin/sh

             then start  tcpdump:  tcpdump -tt -s0 -i vt0 port 8888 -X

PC: nc to qnx:  nc 172.31.1.110 8888,  then enter  date +%s.%N  to get QNX timestamp

From the attached picture:

   T1: 1723810894.254562  receive the date command packet, then tcp package passing to netcat server

   T2: 1723810893.643000000 netcat server receive date command, run the command get this timestamp

Why T2 < T1?  this  is normal?  From qnx documents:  https://www.qnx.com/developers/docs/7.0.0/index.html#com.qnx.doc.neutrino.utilities/topic/t/tcpdump.html

       By default, all output lines are preceded by a timestamp. The timestamp is the current clock time, in the form hh:mm:ss.frac and is as accurate as the kernel's clock.

      The timestamp reflects the time io-pkt first saw the packet.

Hope some guru can give some suggestions about this issue, thanks.

Version Info:

QNX SDP 7.1

tcpdump --version

tcpdump version 4.9.3

libpcap version 1.9.1

OpenSSL 1.1.1k  25 Mar 202

netcat:  https://sourceforge.net/projects/netcat/files/

post122571: Re: port grpc to qnx7.1.0

Minh Bui Quang — Wed, 24 Jul 2024 02:17:22 GMT

I've found a public lib on Gitlab (https://gitlab.com/qnx/libs/grpc/)

Tried to build it but not success.

Do you have any update?

post122509: Ethernet controller driver in QNX SDP 6.5.0 SP1 for NIH350 AM4

Ramarao Meda(deleted) — Mon, 18 Mar 2024 13:03:29 GMT

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

Hello,

My name is Ramarao Meda, working as Embedded Software Engineer in National Oilwell Varco, L.P(NOV). Our Ethernet controller has been upgraded from NH82580 to NHI350 because of chip shortage.

I have downloaded the next version of BSP(6.5.0 SP1) and tried to build it by following the steps in BSP user guide. After building it, .ifs image is created, and I have stuck there because our image loading process in target is different.

We are building an .iso image from QNX VM and loading into the target(ATOM processor). We are unable to use this .ifs file to build new image.

I have found the exact driver what we need for this latest controller i.e., devnp-e1000.so. But not able use this driver file directly.

Please support us with suitable steps for including above driver file and building the new image (.iso)for target.

Thanks,

Ramarao Meda,

Embedded Software Engineer,

National Oilwell Varco, L.P.

post122488: Enable Mouse Input in VMware image with graphics

Mario Doerr — Fri, 16 Feb 2024 10:39:09 GMT

Hi,

I've created a VMware image which is running some graphics application. The app is running and fine but I also want to enable mouse or touch input.

Some drivers are loaded/configured by the mkqnximage tool but from my point of view the /graphics/vmwgfx-drm/graphics.conf is not configured properly, it's missing "input = mouse|keyboard".

How can I add these line, or write a new graphics.conf file in the case, when I'm using the mkqnximage tool?

Thanks.

Best Mario

post122468: Interrupt handling in vdev

Jinyong Lee — Wed, 31 Jan 2024 07:11:06 GMT

Hello,

I'm working for developing vdev of gpio for SS.

The interrupt feature of gpio is added in virtio v1.2.

For the host side, there is a gpio resource manager.

There is no code for interrupt in the gpio resource manager, like InterruptAttachEvent(), InterruptWait().

So, I added a thread with InterruptWait() in vdev code.

I wonder the interrupt handling in the vdev code is ok or not.

Can anyone also tell me this is ok or not.

If not, tell me better method please.

post122466: Build GDB Source Code for QNX (QNX App Debugging alternative)

Shail Shah — Fri, 26 Jan 2024 09:50:00 GMT

I’ve retrieved the GDB source from QNX CVS.

The error message indicates that the linker (ld) cannot find the definitions for various functions, such as i386_linux_read_description, amd64_linux_read_description, amd64_collect_fxsave, amd64_collect_xsave, amd64_supply_fxsave, amd64_supply_xsave, amd64_linux_gregset_reg_offset, amd64_supply_fxsave, and amd64_collect_fxsave.

1) Where to find the source code of above function or is it available in some library how to link that library?

2) IS any script availbe to use for the same?

For Example : Using Standard GDB with pdebug:While a standard GDB client wouldn't directly work with pdebug, it's possible to bridge the gap using additional tools:

QNX SDK: The QNX Software Development Platform (SDK) includes a wrapper script that allows executing standard GDB commands over the pdebug connection. This essentially translates GDB commands into the pdebug protocol, enabling limited use of regular GDB with pdebug.

3)Custom scripts: With advanced knowledge of both GDB and pdebug protocols, it's possible to write custom scripts or wrappers that translate commands and bridge the communication gap.

Is any customer scripts availbe for use ?

4) We just need to avoid QNX Debugger. We just want to use open source debugger for QNX application if any other approched can u please inform?

post122439: Gles2 gears failed to start in VmWare

ye changhua(deleted) — Thu, 23 Nov 2023 13:43:20 GMT

I am using SDP version 7.0 and the host is a Windows system. I started the system on VMware Workstation Pro and started the screen graphic system. I saw a black screen, and then I logged in to the system through SSH and ran the gles gears application, but there was an error（EglInitialize: EGL is not initialized, or could not be initialized, for the specified display)!

post122434: qvm vs qvm-safety

Mustafa Alaa(deleted) — Sun, 12 Nov 2023 06:50:35 GMT

we have host vm that runs in safety context, and load all safety variants of all virtual devices and libmod_qvm-safety, in addition we have guest vm which DOESN'T run in safety context

the question is

is there any problem if we loaded the safety variants of all virtual devices for the host vm then we used the qvm binary to run the guest vm (not the qvm-safety) ?

or since we ran the safety variants for the host ,we must use qvm-safety to run the guest ?

post122433: Re: license key not recieved

Bowen Li(deleted) — Fri, 10 Nov 2023 03:12:25 GMT

Has your problem been fixed?

I have also encountered the same issue.

post122423: BSP Patch Data

Muhammed Fasal — Fri, 03 Nov 2023 09:38:09 GMT

I was looking for the QNX x86 BSP patch for Generic x86 BIOS / APIC board which was available since 27th April 2016.

But I couldn't find the downloadable file in below path.

https://community.qnx.com/sf/wiki/do/viewPage/projects.bsp/wiki/BSPAndDrivers

Can I someone help me to know the content of this patch/where I can find the downloadable.

post122399: Re: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Albrecht Uhlmann — Tue, 17 Oct 2023 08:35:29 GMT

Ok. then I fear you have to put the file in /lib/dll. Maybe a symlink can help. It is quite possible that only a hard-coded set of paths is used to search for dynamic modules, might be a security issue.

Regards,

Albrecht

post122391: Re: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Vic One — Thu, 12 Oct 2023 09:41:26 GMT

No, I've done the test. setconf to _CS_LIBPATH variable still doesn't work for dynamic link library finding.

post122390: Re: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Albrecht Uhlmann — Thu, 12 Oct 2023 08:09:13 GMT

That's why I suggested to additionally add your lib path to the _CS_LIBPATH variable. These two variables are not synced by QNX OS. Also /etc/profile may not be executed on system boot, only in interactive login shell.

post122389: Re: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Vic One — Thu, 12 Oct 2023 08:06:50 GMT

Hi,

I didn't use the utilty setconfig/getconfig. I edit the /etc/profile directly and add my library folder to the value of LD_LIBRARY_PATH. But it doesn't work.

post122388: Re: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Albrecht Uhlmann — Thu, 12 Oct 2023 07:57:09 GMT

Hi,

have you tried to add the path of your shared object to the system configuration path?

e.g. setconf _CS_LIBPATH `getconf _CS_LIBPATH`:/your/path

(haven't tried this, just the idea)

It is quite possible that io-pkt only evaluates this variable.

Regards,

Albrecht

post122387: Does the io-pkt instance allow to mount a lsm module that will also dynamic link to another lib file?

Vic One — Thu, 12 Oct 2023 07:21:48 GMT

Hi, I get a problem in implementing a packet filtering lsm module for io-pkt.

For example, I implement a lsm-pf.so that is used to be mount to io-pkt instance. The lsm.so will dynamic link to another libtest.so that is implemented by myself.

I found if I didn't put the libtest.so to the system library folder, such as /lib/, /usr/lib, etc, the lsm-pf.so won't find the libtest.so even I already set my library to the LD_LIBRARY_PATH.

Is it a limitation for io-pkt's packet filtering mechanism?

post122386: Re: Virtual image created for VmWare with GUI not booting

Albrecht Uhlmann — Tue, 10 Oct 2023 14:47:22 GMT

Hi Mutyala,

I think this is normal - when QNX Screen starts up, it just presents a black screen. You need to run an application that actually draws something on the screen. gles2-gears is a favourite one.

You can launch this application from a telnet/ssh console, because the local keyboard in the VM is gone after launching QNX Screen.

Regards,

Albrecht

post122384: Virtual image created for VmWare with GUI not booting

Mutyala Jagadeesh Babu — Tue, 10 Oct 2023 09:47:33 GMT

We are using QNX sdp 7.1 and momentics ide to create a virtual target for vmware with an addition argument --graphics=yes to load the gfx libraries. This image doesnt boot completely and is stuck in a black screen. I have attached images. VMware Workstation 17 player is used.

post122365: Re: license key not recieved

Lin Xiang — Fri, 25 Aug 2023 02:23:09 GMT

yes, i also didn't recieved the 30-day evaluate key.

post122363: Encrypted FileSystem

Ankur Rastogi(deleted) — Wed, 16 Aug 2023 11:26:38 GMT

Dear All,

QNX features testing are there some tools available ? If so and Request to kindly share the details in response.

For example if I want to check encryption in file system and wanted to do penetration testing. what are all tools, environment can be used.

I went through the available documentation , but I am unable to identify tools for testing the features like EES or similar.

Thanks for all of you!!

Regards.

Ankur

post122314: Filesystem forms an infinite loop

Jeevasubramaniam Manoharan — Wed, 07 Jun 2023 04:56:31 GMT

Hi,

I am using QNX 710 for development.

I have observed an infinite loop response when executing below FIND command in TI Evaluation board(J721EXSOMXEVM:).

Request : find / -type l

Response : find: Cannot recurse into '/dev/sem' - filesystem forms an infinite loop

Please refer the attached file for response.

Could you please let me know why there is a infinite loop formation. Is this a known issue.

Thanks and Regards,

Jeeva M

post122269: license key not recieved

Teena J(deleted) — Thu, 18 May 2023 03:21:14 GMT

i have applied for 30day evaluation license i havent received the key yet how many days it would take generally to receive this and is there any other procedure to be done from my side other than filling the evaluation license form to get software center access for sdp

post122268: Re: How to build IOZONE tool on QNX SDP7.1

Lenin T — Wed, 17 May 2023 03:39:20 GMT

I have made changes in the make file for some qnx related defines .

Still i am facing errors as below :

11:35:28 **** Incremental Build of configuration Debug for project current ****

Info: Internal Builder is used for build

qcc -o libasync.o "..\\libasync.c" -V8.3.0,gcc_ntoaarch64le -c -O0 -g

..\libasync.c: In function 'async_init':

..\libasync.c:283:3: warning: implicit declaration of function 'exit' [-Wimplicit-function-declaration]

   exit(174);

   ^~~~

..\libasync.c:283:3: warning: incompatible implicit declaration of built-in function 'exit'

..\libasync.c:283:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:116:1:

+#include <stdlib.h>

#endif

..\libasync.c:283:3:

   exit(174);

   ^~~~

..\libasync.c:285:2: warning: implicit declaration of function 'bzero' [-Wimplicit-function-declaration]

  bzero(*gc,sizeof(struct cache));

  ^~~~~

..\libasync.c:285:2: warning: incompatible implicit declaration of built-in function 'bzero'

..\libasync.c: In function 'end_async':

..\libasync.c:306:2: warning: implicit declaration of function 'del_cache'; did you mean 'incache'? [-Wimplicit-function-declaration]

  del_cache(mygc);

  ^~~~~~~~~

  incache

..\libasync.c: In function 'async_read':

..\libasync.c:436:3: warning: implicit declaration of function 'takeoff_cache'; did you mean 'alloc_cache'? [-Wimplicit-function-declaration]

   takeoff_cache(gc,ce);

   ^~~~~~~~~~~~~

   alloc_cache

..\libasync.c: In function 'alloc_cache':

..\libasync.c:545:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(175);

   ^~~~

..\libasync.c:545:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:547:2: warning: incompatible implicit declaration of built-in function 'bzero'

  bzero(ce,sizeof(struct cache_ent));

  ^~~~~

..\libasync.c:554:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(1);

   ^~~~

..\libasync.c:554:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:563:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(176);

   ^~~~

..\libasync.c:563:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c: At top level:

..\libasync.c:622:1: warning: conflicting types for 'takeoff_cache'

takeoff_cache(struct cache *gc, struct cache_ent *ce)

^~~~~~~~~~~~~

..\libasync.c:436:3: note: previous implicit declaration of 'takeoff_cache' was here

   takeoff_cache(gc,ce);

   ^~~~~~~~~~~~~

..\libasync.c:696:1: warning: conflicting types for 'del_cache'

del_cache(gc)

^~~~~~~~~

..\libasync.c:306:2: note: previous implicit declaration of 'del_cache' was here

  del_cache(mygc);

  ^~~~~~~~~

..\libasync.c: In function 'async_read_no_copy':

..\libasync.c:799:4: warning: incompatible implicit declaration of built-in function 'exit'

    exit(1);

    ^~~~

..\libasync.c:799:4: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:819:3: warning: implicit declaration of function 'putoninuse' [-Wimplicit-function-declaration]

   putoninuse(gc,ce);

   ^~~~~~~~~~

..\libasync.c:960:4: warning: incompatible implicit declaration of built-in function 'exit'

    exit(1);

    ^~~~

..\libasync.c:960:4: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c: In function 'async_release':

..\libasync.c:986:2: warning: implicit declaration of function 'takeoffinuse'; did you mean 'takeoff_cache'? [-Wimplicit-function-declaration]

  takeoffinuse(gc);

  ^~~~~~~~~~~~

  takeoff_cache

..\libasync.c: At top level:

..\libasync.c:1000:1: warning: conflicting types for 'putoninuse'

putoninuse(gc,entry)

^~~~~~~~~~

..\libasync.c:819:3: note: previous implicit declaration of 'putoninuse' was here

   putoninuse(gc,ce);

   ^~~~~~~~~~

..\libasync.c:1021:1: warning: conflicting types for 'takeoffinuse'

takeoffinuse(gc)

^~~~~~~~~~~~

..\libasync.c:986:2: note: previous implicit declaration of 'takeoffinuse' was here

  takeoffinuse(gc);

  ^~~~~~~~~~~~

..\libasync.c: In function 'async_write':

..\libasync.c:1102:4: warning: incompatible implicit declaration of built-in function 'exit'

    exit(177);

    ^~~~

..\libasync.c:1102:4: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c: In function 'allocate_write_buffer':

..\libasync.c:1134:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(178);

   ^~~~

..\libasync.c:1134:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:1142:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(179);

   ^~~~

..\libasync.c:1142:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:1144:2: warning: incompatible implicit declaration of built-in function 'bzero'

  bzero(ce,sizeof(struct cache_ent));

  ^~~~~

..\libasync.c:1153:4: warning: incompatible implicit declaration of built-in function 'exit'

    exit(1);

    ^~~~

..\libasync.c:1153:4: note: include '<stdlib.h>' or provide a declaration of 'exit'

..\libasync.c:1167:3: warning: incompatible implicit declaration of built-in function 'exit'

   exit(180);

   ^~~~

..\libasync.c:1167:3: note: include '<stdlib.h>' or provide a declaration of 'exit'

qcc -o current fileop.o iozone.o libasync.o libbif.o pit_server.o -lc -lsocket -V8.3.0,gcc_ntoaarch64le -lang-c++ -g "-LC:\\Users\\ssdrive\\qnx710\\target\\qnx7\\aarch64le\\usr\\lib"

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: warning: alignment 4 of symbol `splash' in fileop.o is smaller than 8 in iozone.o

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: iozone.o: in function `main':

C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:1853: multiple definition of `main'; fileop.o:C:\Lenin\iozone\iozone3_506\src\current\Debug/..\fileop.c:176: first defined here

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: iozone.o: in function `purge_buffer_cache':

C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:12964: multiple definition of `purge_buffer_cache'; fileop.o:C:\Lenin\iozone\iozone3_506\src\current\Debug/..\fileop.c:147: first defined here

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: libasync.o: in function `async_read':

C:\Lenin\iozone\iozone3_506\src\current\Debug/..\libasync.c:392: multiple definition of `async_read'; iozone.o:C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:20088: first defined here

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: pit_server.o: in function `main':

C:\Lenin\iozone\iozone3_506\src\current\Debug/..\pit_server.c:131: multiple definition of `main'; fileop.o:C:\Lenin\iozone\iozone3_506\src\current\Debug/..\fileop.c:176: first defined here

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: iozone.o: in function `write_perf_test':

C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:7989: undefined reference to `signal_handler'

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:7996: undefined reference to `signal_handler'

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:8160: undefined reference to `signal_handler'

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:8236: undefined reference to `signal_handler'

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:8257: undefined reference to `signal_handler'

C:\Users\ssdrive\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld: iozone.o:C:\Lenin\iozone\iozone3_506\src\current\Debug/..\iozone.c:8281: more undefined references to `signal_handler' follow

cc: C:/Users/ssdrive/qnx710/host/win64/x86_64/usr/bin/aarch64-unknown-nto-qnx7.1.0-ld caught signal 1

11:35:28 Build Failed. 15 errors, 26 warnings. (took 296ms)

Could some one suggest how to build the tool on qnx7.1.0 ?

post122257: Re: How to build IOZONE tool on QNX SDP7.1

Lenin T — Fri, 12 May 2023 10:46:43 GMT

> Hi All,

>

> I want to build IOZONE tool for QNX OS systems.

>

> I have tried to build the tool on QNX momentics IDE but facing some build

> errors as below :

>

>

> C:\iozone\iozone3_506\src\current>make qnx

> which: no ndk-build in (/cygdrive/c/QNX/QNX_Momentics_IDE_2/features/com.qnx.

> tools.jre.win32.x86_64_11.0.10/jdk-11.0.10+9-jre/bin/server:/cygdrive/c/QNX/

> QNX_Momentics_IDE_2/features/com.qnx.tools.jre.win32.x86_64_11.0.10/jdk-11

> .0.10+9-jre/bin:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management

> Engine Components/iCLS:/cygdrive/c/Program Files/Intel/Intel(R) Management

> Engine Components/iCLS:/cygdrive/c/windows/system32:/cygdrive/c/windows:/cygdr

>

> ive/c/windows/System32/Wbem:/cygdrive/c/windows/System32/WindowsPowerShell/v1.

> 0:/cygdrive/c/windows/System32/OpenSSH:/cygdrive/c/Users/ssdrive/AppData/Local

> /Microsoft/WindowsApps:/cygdrive/c/Python27:/cygdrive/c/Python27/Script

> s:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/

> DAL:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/DAL:

> /cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Compon

> ents/IPT:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components

> /IPT:/cygdrive/c/ProgramData/chocolatey/bin:/usr/bin:/cygdrive/c/Windows/

> System32/drivers:/cygdrive/c/Users/ssdrive/AppData/Local/Microsoft/WindowsAp

> ps:/cygdrive/c/Users/ssdrive/Downloads/MinGW:/cygdrive/c/QNX/

> QNX_Momentics_IDE_2)

>

> Build iozone for QNX

>

> cc -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO

> iozone.c -o iozone_qnx.o

> make: cc: No such file or directory

> make: *** [makefile:1568: iozone_qnx.o] Error 127

>

>

> Need some suggestions on building the tool.

>

Makefile :

iozone_qnx.o: iozone.c libbif.c  libasync.c

@echo ""

@echo "Build iozone for QNX"

@echo ""

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) iozone.c -o iozone_qnx.o

$(CC) -c -O -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO  $(CFLAGS) libbif.c -o libbif.o

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) libasync.c -o libasync.o

fileop_qnx.o: fileop.c

@echo ""

@echo "Building fileop for QNX"

@echo ""

$(CC) -Wall -c -O $(CFLAGS) -D_FILE_OFFSET_BITS=64 fileop.c -o fileop_qnx.o

qnx: iozone_qnx.o libbif.o libasync.o fileop_qnx.o pit_server.o

$(CC) -O $(LDFLAGS) iozone_qnx.o libbif.o libasync.o -o iozone

$(CC)  -O fileop_qnx.o -o fileop

$(CC)  -O pit_server.o -o pit_server

> Regards,

post122256: How to build IOZONE tool on QNX SDP7.1

Lenin T — Fri, 12 May 2023 10:45:43 GMT

Hi All,

I want to build IOZONE tool for QNX OS systems.

I have tried to build the tool on QNX momentics IDE but facing some build errors as below :

C:\iozone\iozone3_506\src\current>make qnx

which: no ndk-build in (/cygdrive/c/QNX/QNX_Momentics_IDE_2/features/com.qnx.tools.jre.win32.x86_64_11.0.10/jdk-11.0.10+9-jre/bin/server:/cygdrive/c/QNX/QNX_Momentics_IDE_2/features/com.qnx.tools.jre.win32.x86_64_11.0.10/jdk-11

.0.10+9-jre/bin:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/windows/system32:/cygdrive/c/windows:/cygdr

ive/c/windows/System32/Wbem:/cygdrive/c/windows/System32/WindowsPowerShell/v1.0:/cygdrive/c/windows/System32/OpenSSH:/cygdrive/c/Users/ssdrive/AppData/Local/Microsoft/WindowsApps:/cygdrive/c/Python27:/cygdrive/c/Python27/Script

s:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/DAL:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/DAL:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Compon

ents/IPT:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/IPT:/cygdrive/c/ProgramData/chocolatey/bin:/usr/bin:/cygdrive/c/Windows/System32/drivers:/cygdrive/c/Users/ssdrive/AppData/Local/Microsoft/WindowsAp

ps:/cygdrive/c/Users/ssdrive/Downloads/MinGW:/cygdrive/c/QNX/QNX_Momentics_IDE_2)

Build iozone for QNX

cc -c -O -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO iozone.c -o iozone_qnx.o

make: cc: No such file or directory

make: *** [makefile:1568: iozone_qnx.o] Error 127

Need some suggestions on building the tool.

Regards,

post122255: Re: IOzone run on QNX

Lenin T — Fri, 12 May 2023 10:36:23 GMT

Hi All,

I have tried to build IOZONE on qnx sdp 7.1 , but facing some errors:

C:\Lenin\iozone\iozone3_506\src\current>make qnx

which: no ndk-build in (/cygdrive/c/QNX/QNX_Momentics_IDE_2/features/com.qnx.tools.jre.win32.x86_64_11.0.10/jdk-11.0.10+9-jre/bin/server:/cygdrive/c/QNX/QNX_Momentics_IDE_2/features/com.qnx.tools.jre.win32.x86_64_11.0.10/jdk-11

.0.10+9-jre/bin:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/windows/system32:/cygdrive/c/windows:/cygdr

ive/c/windows/System32/Wbem:/cygdrive/c/windows/System32/WindowsPowerShell/v1.0:/cygdrive/c/windows/System32/OpenSSH:/cygdrive/c/Users/ssdrive/AppData/Local/Microsoft/WindowsApps:/cygdrive/c/Python27:/cygdrive/c/Python27/Script

s:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/DAL:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/DAL:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Compon

ents/IPT:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/IPT:/cygdrive/c/ProgramData/chocolatey/bin:/usr/bin:/cygdrive/c/Windows/System32/drivers:/cygdrive/c/Users/ssdrive/AppData/Local/Microsoft/WindowsAp

ps:/cygdrive/c/Users/ssdrive/Downloads/MinGW:/cygdrive/c/QNX/QNX_Momentics_IDE_2)

Build iozone for QNX

cc -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO  iozone.c -o iozone_qnx.o

make: cc: No such file or directory

make: *** [makefile:1568: iozone_qnx.o] Error 127

Could you please suggest if anything i am missing here ?

Make file :

iozone_qnx.o: iozone.c libbif.c  libasync.c

@echo ""

@echo "Build iozone for QNX"

@echo ""

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) iozone.c -o iozone_qnx.o

$(CC) -c -O -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO  $(CFLAGS) libbif.c -o libbif.o

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) libasync.c -o libasync.o

fileop_qnx.o: fileop.c

@echo ""

@echo "Building fileop for QNX"

@echo ""

$(CC) -Wall -c -O $(CFLAGS) -D_FILE_OFFSET_BITS=64 fileop.c -o fileop_qnx.o

qnx: iozone_qnx.o libbif.o libasync.o fileop_qnx.o pit_server.o

$(CC) -O $(LDFLAGS) iozone_qnx.o libbif.o libasync.o -o iozone

$(CC)  -O fileop_qnx.o -o fileop

$(CC)  -O pit_server.o -o pit_server

Regards,

post122251: Re: IOzone run on QNX

Lenin T — Thu, 11 May 2023 02:21:27 GMT

Hi Janardhan & Will ,

I am trying to port iozone tool to QNX 7.1

Can i use the source from 'iozone.org' or is there any other soruce package for unix / qnx separately.

Could you please share the required changes to be done for porting the tool.

Regards,

Lenin

post122199: Unable to Enable Jumbo Frames for QNX 710 BSP

Bijo Philip(deleted) — Wed, 05 Apr 2023 07:00:41 GMT

HW information: NXP S32G VNP RDB2 (EVB)

BSP: BSP_nxp-s32g-evb_br-710_be-710 (SVN941162_JBN31)

I am trying to enable jumbo frames transmission and reception on NXP S32G Evaluation board, however, the MTU size is not changing above 1500. The following steps are done in dwcqeos.c to enable Jumbo Frames:

1) Enabled JE (Jumbo Frames Enable) and GPSLCE ( Giant Packet Size Limit Control Enable)

2) Disabled Store & Forward Mechanism (RSF, TSF)

3) MTU reconfigured from ETHERMTU to ETHERMTU_JUMBO

4) Media reconfigured from NIC_MEDIA_802_3 to NIC_MEDIA_CUSTOM

I have also tried executing the steps mentioned in the below link to change MTU manually, however, it is giving an error stating " SIOCSIFMTU: Invalid argument" for values above 1500:

http://www.qnx.com/developers/docs/qnxcar2/index.jsp?topic=%2Fcom.qnx.doc.core_networking.user_guide%2Ftopic%2Fdrivers_Jumbo.html

Any help would be greatly appreciated. Thank you in advance for you support!

post122178: Re: How to use Clang-Tidy with QNX code?

Jinlong Shen — Tue, 28 Mar 2023 20:41:03 GMT

I am trying this without success. I end up with not finding the standard c++ headers.

Have you succeeded?

post122163: How to change IRQ assignments to devices in QNX 6.3.2

Hardik Baldaniya — Mon, 13 Mar 2023 09:00:16 GMT

I have an x86 target (Advantech PCM-3353) running on QNX 6.3.2 Neutrino. System has 1 onboard Ethernet port (intel 82544) and 3 PCI cards - each having 2 Ethernet NICs (intel 82551T). The process 'io-net' with 'devn-speedo.so' driver module is used to get these PCI NIC cards working. We have observed that sometimes, io-net process fails with segmentation fault without disturbing other processes and sometimes entire OS freezes (WDT reset).

1. Can anybody help find root-cause OR solution to this issue?

2. I believe this has to do something with IRQ assignment by BIOS. BIOS assigns same IRQ to 2 of the 3 PCI NIC cards. Though this is valid theoretically, I suspect driver may not have handled this properly. I tried various BIOS settings without success to assign different IRQs to all PCI NIC cards. Is there any way in QNX to change IRQ assignments after QNX kernel has started?

post122100: 4G Modem interface in Linux Guest OS

Sachin Gole — Wed, 04 Jan 2023 13:04:14 GMT

Hello All,

4G Simcom A7672S modem is connected through USB on R-Car H3e. Inside QNX hypervisor host with devc-serusb driver, we are able to run pp0 interface and it gives ip from sim service provider. Internet is working.

Next, same we want to get inside Linux Guest VM,

1) Created a vdev as follows,

vdev virtio-console

     sched 7

     loc 0x5b130000

     intr gic:303

     hostdev /dev/serusb1

2) On linux guest we can see the device as hvc0. The path is /dev/hvc0. Logs shows incoming calls.

For ppp interface connect script is failing every time. Logs shows that AT command is not sent to modem as per below error log,

root@root:~# pppd

[   28.961403] PPP generic driver version 2.4.2

Connect script failed

root@root:~# plog

Jan  2 00:49:23 root chat[540]: abort on (ERROR)

Jan  2 00:49:23 root chat[540]: abort on (+CGATT: 0)

Jan  2 00:49:23 root chat[540]: send (AT^M)

Jan  2 00:49:23 root chat[540]: timeout set to 12 seconds

Jan  2 00:49:23 root chat[540]: expect (OK)

Jan  2 00:49:35 root chat[540]: alarm

Jan  2 00:49:35 root chat[540]: Failed

Jan  2 00:49:35 root pppd[533]: Connect script failed

Jan  2 00:49:35 root pppd[533]: Exit.

Our conclusion is that the AT command is not getting passed from guest to the modem device connected on host.

Question :

    Approach taking dev-serusb1 to guest os via vdev console is correct way? Or there is another approach.

    hvc0 in the dev is write protected or something need to be changed over there?

    We are also getting the device as /dev/vport0p0. But we are unable to use the device & what is the functionality of same. If we try to cat the vport0p0 then it says No such device or addresses

Thanks in Advance

post122080: Re: qnx-setup-2.0-202209011607-linux.run

Albrecht Uhlmann — Fri, 23 Dec 2022 12:48:13 GMT

QNX Software Center is a graphical application. You need a desktop environment running. Alternatively, you can use the command line version, run qnxsoftwarecenter_cli (or something) from the same folder.

-Al

post122079: Re: qnx-setup-2.0-202209011607-linux.run

ShAILESH AGRAHARI — Fri, 23 Dec 2022 11:10:58 GMT

Please help it is urgently required...if anybody can help.

post122078: qnx-setup-2.0-202209011607-linux.run

ShAILESH AGRAHARI — Fri, 23 Dec 2022 11:09:29 GMT

Hi Team,

while running installer "qnx-setup-2.0-202209011607-linux.run".

I am getting some error, please find the attachment,

I am running this installer on WSL2.

Thanks!

Shailesh

post122018: gcc Fortran on QNX

raymond balister — Thu, 08 Dec 2022 00:36:15 GMT

I've scoured the forums and discussion groups for any information regarding fortran on QNX, preferrably 6.5.0. I found several good discussions, but they all lead back to the same dead links. Is there any good, recent information on how to get fortran compiled on QNX? It seem that others have done this, I'd appreciate info on how to get this done. Note that f2c isn't an option as this is for a large library, and the C++ version of the library isn't scheduled for release until late 2023, if at all. Linux versions are available but I've had no success in getting Linux objects or libs to link on QNX. Not surprised at that.

Thanks.

Raymond Balister

ray.balister@l3harris.com

post122016: Lilliput display as a pass-through device to Linux(Bullseye) guest

Sachin Gole — Tue, 06 Dec 2022 13:47:39 GMT

Hello,

I want to give Lilliput display as a pass-through device to Linux guest(Bullseye) running on QNX host on NXP IMX8 QXP board. As per the folliwng QNX guide i made changes in the qvmconf file but it gives No such device on Host error:

http://www.qnx.com/developers/docs/7.1/index.html#com.qnx.doc.hypervisor.user/topic/vm/pass.html

I made following changes in the qvmconf file :

pass loc pci:0:2.0=pci:0:2.0

I want to know how can i get the pci details of Lilliput display. Is there any other way to pass it to guest? Kindly let me know. Thanks!!

post122015: Steps to Enable GSM Modem on QNX7.1.0 using PPPD

Sachin Gole — Thu, 01 Dec 2022 10:48:16 GMT

Hello QNX,

I am using QNX7.1.0 on i.Mx8QXP MEK board.

For connectivity I am using the GSM MODEM.

I have Enabled the PPPD daemon on QNX7.1.0

I am connecting the modem to board via USB, and following the below steps to start.

STEPS :-

1.Enable serusb to detect the modem connected to USB board.

   devc-serusb -d did=0x2303 vid=0x067b -b 115200

2.Enable the PPPD daemon to start the PPP session with modem.

   pppd /dev/serusb1 115200 nodetach noauth user ppp debug

From above command I am getting below logs.

   Using interface ppp0

   Connect: ppp0 <--> /dev/serusb1

   LCP: timeout sending Config-Requests

   auth.c: Connection terminated.

As per my understanding the LCP negotiation is getting failed.

My findings are PPPD is not able to start session due to PAP (Password Authentication Protocol) & CHAP (Challenge Handshake Authentication protocol) authentication gets failed.

I want to know the below points -

1.What content is required to put in PAP & CHAP scripts.

2.Path to keep the above scripts OR need to include in build file.

3.Any Example or document on how to enable MODEM on QNX with pppd daemon.

Thanks in advance

post122000: Interconnect applications deployed on different VMs (guests) in qnx hypervisor

Ayush Kumar(deleted) — Fri, 18 Nov 2022 12:09:05 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

post121998: Re: Target to build OpenDDS with QNX tool chain

Albrecht Uhlmann — Fri, 18 Nov 2022 12:04:02 GMT

I think there are at least two problems here:

- the build includes files from /usr/include/..., that is your Linux host. You need to tell it to llok into $QNX_TARGET/usr/include/.... What happens if you do "echo $QNX_TARGET" in the build console? It should point to your SDK installation. You may also check the generated env scripts.

- It appears the build expects gcc v9. QNX SDP supports gcc v8.3.0 - you find headers under $QNX_TARGET/usr/include/c++/8.3.0/*.

Regards,

Albrecht

post121996: led driver for Reneses Rcar H3 in QNX

kanchan devi(deleted) — Fri, 18 Nov 2022 07:10:26 GMT

Hello Sir,

I want to toggle led from QNX in RCAR H3 board can you please help with proper documentation or Driver required.

post121995: Target to build OpenDDS with QNX tool chain

Sachin Gole — Wed, 16 Nov 2022 15:07:52 GMT

Hello All,

we are trying to compile OpenDDS (v 3.20) with qnx compiler tool chain.

Here in OpenDDS having makefile based (MPC based) build system with configure script.

So we are calling configure script with cross-compilation option.

And following below command:

1 ./configure --target=linux-cross --target-compiler=/home/sdv-1011/qnx710/host/linux/x86_64/usr/bin/aarch64-unknown-nto-qnx7.1.0-g++

2 make

Here for more information we have attached logs files for both above mentioned commands

Actually we are getting error in make. It seems like make doesn't get files from qnx710.

So request to all, Anyone having any solution to build OpenDDS in qnx please suggest.

Thanks in advance

post121973: Anyone port or use a Kubernetes container like K3s to QNX / QNX7 ?

Jim P — Wed, 26 Oct 2022 16:56:29 GMT

Kubernetes, also known as K8s, is an open-source system for automating deployment, scaling, and management of containerized applications.

K3s is capable of nearly everything Kubernetes can do. It is just a more lightweight version.

Has anyone tried to port a Kubernetes (or lightwieght K3s version of a Kubernetes container to QNX?

References:

https://en.wikipedia.org/wiki/Kubernetes

https://docs.k3s.io/faq

https://www.kdab.com/wp-content/uploads/stories/Containers-in-Embedded.pdf

post121968: Performance issues when using conditional variables in QNX7

Emiliano Firmino(deleted) — Fri, 21 Oct 2022 17:47:58 GMT

Hi,

I notice issues when using notify_all() in conditional variables in QNX7 compare to others platforms like Linux. Causing in some cases to couple of milliseconds lost during a cycle. The program in question has over 100 conditional variables and can run in single or multi-thread by configuration but the conditional variable are not disabled or skipped when using single thread. There is noticeable lost time when calling multiple notify_all() in single threaded executions.

In QNX, std::condition_variable::notify_all() calls SyncCondvarSignal() that emits a system call that is clearly visible in the Kernel Tracing and Enter and Exit that takes a couple of microseconds usually 5us or more. When measuring using the attached cpp file without Kernel Tracing it takes about 1us. Compare to Linux it seems to be about 100 or 1000 slower where it is around 5ns.

Why notify_all() even when no threads are waiting for that signal takes so long? Is there any good reference on internal implementation in the kernel and expected performance for these system calls? What is the impact of kernel tracing on the performance of system calls like the ones from conditional variables?

Thanks

post121967: Can I generate external symbol files to do a postmortem?

Adrian Hawryluk — Fri, 21 Oct 2022 15:09:22 GMT

MS has .pdb files allowing you to do a postmortem on a production build without releasing the symbols to the public. Does the qnx build chain allow for something similar, and if so, how would I do this?

Thx.

post121966: Errcc: spawn of C:\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld failed: No such file or directory

Preethi Selvaraju — Thu, 20 Oct 2022 13:14:26 GMT

May I know , in general, When we will get this error?

I've cc: spawn of C:\qnx710\host\win64\x86_64\usr\bin\aarch64-unknown-nto-qnx7.1.0-ld failed file present in the mentioned path.

post121965: NXP S32G2 RDB2 BSP

Josh Hsiao — Thu, 20 Oct 2022 01:30:01 GMT

Hi,

We are using NXP S32G2 RDB2 for development. I found SDP 7.1 BSP for NXP S32G274A EVB/RDB in software center. I followed the instruction to build and boot from SD card, but it will stuck in "kernel start" and show "CPU 1 start failure" then reboot. Since the evb/rdb board is no longer available. Is there any plan to release SDP 7.1 BSP for NXP S32G2 RDB2?

Thanks,

Josh

post121943: fifo open failure

Annamalai Lakshmanan — Wed, 28 Sep 2022 12:14:04 GMT

HI,

I am getting an errno ENOSYS while trying to open a fifo created example like mkfifo /ramdisk/fifos/sc_daemon,

Is the FIFO not supported in SDP 7.1 ?

Regards,

Annamalai

post121893: How to flash QNX-IFS into eMMC using dfu-util

Nevin Joe Jerald — Thu, 11 Aug 2022 10:24:22 GMT

Hi,

We have J721EVM where we are trying to flash eMMC in it using dfu-util... We were able to successfully flash the other bootloader images using dfu-util the below dfu-util command... But, for Qnx-ifs, we don't see any command and not sure how to flash... Can anyone please suggest ?

HOST $ sudo dfu-util -a tiboot3.bin.raw -D /tiboot3.bin

HOST $ sudo dfu-util -a tispl.bin.raw -D /tispl.bin

HOST $ sudo dfu-util -a u-boot.img.raw -D /u-boot.img

HOST $ sudo dfu-util -a sysfw.itb.raw -D /sysfw.itb

Thanks,

Muthu

post121835: Re: How to pass EGL surface to another program through stream on 7.1

Hank Chen(deleted) — Fri, 24 Jun 2022 00:43:41 GMT

update context

//on static int initEGL(void) function change

//egl_surface = eglCreateWindowSurface(egl_display, egl_conf, screen_win, 0);

egl_surface = eglCreateWindowSurface(egl_display, egl_conf, screen_win, (EGLint *)&egl_surf_attr);

post121833: Re: How to pass EGL surface to another program through stream on 7.1

Hank Chen(deleted) — Thu, 23 Jun 2022 08:05:15 GMT

on gles-3d-cube.c  modify

screen_window_t screen_win;

screen_stream_t stream_p;//add stream

struct

{

    EGLint render_buffer[2];

    EGLint none;

} egl_surf_attr = {

    .render_buffer = {EGL_RENDER_BUFFER, EGL_BACK_BUFFER}, /* Ask for double-buffering */

    .none = EGL_NONE                                       /* End of list */

};

//on static int initScreen() function before return EXIT_SUCCESS;

//add stream-->

    int success = screen_create_stream(&stream_p, screen_ctx);

if (success == -1) {

printf("failed to create stream \n");

return -1;

}

int buffer_size[2] = {1080, 720};

screen_set_stream_property_iv(stream_p, SCREEN_PROPERTY_BUFFER_SIZE,buffer_size);

screen_set_stream_property_iv(stream_p, SCREEN_PROPERTY_FORMAT,(const int[]){SCREEN_FORMAT_RGBA8888|SCREEN_FORMAT_RGBX8888});

screen_set_stream_property_iv(stream_p, SCREEN_PROPERTY_USAGE,(const int[]){SCREEN_USAGE_OPENGL_ES2 | SCREEN_USAGE_WRITE |SCREEN_USAGE_NATIVE|SCREEN_USAGE_READ});

success = screen_create_stream_buffers(stream_p, 2);

if (success == -1) {

printf("failed to create stream buffer\n");

return -1;

}

int permissions;

screen_get_stream_property_iv(stream_p, SCREEN_PROPERTY_PERMISSIONS,&permissions);

// Allow processes in the same group to access the stream

permissions |= SCREEN_PERMISSION_IROTH;

permissions |= SCREEN_PERMISSION_IWOTH;

permissions |= SCREEN_PERMISSION_IXOTH;

screen_set_stream_property_iv(stream_p, SCREEN_PROPERTY_PERMISSIONS,&permissions);

//<--add stream

//on static int initEGL(void) function change

//egl_surface = eglCreateWindowSurface(egl_display, egl_conf, screen_win, (EGLint *)&egl_surf_attr);

    egl_surface = eglCreateWindowSurface(egl_display, egl_conf, screen_win, &egl_surf_attr);

//on void render() function change make 3d-cube work

// glDrawArrays(GL_LINES, 0, 6);

glDrawArrays(GL_POINTS, 0, 6);

//on int main(int argc, char *argv[]) befor while(1) add

screen_buffer_t stream_buf = nullptr;

//in while loop after check if (rc != EGL_TRUE) add

//add stream-->

        success = screen_get_stream_property_pv(  // buffers property

// A handle to the buffer or buffers available for rendering.

stream_p, SCREEN_PROPERTY_RENDER_BUFFERS, (void **)&stream_buf);

if (success == -1) {

printf("failed to get stream buffer\n");

  //return -1;

}

if (stream_buf == nullptr) {

  printf("failed to get stream buffer form stream\n");

  //return -1;

}else{

printf("%s:%d,stream_buf is not null\n",__FUNCTION__,__LINE__);

}

void *pointer = NULL;

success = screen_set_buffer_property_pv(stream_buf,SCREEN_PROPERTY_EGL_HANDLE,&pointer);

pointer = &egl_surf_attr.render_buffer[0];

if(pointer == NULL) printf("%s:%d,pointer == NULL\n",__FUNCTION__,__LINE__);

else printf("%s:%d,pointer is not NULL\n",__FUNCTION__,__LINE__);

success = screen_post_stream(stream_p, stream_buf, 0, nullptr, 0);

if (success == -1) {

printf("%s:%d,failed to post stream\n",__FUNCTION__,__LINE__);

}else{

printf("%s:%d,post stream success\n",__FUNCTION__,__LINE__);

}

        //<--add stream

==============================================================================

on Consumer  process refer from https://github.com/youngslab/qnx-stream-api-demo/blob/master/consumer/main.cpp

but that not work on SCREEN_PROPERTY_EGL_HANDLE stream

post121832: How to pass EGL surface to another program through stream on 7.1

Hank Chen(deleted) — Thu, 23 Jun 2022 06:06:14 GMT

I found this example from GitHub

https://github.com/kyo504/gles-tutorials/tree/master/gles-3d-cube

Excuse me, how to pass this 3D animation to another program through the stream buffer.

I tried using the SCREEN_PROPERTY_EGL_HANDLE parameter of screen_get_buffer_property_pv but it didn't work

I refer to Screen Graphics Subsystem Developer's Guide ,Chapter 5 Resource Sharing, but can't understand

Producer Stream --------> Consumer Stream --> Display

(gles-3d-cube) (consumer process)

post121826: Re: LVDS 8bit/4channel to LVDS 6bit/3channel

Pochieh Chuang(deleted) — Thu, 16 Jun 2022 07:51:37 GMT

Hi, I am also looking for wfd source of iMX8QM.

Anyone knows how to get it?

post121798: RAID Support and Use Case

Niraj Desai — Tue, 03 May 2022 18:48:55 GMT

I think we support intelligent HW RAID in QNX 7.1 but do not support SW RAID, right ?

Having said that, do we have any use case for RAID being used for Target failure ? One of our customer is facing with

both harddisk and other components failure resulting in huge downtime.

They are thinking to use RAID as below

"In our case it is most important that we keep the computer (controller) up and running or bring another computer back

up quickly with the old RAID array, when the computer fails. In all cases having RAID helps because it reduces the tool

down time as you can simply put the old RAID which may not have failed in a new computer and off you go. "

But they want to find out if any QNX customers have done this usecase and how is their experience ?

Any help, pointers ?

post121792: Re: problem with fslSATA detecting ssd mSATA

LeviSais213 LeviSais213(deleted) — Tue, 26 Apr 2022 23:41:44 GMT

Many users may be concerned about the slow speed of writing files to the disk when copying from an external drive. However, the cause is not the SSD itself but its load more often than not. Often, the drive performs several tasks that the user may not be aware of while the computer is running. For example, I recently had a problem with fsl SATA SSD mSATA detection. At a glance, it is not possible to determine the cause of the SSD's slow performance. Therefore, the most effective method is diagnostics through testing. It is not easy for me to do it myself, so for RAID RECOVERY(https://www.salvagedata.com/raid-data-recovery/), I turned to a company I know.

post121786: Re: How to support colored text with sshd?

Sorin Lazareanu(deleted) — Tue, 19 Apr 2022 17:04:01 GMT

Actually ANSI colors work. My mistake.

post121783: Re: How to support colored text with sshd?

Sorin Lazareanu(deleted) — Fri, 15 Apr 2022 16:07:28 GMT

ANSI escape actually works. My mistake.

post121782: How to support colored text with sshd?

Sorin Lazareanu(deleted) — Fri, 15 Apr 2022 14:07:26 GMT

Is it possible to support color, like ANSI escape sequences, for remote terminals? I am using GNOME Terminal to remotely connect over ssh to a QNX 7.0 VM and the escape sequences do not work.

Thank you.

post121760: I need help with QNX3.4

mat bek(deleted) — Fri, 11 Mar 2022 14:49:58 GMT

We have an old PC in the factory running QNX 3.4 we have an application inbaded to the QNX OS when it boots up.

The application is a user interface used to load some Lot numbers and masks for a specific lots, keep track of data from the Lots like rejects etc ... on the database.re

our problem is that the application freeze whenever we enter the lot number on the user interface, we have to re-reboot QNX every time to move to the next screen

I was wondering it there is codes to enter in CMD or anything that I can do to solve this problem, or a way to find out the original application inside qnx. Or someone that inderstande this software that I can call or we can bring him/her in the fix this problem for us. located in Rochester NY USA.

Thanks

post121723: Re: QNX 6.6 Monitor Settings front/back porch

Silva Baeva(deleted) — Sun, 13 Feb 2022 21:08:01 GMT

Hi Albrecht,

One more thing - I am attaching the usage message for the libWFDintel-drm.so library.

Silva

post121722: Re: QNX 6.6 Monitor Settings front/back porch

Silva Baeva(deleted) — Sun, 13 Feb 2022 20:55:56 GMT

Hi Albrecht,

I am sorry for the late reply. I am attaching a copy of my graphics.conf file. I can see libwfdcfg-sample.so and libWFDintel-drm.so. There is a use message for libWFDintel-drm.so, but there is no information on how to fix my problem. Any more ideas?

Thanks.

Silva

post121719: Re: QNX 6.6 Monitor Settings front/back porch

Albrecht Uhlmann — Wed, 09 Feb 2022 20:15:18 GMT

Oh you run on x86, is that so? Can you check the graphics.conf that you are using? It will name the libwfdcg-*.lib which provides the modes and timings to QNX Screen. Maybe it has a "use" message that explains some parameters that can be changed.

-Al

post121718: Re: QNX 6.6 Monitor Settings front/back porch

Silva Baeva(deleted) — Wed, 09 Feb 2022 19:39:04 GMT

Hi Albrecht,

Thank you for your reply. It's a VGA monitor (not HDMI). Does it make a difference?

Silva

post121717: Re: QNX 6.6 Monitor Settings front/back porch

Albrecht Uhlmann — Wed, 09 Feb 2022 15:49:56 GMT

Am I right in assuming that this is an HDMI monitor? Do you use the default HDMI wfdcfg library from QNX as the modelib?

You would need to modify its code to read the values from a file rather than using the hard-coded values.

Best regards,

Al

post121716: QNX 6.6 Monitor Settings front/back porch

Silva Baeva(deleted) — Wed, 09 Feb 2022 15:39:15 GMT

Hi,

We have developed an application under QNX 6.6 which uses the QNX Screen API to display graphics on the screen. Our client uses a monitor where the GUI screen appears skewed/shifted to the right. They can manually adjust the monitor settings to correct the problem. However, after a power down of the monitor, these settings are lost and they have to adjust them again. Is there a way to put the front/back porch settings in a configuration file, so they don't have to be adjusted every time?

Thanks you in advance for your help.

Silva

post121680: port grpc to qnx7.1.0

Yufei Xie(deleted) — Fri, 21 Jan 2022 23:41:09 GMT

Is it possible to port grpc to qnx7.1? I recently tried to port it, it seems that its dependency abseil-cpp has a lot of issues. Any success story or hints? Thanks

post121669: Re: probs with std c++ libs

David Haufe — Tue, 11 Jan 2022 20:30:39 GMT

Okay, this Problems with the STL are only if I try to compile with GNU 4.8.3.

When I use zur default 4.7.3 the cmath and string will BE linkable.

Someone has a Idea to fix the STLs with the 4.8.3 Version?

post121668: probs with std c++ libs

David Haufe — Tue, 11 Jan 2022 09:33:43 GMT

Hello,

I would like to use the <a href="https://github.com/nlohmann/json">nlohmann/json</a> library under QNX 6.6 for an ARMv7le.

Since this library is supposed to run under Vanilla C++11, but only under GCC 4.8. <

That's why Iadded <a href="https://community.qnx.com/sf/frs/do/viewRelease/projects.toolchain/frs.gcc.gcc_4_8">this one</a> under QNX 6.6.

Also the <a href="https://community.qnx.com/sf/frs/do/viewRelease/projects.toolchain/frs.binutils.binutils_2_25">Binutils 2.25</a> under qnx 6.6.

Now I can build the whole thing "a little" further. But I now get errors with some C++11 standard libraries under qnx, affected seem to be cmath, underlying math.h; cstdio; string;

The compiler outputs the following:

Description Resource Path Location Type

'isfinite' is not a member of 'std' parser.hpp ../external/json/include/nlohmann/detail/input line 261 C/C++ Problem

Description Resource Path Location Type

'isinf' is not a member of 'std' binary_writer.hpp ../external/json/include/nlohmann/detail/output line 199 C/C++ Problem

Description Resource Path Location Type

'signbit' is not a member of 'std' to_chars.hpp ../external/json/include/nlohmann/detail/conversions line 1063 C/C++ Problem

Description Resource Path Location Type

'snprintf' is not a member of 'std' binary_reader.hpp ../external/json/include/nlohmann/detail/input line 322 C/C++ Problem

Description Resource Path Location Type

'std::to_string' has not been declared iteration_proxy.hpp ../external/json/include/nlohmann/detail/iterators line 20 C/C++ Problem

Description Resource Path Location Type

'stoull' is not a member of 'std' json_pointer.hpp ../external/json/include/nlohmann/detail line 191 C/C++ Problem

For example, when I remove std:: I get this error message :'isfinite' was not declared in this scope

sincerely david haufe

post121667: Re: Developer for NUC980

Romanov Mishel — Wed, 22 Dec 2021 14:19:42 GMT

But we don't have time and free stuff to design it ourselves, so we want to hire remote developer(s)

post121666: Re: Developer for NUC980

Romanov Mishel — Wed, 22 Dec 2021 14:16:04 GMT

Thanks for the information, Albrecht! We assumed to use 6.4 or 6.5 as a reference design. We have already used at91sam9g-ekes BSP in out projects, and now we are thinking of using it.

post121665: Re: Developer for NUC980

Albrecht Uhlmann — Wed, 22 Dec 2021 12:56:17 GMT

Hi Romanov,

thanks for suggesting this. I am wondering if you are aware that current QNX versions do no longer support ARMv6. You would need to stick with 6.5, released 2012.

As far as I understand ARM926EJS is ARMv6. The lowest ARM version currently supported is ARMv7 (i.MX53, i.MX6 etc.)

-Al

post121663: Developer for NUC980

Romanov Mishel — Tue, 21 Dec 2021 08:43:55 GMT

Good afternoon! Our company wants to hire a developer(s) to make BSP for Nuvoton NUC980 chip. NUC980 - ARM926EJS based SoC with integrated RAM. For acquaintance: datashet https://www.nuvoton.com/resource-files/NUC980_Datasheet_EN_Rev1.0.pdf, all sources for Linux https://github.com/OpenNuvoton. We can provide free development board also. All interested - write here.This is remote development with payment by mutual agreement.

If You want to mail us, mail to dmitrii.grigorev@trei.biz

Thanks.

post121660: Porting Apache MiNiFi (C++) to QNX

Yan Titarenko(deleted) — Fri, 03 Dec 2021 01:19:35 GMT

Hello.

I was curious how to get started with porting https://github.com/apache/nifi-minifi-cpp to QNX.

Yan

post121659: QNX 7.1 Web Browser - PDF browser plug in ?

Jim P — Mon, 22 Nov 2021 17:24:56 GMT

Does anyone have any information on whether there is a PDF viewer browser plug in available for QNX Web Brower on 7.1 ?

Using x86.

post121628: Re: help!!!!!!!Unable to access /dev/hd0

Albrecht Uhlmann — Tue, 19 Oct 2021 08:49:05 GMT

Hello Anu,

we are unaware of any current hardware that supports QNX 6.3. To be honest, I am surprised to see that you got *that* far with the installation.

I think the Intel 945 chipset was one of the last that supports 6.3.

You could try to virtualize the system, which works pretty well if you can switch all communications to IP-based protocols.

Another approach would be a forward port to recent QNX versions.

Regards,

Albrecht

post121626: help!!!!!!!Unable to access /dev/hd0

Anu mohan m(deleted) — Sat, 16 Oct 2021 07:34:31 GMT

I am trying to install QNX 6.3.0 on PC

PC is equipped with SATA hard disk (1 tb )with 7th gen mother board and I3 6th gen processor.

It is able to load the software version and application but while booting in routine mode it shows this error

unable to acess "/dev/hdo"(2)

mount : can't mount /(type qnx4)

mount : possible reason : no such device or address

unable to start "mount" (2)

unable to start "ksh"(2)

Quires

1. Software version 6.3 is compactable with 7th gen mother board and I3 6th gen processor, if so how to resolve above error

2. If it not compatible which processor and mother board can be used ?

Kindly quide

Thanks in advance

Anu Mohan M

post121412: Re: syslog() function cannot log the message over 200 bytes???

Brian Schuth — Wed, 12 May 2021 19:15:41 GMT

Picking this up a few years later --

I just hit this today, was building up a logging framework on syslog then nearly fainted when I saw that the message length was 200 bytes and apparently not configurable (stupid me for thinking that messages could surely be at least 1024 bytes long).

Is there any way to get either a) source to this, or b) someone to update so that we can have a usable logging system? 200 bytes is barely enough to say "hello" in some of our domain work. Or, alternatively, has anyone used any other alternative daemon?

post121408: Re: How to use Eigen in QNX 6.4.1

Madhav Cham(deleted) — Fri, 07 May 2021 04:52:36 GMT

just compile this program using command line interface like this

go to the QNX installtion folder and run

source qnxsdp-env.sh

q++ -Vgcc_ntox86_64 -std=c++14 -I /home/madhav/eigen-3.3.9/ eigentest.cpp -o eigenexe1

thanks

post121267: Re: Block Device Level Encryption

Yi Li(deleted) — Wed, 17 Feb 2021 14:59:22 GMT

Is there a link to download fsf-merkle.so? Thanks.

post121265: Re: Specification or documentation related to QNX6 file system

Jean de Bonfils Lavernelle(deleted) — Tue, 16 Feb 2021 13:32:42 GMT

> I'd start with the Filesystems chapter of the System Architecture guide. Note

> that the documentation never refers to the QNX 6 filesystem; we use the

> official name, which is the Power-Safe filesystem.

Thank you sir, I am going to take a look at the System Architecture guide

Jean

post121264: Re: Specification or documentation related to QNX6 file system

Jean de Bonfils Lavernelle(deleted) — Tue, 16 Feb 2021 13:30:24 GMT

post121255: Re: Specification or documentation related to QNX6 file system

Steve Reid(deleted) — Wed, 10 Feb 2021 18:29:06 GMT

I'd start with the Filesystems chapter of the System Architecture guide. Note that the documentation never refers to the QNX 6 filesystem; we use the official name, which is the Power-Safe filesystem.

post121254: Specification or documentation related to QNX6 file system

Jean de Bonfils Lavernelle(deleted) — Wed, 10 Feb 2021 18:22:24 GMT

Hello. For a project, I have to study how the QNX6 file system works. I don't really know where to start. I have identified fs-qnx6.so which I should analyze by reverse-engineering. Does anyone have any documentation, code or specification related to QNX6 file system? Or any information which could help me... This would be very useful to me. Thanks. Have a nice day.

Jean

post121181: 2012 Fisker Karma EV Infotainment QNX

shawn shawn(deleted) — Thu, 28 Jan 2021 01:46:18 GMT

Does anyone know what QNX version was in the Fisker Karma in 2010?

post121154: fsencrypt now works

li weihua — Thu, 17 Dec 2020 05:14:24 GMT

Hello QNX experts,

I'm experimenting with fsencrypt on my board.

The storage medium is sdcard.

I start the sdcard driver with"devb-sdmmc-mx8x cam pnp,cache,quiet blk ra=64k:2048k,memory="sysram&below4G:sysram",maxio=256,cache=4m disk name=sd sdio idx=1,bs=cd=0x5D0C0000^22^694,verbose=1 qnx6 crypto=enable"

On the board:

# mkqnx6fs -E /dev/sd0t131.1

All files on /dev/sd0t131.1 will be lost!

Confirm filesystem re-format (y) or (n): y

Format fs-qnx6: 131068 blocks, 16384 inodes, 4 groups

# mount -t qnx6 /dev/sd0t131.1 /mnt

mount: Can't mount /mnt (type qnx6)

mount: Possible reason: Not supported

What's wrong？

Thanks.

Best regards.

LiWeiHua

post121148: Re: display configuration issue with QNX 6.5.0 SP1 x86 BIOS/APIC BSP rel3653

Stuart Little(deleted) — Wed, 09 Dec 2020 18:33:53 GMT

Thanks for the confirmation.

post121147: Re: display configuration issue with QNX 6.5.0 SP1 x86 BIOS/APIC BSP rel3653

Stuart Little(deleted) — Wed, 09 Dec 2020 18:33:29 GMT

Just noticed a typo/error:-

"but then does come back to show the Accept/

Revert popup."

should be "doesn't come back" i.e. remains blanked

post121145: Re: display configuration issue with QNX 6.5.0 SP1 x86 BIOS/APIC BSP rel3653

Albrecht Uhlmann — Tue, 08 Dec 2020 09:33:48 GMT

I also experienced that changing the resolution does not work on newer boards.

I configure the resolution explicitly in display.conf and leave it at that.

-Al

post121144: display configuration issue with QNX 6.5.0 SP1 x86 BIOS/APIC BSP rel3653

Stuart Little(deleted) — Mon, 07 Dec 2020 14:05:26 GMT

Have an unexpected display configuration problem while checking the BSP on new hardware.

When the display resolution is changed the screen blanks as per normal, but then does come back to show the Accept/Revert popup. The new display resolution can be blindly accepted as works OK after a forced reboot.

Basic 650 installation is OK.

650SP1 is OK

An ifs created with the BSP is OK, with no patches.

With full patches the fault is witnessed and the following errors logged:-

pci_init: found PCI device 8086:f31

io-display: aperture shm_ctl failed! (0x00000000)

/usr/photon/bin/io-graphics: gf_dev_attach('/dev/io-display/8086,0f31,0'): gf errno 7

phgrafx: restart.status = 0, restart.n_devices = 1

phgrafx: 0: vid=0x8086,did=0xf31,deviceindex=0 = [0: No error]

VGA primary : bus 0x0 dev/func 0x10

Found 1 PCI/AGP display devices

pci_init: found PCI device 8086:f31

io-display: aperture shm_ctl failed! (0x00000000)

Given the patches do not include any graphics drivers with the libc* libraries only applied the resolution change still works as expected, but with procnto included in the ifs the problem occurs.

Is this a known issue with a workaround?

post121048: Re: RTL8168 Network Card Driver

John Fortune(deleted) — Fri, 23 Oct 2020 08:16:46 GMT

Thanks for that , I am not that familiar with qnx. Any Suggestions where I might find the correct driver.

John

post121047: Re: RTL8168 Network Card Driver

Albrecht Uhlmann — Fri, 23 Oct 2020 07:54:15 GMT

This driver is for QNX 4.

post121046: Re: RTL8168 Network Card Driver

John Fortune(deleted) — Fri, 23 Oct 2020 07:46:58 GMT

Here is the attachment from the post

post121045: Re: RTL8168 Network Card Driver

John Fortune(deleted) — Fri, 23 Oct 2020 07:44:52 GMT

The driver on the link is not a .so file. so cant try it.

https://community.qnx.com/sf/discussion/do/listPosts/projects.community/discussion.qnx4_community_support.topc8226?pageSize=-1#post_post32091

post121044: Re: RTL8168 Network Card Driver

Albrecht Uhlmann — Fri, 23 Oct 2020 07:38:52 GMT

Can you send a link to that post please.

If it is that old, there is a considerable risk that the driver won't work properly with current hardware. Just as a hint.

-Al

post121043: Re: RTL8168 Network Card Driver

John Fortune(deleted) — Fri, 23 Oct 2020 07:16:46 GMT

Hi Al,

I have RTOS Runtime 6.3.2. The Archive was Post 31718 from Hugh Brown in 2009 with a attachment.

post121042: Re: RTL8168 Network Card Driver

Albrecht Uhlmann — Fri, 23 Oct 2020 06:48:06 GMT

Which QNX version are you looking at?

Which archive are you referring to?

-Al

post121041: RTL8168 Network Card Driver

John Fortune(deleted) — Fri, 23 Oct 2020 05:33:07 GMT

I am looking for the RTL8168.so Network Driver file. There is one in the archive but it is not a .so file.

post121033: Ping delays on network re-coonect

hamed davaneh(deleted) — Wed, 14 Oct 2020 11:37:46 GMT

Hi,

We have QNX 6.5 on a Freescale P1013 based board (A21B from Duagon, formerly MEN).

When we disconnect a LAN connection and then reconnect while for example we have a ping test, we have huge delays:

64 bytes from 192.168.101.130: icmp_seq=18 ttl=255 time=0 ms

64 bytes from 192.168.101.130: icmp_seq=19 ttl=255 time=0 ms

64 bytes from 192.168.101.130: icmp_seq=24 ttl=255 time=7547 ms

64 bytes from 192.168.101.130: icmp_seq=25 ttl=255 time=6548 ms

64 bytes from 192.168.101.130: icmp_seq=26 ttl=255 time=5548 ms

64 bytes from 192.168.101.130: icmp_seq=27 ttl=255 time=4548 ms

64 bytes from 192.168.101.130: icmp_seq=28 ttl=255 time=3548 ms

64 bytes from 192.168.101.130: icmp_seq=29 ttl=255 time=2549 ms

64 bytes from 192.168.101.130: icmp_seq=30 ttl=255 time=1549 ms

64 bytes from 192.168.101.130: icmp_seq=31 ttl=255 time=549 ms

64 bytes from 192.168.101.130: icmp_seq=32 ttl=255 time=0 ms

64 bytes from 192.168.101.130: icmp_seq=33 ttl=255 time=0 ms

Also when we have pings with large size messages, when we re-connect we get the following message:

ping: sendto: No buffer space available

ping: sendto: No buffer space available

ping: sendto: No buffer space available

.

.

And it never reconnects.

The BSP is used is similar to "Freescale P1010RDB", Any hint on how we can solve this problem?

post121006: QNX SDP 7.0 x86_64 vmware image - root passwords

Daniel Amsalem(deleted) — Fri, 02 Oct 2020 08:38:54 GMT

Hello

What is the passwords for root user in the above QNX vmware image ?

Kind regards,

Dany Amsalem

Embedded software

Arilou Cyber Security (part of NNG Group)

Arilou Technologies Ltd. | 7 HaAhim Bejerano st., Ramat-Gan, Israel |

M: +972 (54) 5750205 | Tel.: +972 (72)-222-1337 | Fax.: +972 (72)-222-1338 |

E-mail: dany.amsalem@nng.com<mailto:dany.amsalem@nng.com> | www.ariloutech.com<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ariloutech.com_&d=DwIFAg&c=yzoHOc_ZK-sxl-kfGNSEvlJYanssXN3q-lhj0sp26wE&r=YtmGHBrMV2UwJGa3fbjrG8vlLowY6D6-nYw70ZGhL7lPYiK5mgwrfx5h2ziZzct9&m=kKRMsHatdJOlx30dzdsAwSj2r14GZtQUiaYc4F8lN-s&s=VDAGEqZkv8ijh8Rgt5mEQSrprj2cH6EiNNdRas5k8MI&e= >

Follow us on LinkedIn<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.linkedin.com_company_3208092_&d=DwIFAg&c=yzoHOc_ZK-sxl-kfGNSEvlJYanssXN3q-lhj0sp26wE&r=YtmGHBrMV2UwJGa3fbjrG8vlLowY6D6-nYw70ZGhL7lPYiK5mgwrfx5h2ziZzct9&m=kKRMsHatdJOlx30dzdsAwSj2r14GZtQUiaYc4F8lN-s&s=61RAswxByfTYYkDuqF7TVD4Mgpf3U5_6bG37_pnuPBQ&e= >

This message, including its attachments, is confidential and proprietary to Arilou and NNG Group and the email address belongs to NNG Group. By replying to this message, you expressly acknowledge and consent that in addition to the addressee of the message, NNG Group's members and their employees and professionals may have access to the correspondence and its content.

Title and full ownership rights to any content in this message, including its attachments is and shall remain with Arilou and NNG Group and this message shall not be construed as creating, conveying, transferring, granting or conferring upon the recipient any rights, license or authority in or to the message and any content thereof, including its attachments.

________________________________

This message, including its attachments, is confidential and the email address belongs to NNG Kft. By replying to this message, you expressly acknowledge and consent that in addition to the addressee of the message, NNG Kft's other employees and professionals may have access to the correspondence and its content. Additional terms of NNG's email policy can be found here:

https://urldefense.proofpoint.com/v2/url?u=https-3A__www.nng.com_email-2Dpolicy_&d=DwIFAg&c=yzoHOc_ZK-sxl-kfGNSEvlJYanssXN3q-lhj0sp26wE&r=YtmGHBrMV2UwJGa3fbjrG8vlLowY6D6-nYw70ZGhL7lPYiK5mgwrfx5h2ziZzct9&m=kKRMsHatdJOlx30dzdsAwSj2r14GZtQUiaYc4F8lN-s&s=PNi26FFwQuQf1kQDx2tIYNmWjF4ZpWHJ4RKfWe_ALos&e=

By responding to this email you accept the email policy.

Ez az ?zenet - a mell?kleteit is ide?rtve - bizalmas ?s szakmai jelleg?, ?s az e-mail-c?m az NNG Kft.-hez tartozik. Amennyiben v?laszol az ?zenetre, kifejezett hozz?j?rul?s?t adja ahhoz, hogy a levelez?shez ?s annak tartalm?hoz a c?mzetten k?v?l az NNG Kft. b?rmely m?s alkalmazottja is hozz?f?rjen. Az NNG e-mail-szab?lyzat?val kapcsolatos egy?b rendelkez?seket itt tal?lja:

https://urldefense.proofpoint.com/v2/url?u=https-3A__www.nng.com_email-2Dpolicy_&d=DwIFAg&c=yzoHOc_ZK-sxl-kfGNSEvlJYanssXN3q-lhj0sp26wE&r=YtmGHBrMV2UwJGa3fbjrG8vlLowY6D6-nYw70ZGhL7lPYiK5mgwrfx5h2ziZzct9&m=kKRMsHatdJOlx30dzdsAwSj2r14GZtQUiaYc4F8lN-s&s=PNi26FFwQuQf1kQDx2tIYNmWjF4ZpWHJ4RKfWe_ALos&e=

Az e-mailre t?rt?n? v?laszad?ssal ?n elfogadja az e-mail-szab?lyzatot.

post120972: Getting loopback packet in pfil filter

manas sahoo — Wed, 30 Sep 2020 09:16:14 GMT

Hi,

In io-pkt I am using pfill hooks to get output and input packets on an ethernet interface. But when I sent the packet on local loopback port (lo0), the packet does not show of in pfill hooks. But the packet is received through tcpdump.

How to receive the loopback packets sent on local loopback port?

post120962: Receiving own sent UDP packet through nfm filter

manas sahoo — Tue, 29 Sep 2020 08:19:49 GMT

Hi,

In my controller, I am using QNX 6.5 SP1. I am sending a multicast packet to outside through 1 Ethernet interface, through nfm filter. The packet goes outside and consumed by other controllers on the same multicast group.

The sending controller is also waiting for packets on same multicast group. But this controller never gets its own sent packet (specifically sent from nfm filter). Is there anything specific I am missing, due to which the packet is not received in the same controller?

Regards,

Manas

post120922: Re: porting bcm43455 wifi driver(bcmdhd) to imx6 + qnx6.6.0

jing feng(deleted) — Mon, 07 Sep 2020 12:19:21 GMT

hi Albrecht Uhlmann

i am working on porting bcm wifi driver from linux to QNX7.0, would you please give me some example to do this job?

post120871: BSP7.0 altera cyclone 5

Markus Gäumann(deleted) — Tue, 28 Jul 2020 08:35:56 GMT

Is there an existing bsp for an altera cyclone 5 development kit for QNX7.0.0?

If not what do I have to change from a bsp for QNX6.6.0?

post120840: Re: Error in Final Launch Sequence (Unknown Response Error Attempting to Exec Shared Lib)

Emily Lin(deleted) — Tue, 14 Jul 2020 08:49:38 GMT

Same error here. Already tried IDE 2.1.2 && 7.0.4., both got same failure....

post120814: Re: File system encryption using fsencrypt

Nicholas Berryrdl(deleted) — Mon, 06 Jul 2020 13:41:11 GMT

https://www.google.com/

post120813: Re: File system encryption using fsencrypt

Nicholas Berryrdl(deleted) — Mon, 06 Jul 2020 13:40:49 GMT

I think it works! I encrypted it like this, but I think it was a long time ago

post120792: QNX6.6,emmc driver,GUID Partition Table problems for Jacinto 6

Su Jianbang — Fri, 19 Jun 2020 05:40:51 GMT

I use this command below to start MMC device:

devb-sdmmc-omap_generic sdmmc busno=1,verbose=7,partitions=on,cache=on blk naming=0#@,noatime,cache=2M cam cache,async

there are only emmc‘s physical partitions in /dev folder, there is no uda’s gpt partitions.

Does devb-sdmmc-omap_generic support to emmc's gpt partitions in QNX 6.6 platform?

If does, please tell me how to use devb-sdmmc-omap_generic.

If NOT, is there any patches can support this feature in QNX 6.6 platform

post120771: Re: File system encryption using fsencrypt

Daniel Wang(deleted) — Thu, 11 Jun 2020 21:39:35 GMT

> My understanding isthat fsencrypt works only in conjunction with the fs-qnx6.

> so on a block device.

> FFSv3, if there is support for encryption, must have its own solution, but I

> am not aware of any.

>

> What you could do is run devb-loopback on top af a large-enough file inside

> the FFSv3 partition, and launch encryption on that. Of course there is a

> performance penalty, but maybe it is acceptable.

>

> Regards,

> Albrecht

Hi Albrecht, Thank you for your response. Sorry I'm new to QNX. It might work for me. Could you explain a little bit more? Right now I start like this:

# Starting QSPI flash driver

on -R 0xD devf-qspi-xzynq-zcu102

flashctl -p /dev/fs0p0 -o 0xee00000 -l 0x400000 -e -f -v -U

flashctl -p /dev/fs0p1 -U

flashctl -p /dev/fs0p1 -v -m -n /mnt/flash-partition

How can I create a devb-loopback and encrypt it? Thank you!

post120765: Does SLM support additional groups?

Ilyas Hamadouche(deleted) — Thu, 11 Jun 2020 17:10:25 GMT

I have two groups defined in /etc/groups as:

  file-owner:x:300:user

  user:x:400:

when I run the command id on the terminal, I see file-owner is added as an additional group.

  # id user

  uid=300(user) gid=300(user) groups=300(user),400(file-owner)

However, running id from a script started bu SLM with <SLM:user>user:user</SLM:user> gives:

  uid=300(user) gid=300(user)

Is this a known bug?

This is causing me a problem because I cannot read a configuration file that is owned by the user/group file-owner, with zero permissions for others

   ls -l /tmp/my-configs.txt

   -r--r-----   1   file-owner   file-owner   1054552  Jun 11  12:22  /tmp/my-configs.txt

I run a process with id user:user but I cannot read the configuration file owned by file-owner.

post120760: Re: File system encryption using fsencrypt

Albrecht Uhlmann — Thu, 11 Jun 2020 12:16:56 GMT

My understanding isthat fsencrypt works only in conjunction with the fs-qnx6.so on a block device.

FFSv3, if there is support for encryption, must have its own solution, but I am not aware of any.

What you could do is run devb-loopback on top af a large-enough file inside the FFSv3 partition, and launch encryption on that. Of course there is a performance penalty, but maybe it is acceptable.

Regards,

Albrecht

post120759: Re: MMC util to read extcsd from emmc devices

Albrecht Uhlmann — Thu, 11 Jun 2020 12:14:01 GMT

Hi, I did a quick scan of the dcmd_*.h header files, but I didn't come across a device control command that would retrieve this information.

$QNX_TARGET/usr/include/mmc contains a header file with some structures, but I think this is what the devb-sdmmc uses internally.

-Al

post120757: Re: MMC util to read extcsd from emmc devices

Sriya Pradhan(deleted) — Thu, 11 Jun 2020 07:41:36 GMT

Hi, did you get any answer to what you had asked ? Using devb-sdmmc-* , could you read the emmc devices?

post120756: File system encryption using fsencrypt

Daniel Wang(deleted) — Wed, 10 Jun 2020 21:02:16 GMT

It seems to me that this is no block level encryption solution (like dm-crypt in Linux). I found fsencrypt in the QNX document. However, in the document it mentioned "The fsencrypt utility manages the encryption of a Power-Safe (fs-qnx6.so) filesystem. In order to use fsencrypt, you must have specified crypto=enable for fs-qnx6.so."

I have a flash device that is currently using FF3S file-system. Does fsencrypt works in this file system?

post120751: QNX7 egalax (devh-egalax.so) touchscreen driver

Daryl Tyson(deleted) — Tue, 09 Jun 2020 18:46:59 GMT

Hi I'm looking for the driver for this, the documentation for qnx 7 seems to indicate this exists but I cannot find the driver in my installation path for image creation. (devh-egalax.so)

post120386: Re: QCC: license check failed

Albrecht Uhlmann — Sun, 22 Mar 2020 17:40:26 GMT

QNX is a commercial OS, and you need to have a license to run the toolchain. There are some licenses available free of cost, e.g. for academic use, but you need to apply for them on the QNX homepage.

-Al

post120385: QCC: license check failed

Ilyas Hamadouche(deleted) — Sat, 21 Mar 2020 10:29:43 GMT

I am trying to compile some automotive c++ code targeting QNX. I always get the error license check failed.

QCC -V gives me the same error:

source qnx700/qnxsdp-env.sh

QCC -V

Anybody faced this before? Any idea how to solve it?

- Ilyas

post120216: Re: BSP 6.5SP1 to 6.5 port

Albrecht Uhlmann — Mon, 03 Feb 2020 09:14:43 GMT

Hi Ralph,

if the BSP is supplied by Adlink, I suggest to contact them if a downgrade is possible and at what cost.

The other option is to use the SP1 BSP supplied by Adlink, enjoy their support, and bring your application forward to SP1. This would give you a mucher better maintainable solution in the mid and long term.

Looking at the Adlink platform you mentioned, it's based on Baytrail SoC and Intel i210 network controllers, and getting this to work on 6.5.0 will probably cause a lot of headache.

Regards,

Albrecht

post120213: Re: BSP 6.5SP1 to 6.5 port

Ralph Kanzler(deleted) — Fri, 31 Jan 2020 17:49:25 GMT

Thanks, Here is more info...

Currently have 6.5.0 apps developed/installed, in the field, with photon.

Replacing, through attrition, a Lippert N270/945GSE PC104+ board(6.5.0 BSP) with an ADLINK CM2-BT2-E3825 board. The ADLINK board has, manufacture supplied, 6.5.0SP1 BSP and 7. We need a 6.5.0 BSP.

My initial thought was to mod the 6.5.0SP1 to 6.5.0.

Can a generic X86 BSP work? But the current list does not show one for 6.5.0. Is that still available?

post120212: Re: BSP 6.5SP1 to 6.5 port

Albrecht Uhlmann — Fri, 31 Jan 2020 16:27:03 GMT

I guess it can be built, because the layout and structure of BSPs and the QNX Momentics SDP have not changed between 6.5.0 and SP1. But I would not be surprised if it either

- does not boot up at all

- boots up and looks ok but behaves strangely later on

SP1 was a quite substantial step, and any subsequent releases/patches are mostly only targeted towards 6.5.0SP1.

It will very much depend on the platform that you are looking at. Can you give more details? Do you have a particular reason why you want to stick with 6.5.0?

Regards,

Albrecht

post120199: BSP 6.5SP1 to 6.5 port

Ralph Kanzler(deleted) — Wed, 29 Jan 2020 18:04:41 GMT

Can a 6.5.0 SP1 BSP be built on a 6.5.0 momentics IDE environment?

Is it just a mater of copying files from a similar BSP as the 6.4 to 6.5 migration?

post120181: Embedded GUI tools supported by QNX

Nagarajan Ramanantham(deleted) — Mon, 27 Jan 2020 14:09:31 GMT

Hi,

We are in the stage of analyzing and choosing the suitable embedded GUI framework for our medical device project which will run on QNX 6.6.

Can anyone please provide the list of GUI frameworks supported by QNX 6.6.

I am aware of Qt and PEG pro.

Your responses are really appreciated.

thanks

Nagaraj

post120160: Re: Does QNX 7 SDK VsomeIP and Boost libraries are open source for access?

Ilyas Hamadouche(deleted) — Mon, 20 Jan 2020 09:24:11 GMT

Hi,

If you want to use Boost and VsomeIP for QNX, download there source code and build them locally targeting QNX.

I don't think QNX ships VsomeIP and Boost.

- Ilyas

post120156: Does QNX 7 SDK VsomeIP and Boost libraries are open source for access?

Daniel Medina(deleted) — Fri, 17 Jan 2020 09:54:12 GMT

Hello,

I am working on a project with QNX as the target platform.

For enabling phase of the project, the project is using Open Source VsomeIP and Boost 7.

But in order to enable validation for QNX platform, we require QNX version of the software libraries.

Does the QNX 7 SDK Boost and VSomeIP software libraries are Open Source for download?

post120138: Re: IOzone run on QNX

Will Miles — Fri, 10 Jan 2020 19:09:57 GMT

Hi Shao,

QNX isn't Linux -- iozone's Linux makefile target calls for features like System V shared memory  that aren't available on QNX.   You will need to create a makefile target that enables the right set of features for QNX.    I built it with:

#

# QCC QNX build with threads, largefiles, async I/O

#

qnx: iozone_qnx.o libbif.o libasync.o fileop_qnx.o pit_server.o

$(CC) -O $(LDFLAGS) iozone_qnx.o libbif.o libasync.o -lsocket -o iozone

$(CC)  -O fileop_qnx.o -o fileop

$(CC)  -O pit_server.o -lsocket -o pit_server

iozone_qnx.o: iozone.c libbif.c  libasync.c

@echo ""

@echo "Build iozone for QNX"

@echo ""

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) iozone.c -o iozone_qnx.o

$(CC) -c -O -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO  $(CFLAGS) libbif.c -o libbif.o

$(CC) -c -O  -DNAME='"qnx"' -DHAVE_ANSIC_C -D_FILE_OFFSET_BITS=64 -DASYNC_IO $(CFLAGS) libasync.c -o libasync.o

fileop_qnx.o: fileop.c

@echo ""

@echo "Building fileop for QNX"

@echo ""

$(CC) -Wall -c -O $(CFLAGS) -D_FILE_OFFSET_BITS=64 fileop.c -o fileop_qnx.o

post120136: Re: IOzone run on QNX

Albrecht Uhlmann — Fri, 10 Jan 2020 08:03:13 GMT

Hi Shao,

I looked up the shmget, shmat and shmctl functions. They are POSIX, but that part ("XSI Shared Memory") is apparently not implemented under QNX.

What you will need to do is to re-write the code that operates with shared memory and make use of shm_open, shm_ctl functions as provided by QNX, see here: http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.lib_ref/topic/s/shm_ctl.html.

Maybe someone from QSS can comment on that, or is XSI Shared Memory planned to be supported in a later release?

What about the other guys in this thread, I understand that someone managed to get IOzone built for QNX some time back?

Regards,

Albrecht

post120135: Re: IOzone run on QNX

shao zongfan(deleted) — Fri, 10 Jan 2020 00:58:05 GMT

hi:

there also have a problem;

log as fllow:

qcc -static -O3 iozone_linux.o libbif.o libasync.o -o iozone -lc -lsocket

iozone_linux.o: In function `alloc_mem':

iozone.c:(.text+0x391c): undefined reference to `shmget'

iozone.c:(.text+0x3936): undefined reference to `shmat'

iozone.c:(.text+0x394d): undefined reference to `shmctl'

cc: /home/qt/Desktop/qnx700/host/linux/x86_64/usr/bin/i586-pc-nto-qnx7.0.0-ld error 1

makefile:171: recipe for target 'linux' failed

make: *** [linux] Error 1

post120134: Re: IOzone run on QNX

Albrecht Uhlmann — Thu, 09 Jan 2020 11:49:17 GMT

Hi Janardhan,

try to add libs to linker explicitly:

qcc -static -O3 iozone_linux.o libbif.o libasync.o -o iozone -lc -lsocket

Regards,

Albrecht

post120133: Re: IOzone run on QNX

shao zongfan(deleted) — Thu, 09 Jan 2020 08:24:24 GMT

i also some problem, can i get some help?

log as fllow:

qcc -static -O3 iozone_linux.o libbif.o libasync.o \

-o iozone

iozone_linux.o: In function `alloc_mem':

iozone.c:(.text+0x391c): undefined reference to `shmget'

iozone.c:(.text+0x3936): undefined reference to `shmat'

iozone.c:(.text+0x394d): undefined reference to `shmctl'

iozone_linux.o: In function `start_master_listen':

iozone.c:(.text+0x75eb): undefined reference to `socket'

iozone.c:(.text+0x7612): undefined reference to `setsockopt'

iozone.c:(.text+0x7635): undefined reference to `setsockopt'

iozone.c:(.text+0x765b): undefined reference to `setsockopt'

iozone.c:(.text+0x76ab): undefined reference to `bind'

iozone_linux.o: In function `master_listen':

iozone.c:(.text+0x7811): undefined reference to `accept'

iozone.c:(.text+0x7853): undefined reference to `listen'

iozone_linux.o: In function `child_send':

iozone.c:(.text+0x79b6): undefined reference to `gethostbyname'

iozone.c:(.text+0x7a09): undefined reference to `socket'

iozone.c:(.text+0x7a5a): undefined reference to `bind'

iozone.c:(.text+0x7ad9): undefined reference to `connect'

iozone_linux.o: In function `start_child_listen':

iozone.c:(.text+0x894b): undefined reference to `socket'

iozone.c:(.text+0x8972): undefined reference to `setsockopt'

iozone.c:(.text+0x8995): undefined reference to `setsockopt'

iozone.c:(.text+0x89bb): undefined reference to `setsockopt'

iozone.c:(.text+0x8a1d): undefined reference to `bind'

iozone_linux.o: In function `child_attach':

iozone.c:(.text+0x8ba3): undefined reference to `listen'

iozone.c:(.text+0x8bc3): undefined reference to `accept'

iozone_linux.o: In function `start_child_listen_async':

iozone.c:(.text+0x945b): undefined reference to `socket'

iozone.c:(.text+0x9482): undefined reference to `setsockopt'

iozone.c:(.text+0x94a5): undefined reference to `setsockopt'

iozone.c:(.text+0x94f1): undefined reference to `bind'

iozone_linux.o: In function `start_master_send':

iozone.c:(.text+0x9a81): undefined reference to `gethostbyname'

iozone.c:(.text+0x9ac9): undefined reference to `socket'

iozone.c:(.text+0x9b1a): undefined reference to `bind'

iozone.c:(.text+0x9b80): undefined reference to `connect'

iozone.c:(.text+0x9c7d): undefined reference to `inet_ntoa'

iozone_linux.o: In function `start_master_send_async':

iozone.c:(.text+0x9db5): undefined reference to `socket'

iozone.c:(.text+0x9e06): undefined reference to `bind'

iozone.c:(.text+0x9e50): undefined reference to `connect'

iozone_linux.o: In function `sp_start_master_send':

iozone.c:(.text+0xcc21): undefined reference to `gethostbyname'

iozone.c:(.text+0xcc69): undefined reference to `socket'

iozone.c:(.text+0xccba): undefined reference to `bind'

iozone.c:(.text+0xcd20): undefined reference to `connect'

iozone.c:(.text+0xce1d): undefined reference to `inet_ntoa'

iozone_linux.o: In function `sp_start_child_listen':

iozone.c:(.text+0xcef3): undefined reference to `socket'

iozone.c:(.text+0xcf1a): undefined reference to `setsockopt'

iozone.c:(.text+0xcf3d): undefined reference to `setsockopt'

iozone.c:(.text+0xcf90): undefined reference to `bind'

iozone.c:(.text+0xcfc2): undefined reference to `listen'

iozone.c:(.text+0xcfe7): undefined reference to `accept'

iozone_linux.o: In function `sp_start_master_listen':

iozone.c:(.text+0xd187): undefined reference to `socket'

iozone.c:(.text+0xd1ae): undefined reference to `setsockopt'

iozone.c:(.text+0xd1d1): undefined reference to `setsockopt'

iozone.c:(.text+0xd230): undefined reference to `bind'

iozone.c:(.text+0xd259): undefined reference to `listen'

iozone.c:(.text+0xd27e): undefined reference to `accept'

iozone_linux.o: In function `sp_start_child_send':

iozone.c:(.text+0xd521): undefined reference to `gethostbyname'

iozone.c:(.text+0xd569): undefined reference to `socket'

iozone.c:(.text+0xd5ba): undefined reference to `bind'

iozone.c:(.text+0xd620): undefined reference to `connect'

iozone_linux.o: In function `pit_gettimeofday':

iozone.c:(.text+0xe95f): undefined reference to `getaddrinfo'

iozone.c:(.text+0xe997): undefined reference to `socket'

iozone.c:(.text+0xe9b4): undefined reference to `connect'

iozone.c:(.text+0xe9c9): undefined reference to `freeaddrinfo'

iozone.c:(.text+0xea4f): undefined reference to `freeaddrinfo'

iozone.c:(.text+0xea80): undefined reference to `freeaddrinfo'

iozone.c:(.text+0xeab0): undefined reference to `gai_strerror'

cc: /home/qt/Desktop/qnx700/host/linux/x86_64/usr/bin/i586-pc-nto-qnx7.0.0-ld error 1

makefile:171: recipe for target 'linux' failed

make: *** [linux] Error 1

post120132: How to use Clang-Tidy with QNX code?

Ilyas Hamadouche(deleted) — Tue, 07 Jan 2020 10:19:14 GMT

I have a QNX specific code and I want to use clang-tidy for static analysis. Is this possible?

I am asking this because clang-tidy is used through the clang compiler, while QCC is based on GCC not clang.

post120080: LVDS 8bit/4channel to LVDS 6bit/3channel

Nagarajan Ramanantham(deleted) — Fri, 08 Nov 2019 11:00:08 GMT

Hi All,

We trying to replace 15" LCD/Touch panel display(1024x768) which has LVDS 4channel interface to 7" LCD/Touch panel(1024x600) which has LVDS 3channel interface.

The development board is SABRE Smart platform. We are not sure what are the changes we need to make in QNX BSP code in order for the 7" display to work.

We are also trying to find WFDCfg library in our BSP source tree, but it is not available.

Your response is really appreciated.

Thanks

Nagaraj

post120048: Test suite to validate QNX

Koteswararao Nayudu(deleted) — Wed, 16 Oct 2019 10:01:35 GMT

Hi,

I am looking for the Test suite to validate QNX kernel like we have "LTP" (https://github.com/linux-test-project/ltp)

for Linux kernel.

Please can some one provide your suggestion to get the right test suite to validate QNX?

Regards,

Koti

post120020: QNX - OpenGL issues

Alexandra Mocanu(deleted) — Thu, 03 Oct 2019 14:24:56 GMT

Hello again,

So I managed to get the Screen API to work, and also get the other functionalities. But there are still problems with OpenGL, even though the example apps work correctly. (my previous post can be found here: http://community.qnx.com/sf/discussion/do/listPosts/projects.community/discussion.community.topc27748).

Currently, I'm facing two different behaviors of the application. First, when I'm debugging with Momentics, my app stops in "eglGetDisplay", because it cannot find a proper EGL display and so it returns a "egl_no_display". Second, when I'm just running the app on the board, the application stops in "eglCreateWindowSurface" (and also "eglCreatePlatformWindowSurface", because I also tried both), and the app just stops and returns a "Memory fault".

Do you maybe know why we have these two different behaviors? Or maybe what could cause these problems?

Thank you!

post120009: QNX7 OpenGL Screen Issues

Alexandra Mocanu(deleted) — Mon, 30 Sep 2019 13:45:23 GMT

Hello,

I'm working on a M3N board that is supposed to be running QNX Neutrino and I'm using a built BSP from the H3 BSP Package. I'm booting with U-Boot and my host system is Windows10, I'm also using Momentics with an ethernet connection to debug.

The issue I'm having is that I cannot seem to be getting the Screen API to work correctly with OpenGL. No example app from the usr/bin folder works. They all generate errors like "eglGetDisplay: cannot access a requested resource", "eglInitialize: EGL is not initialized, or could not be initialized, for the specified display". I also have some other issues with the Neutrino, like the mkdir command not working, or screeninfo, also i don't have a display.conf file and I can't get io-display to work (i saw that sometimes you also have to configure this s thats why i tried it).

Do i need some other libraries or some other configurations? I added the graphics libraries from "qnx7\aarch64le\usr\lib\graphics\rcarm3n" in the /tmp folder because it was the only folder i could edit, and also set the $GRAPHICS_ROOT as this folder.

Could you please give me some hints on what I could try to do next? Thank you!

post119926: Re: TARFS Resource Manager support

Muthu Vengataraman — Fri, 13 Sep 2019 05:36:58 GMT

Hi Steve

Did you had a chance to look at this?

Do you have any update?

Thanks

V.Muthusubramanian

post119925: Re: Block Device Level Encryption

Muthu Vengataraman — Fri, 13 Sep 2019 05:35:37 GMT

Hi

Do you have any update on this?

Thanks

V.Muthusubramanian

post119879: virtual serial port

Tejus S(deleted) — Fri, 06 Sep 2019 04:24:08 GMT

How do I implement a virtual serial port on QNX ?

I have an application that is currently listening to a real serial port . I have a change request to fetch the dat from anoter source (TCP) . I do not want to change my existing application that listens to the serial port . Hence I need to create a virtual COM port . Is a virtual serial port idea viable ?

post119808: Re: QNX7: Missing pci_hw-Intel_x86.so

Michael Kurt — Fri, 16 Aug 2019 12:31:40 GMT

That was too easy! ;-) Found it.

Thank You very much!

With best regards,

Michael Kurt.

post119807: Re: QNX7: Missing pci_hw-Intel_x86.so

Albrecht Uhlmann — Fri, 16 Aug 2019 11:49:13 GMT

It has to be explicitly installed via QNX Software Center (look for "PCI HW module Intel" or something, or just type "pci" in the search field of the "Available" tab)

post119806: QNX7: Missing pci_hw-Intel_x86.so

Michael Kurt — Fri, 16 Aug 2019 10:01:55 GMT

Dear community,

with QNX7-SDP (latest updates installed) I'm trying to create a boot image for an x86 target PC with pci-server.

As far as I know, I have to specify a hardware module for pci-server to run.

For my case, this should be pci_hw-Intel_x86.so, but I'm unable to find the file in my installation

(neither in 'host' nor 'target' directories as well not in the bsp-archive (BSP_x86[-64]_br-mainline_be-700_SVN848888_JBN5.zip)).

The VMware-image provided with the SDP (qnx700.x86_64.7.0.0.SGA201703011126) contains the pci_hw-Intel_x86.so library in it's boot image.

Where did this come from? Did I miss to install something?

With best regards,

Michael Kurt.

post119774: Re: 'dumper' running on non-root does not work without secpogenerate

Roger Maclean — Fri, 28 Jun 2019 20:22:20 GMT

I tried it and dumper seems to run with the write uid.  I'm not sure if I'm running exactly the same version as you but I suspect it doesn't matter.

Looking into dumper code, it appears that there is a bug which could cause this.  You'll see the policy includes:

    xprocess_mem_read:100

this is what allows dumper to read the memory of the process that it generates a core file for.  However it is subranged by uid so in this case it has been given the ability only to dump processes with a uid of 100.  If a process with a different uid crashes, dumper switches to uid 0 to dump it's memory.  This however fails and dumper gives up but in doing so neglects to switch back.

If this is the case, if you remove the ':100' from xprocess_mem_read in the policy the issue might go away.  This is actually one of the cases where you do tend to have to alter the policy generated by secpolgenerate.  Secpolgenerate tries to give the minimal abilities which is appropriate in most cases but there are some exceptions.

On 2019-06-28, 10:33 AM, "Gauresh Badve" <community-noreply@qnx.com> wrote:

    Please find the rules for the dumper:



    # === Rules for type dumper_t ============

    allow_attach dumper_t {

        /proc/dumper

    };

    allow dumper_t self:ability {

        unlock  noinherit

        # allow all

        setuid

        # allow all

        setgid

        # allow nonroot

        mem_peer

        # allow all

        priority

        # allow root

        able_priv

    };

    allow dumper_t self:ability {

        pathspace

        prot_exec

        map_fixed

        public_channel

        prot_write_and_exec

    };

    allow dumper_t self:ability {

        channel_connect:slogger2_t,devb_t

        xprocess_mem_read:100

    };









    _______________________________________________



    General

    http://community.qnx.com/sf/go/post119773

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

post119773: Re: 'dumper' running on non-root does not work without secpogenerate

Gauresh Badve(deleted) — Fri, 28 Jun 2019 14:33:28 GMT

Please find the rules for the dumper:

# === Rules for type dumper_t ============

allow_attach dumper_t {

    /proc/dumper

};

allow dumper_t self:ability {

    unlock  noinherit

    # allow all

    setuid

    # allow all

    setgid

    # allow nonroot

    mem_peer

    # allow all

    priority

    # allow root

    able_priv

};

allow dumper_t self:ability {

    pathspace

    prot_exec

    map_fixed

    public_channel

    prot_write_and_exec

};

allow dumper_t self:ability {

    channel_connect:slogger2_t,devb_t

    xprocess_mem_read:100

};

post119772: Re: 'dumper' running on non-root does not work without secpogenerate

Roger Maclean — Fri, 28 Jun 2019 14:06:24 GMT

Can you provide all the rules from your policy for dumper's type.

You shouldn't have to add any rules manually. Any abilities that dumper tries to configure manually should have been left in their default state by the policy so should behave the same with or without a security policy.  Dumper exits with a message to stderr if it fails to switch to the specified uid.

On 2019-06-28, 9:43 AM, "Gauresh Badve" <community-noreply@qnx.com> wrote:

    Yes log_1 represents the behavior when we are running the system without secpolgenerate.



    So here are the steps we followed:



    Situation 1:



    1) Enabled: LD_PRELOAD=/proc/boot/secpol-preload

                          secpolgenerate -u -t 100



    2) Give the command "pidin ar"

         I could see the following log:

       pid Arguments

           1 procnto-smp-instr -v -mr

    4098 secpolgenerate -u -t 100

     8196 /proc/boot/pipe

     12293 /proc/boot/slogger2 -U 111:116

     16390 /proc/boot/dumper -U 112

     20487 /proc/boot/random -pt -U 113:105



    3) Give the command "ps -e -o "user,CMD" and I see the following log:

    USER     CMD

    0        procnto-smp-instr

    0        secpolgenerate

    0        on

    110      pipe

    111      slogger2

    112      dumper

    113      random

        Here I see that the dumper is running through user ID 112.



    4) Check the generated policy at /dev/secpolgenerate/policy.



    Situation 2:



    1) Use the policy generated from Situation 1.



    2) Disabled: LD_PRELOAD=/proc/boot/secpol-preload

                          secpolgenerate -u -t 100



    3) Give the command "pidin ar".

         Here I see the following log

       pid Arguments

           1 procnto-smp-instr -v -mr

    8195 /proc/boot/pipe

    12292 /proc/boot/slogger2 -U 111:116

    16389 /proc/boot/dumper -U 112

    20486 /proc/boot/random -pt -U 113:105



    4) Give the command "ps -e -o "user,CMD" and I see the following log:

    USER     CMD

    0        procnto-smp-instr

    110      pipe

    111      slogger2

    0        dumper

    113      random

    Now I see that the dumper runs on root(UserID 0) instead of userid 112.



    To check further I killed an application "kill -SIGBUS PID_of_App" but I still see the dumper running on root.

    I believe the secpolgenerate should generate the ability required to run the dumper on userid 112. Do we need to add some abilities manually?











    _______________________________________________



    General

    http://community.qnx.com/sf/go/post119771

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

post119771: Re: 'dumper' running on non-root does not work without secpogenerate

Gauresh Badve(deleted) — Fri, 28 Jun 2019 13:43:04 GMT

Yes log_1 represents the behavior when we are running the system without secpolgenerate.

So here are the steps we followed:

Situation 1:

1) Enabled: LD_PRELOAD=/proc/boot/secpol-preload

                      secpolgenerate -u -t 100

2) Give the command "pidin ar"

     I could see the following log:

   pid Arguments

       1 procnto-smp-instr -v -mr

4098 secpolgenerate -u -t 100

8196 /proc/boot/pipe

12293 /proc/boot/slogger2 -U 111:116

16390 /proc/boot/dumper -U 112

20487 /proc/boot/random -pt -U 113:105

3) Give the command "ps -e -o "user,CMD" and I see the following log:

USER     CMD

0        procnto-smp-instr

0        secpolgenerate

0        on

110      pipe

111      slogger2

112      dumper

113      random

    Here I see that the dumper is running through user ID 112.

4) Check the generated policy at /dev/secpolgenerate/policy.

Situation 2:

1) Use the policy generated from Situation 1.

2) Disabled: LD_PRELOAD=/proc/boot/secpol-preload

                      secpolgenerate -u -t 100

3) Give the command "pidin ar".

     Here I see the following log

   pid Arguments

       1 procnto-smp-instr -v -mr

8195 /proc/boot/pipe

12292 /proc/boot/slogger2 -U 111:116

16389 /proc/boot/dumper -U 112

20486 /proc/boot/random -pt -U 113:105

4) Give the command "ps -e -o "user,CMD" and I see the following log:

USER     CMD

0        procnto-smp-instr

110      pipe

111      slogger2

0        dumper

113      random

Now I see that the dumper runs on root(UserID 0) instead of userid 112.

To check further I killed an application "kill -SIGBUS PID_of_App" but I still see the dumper running on root.

I believe the secpolgenerate should generate the ability required to run the dumper on userid 112. Do we need to add some abilities manually?

post119770: Re: 'dumper' running on non-root does not work without secpogenerate

Roger Maclean — Fri, 28 Jun 2019 12:31:23 GMT

It's not completely clear to me what you've done.  You need to set LD_PRELOAD only when you’re using secpolgenerate since its purpose is to communicate to secpolgenerate information about procmgr_ability calls.

Does log_1 represent the behavior when you're running the system without secpolgenerate using the generated policy?

Even if run with the -U option, dumper will run as root at times since it needs to acquire additional privileges when it dumps a core file.  So if you did happen to catch it after something crashed, this might be perfectly normal.

On 2019-06-28, 7:41 AM, "Gauresh Badve" <community-noreply@qnx.com> wrote:

    Thank you for the solution.



    Yes the environment variable "LD_PRELOAD" was not set to /proc/boot/secpol-preload.so. After adding the following change:

    LD_PRELOAD=/proc/boot/secpol-preload.so

    I could see that "pidin ar" command shows the dumper running as below



         pid Arguments

           1 procnto-smp-instr -v -mr

        8195 /proc/boot/pipe

       12292 /proc/boot/slogger2 -U 111:116

       16389 /proc/boot/dumper -U 112

       20486 /proc/boot/random -pt -U 113:105



    But when I run the command " ps -e -o "user,CMD", I get the output as dumper still running on root. Please find the below log_1:



    USER     CMD

    0        procnto-smp-instr

    110      pipe

    111      slogger2

    0        dumper

    113      random



    However with the secpolgenerate enabled I see that dumper running on non-root. Please find the below log_2:



    USER     CMD

    0        procnto-smp-instr

    0        secpolgenerate

    110      pipe

    111      slogger2

    112      dumper

    113      random



    Note:  The log_1 uses the same policy generated by the secpolgenerate where we see the dumper running on non-root in log_2.







    _______________________________________________



    General

    http://community.qnx.com/sf/go/post119769

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

post119769: Re: 'dumper' running on non-root does not work without secpogenerate

Gauresh Badve(deleted) — Fri, 28 Jun 2019 11:41:49 GMT

Thank you for the solution.

Yes the environment variable "LD_PRELOAD" was not set to /proc/boot/secpol-preload.so. After adding the following change:

LD_PRELOAD=/proc/boot/secpol-preload.so

I could see that "pidin ar" command shows the dumper running as below

     pid Arguments

       1 procnto-smp-instr -v -mr

    8195 /proc/boot/pipe

   12292 /proc/boot/slogger2 -U 111:116

   16389 /proc/boot/dumper -U 112

   20486 /proc/boot/random -pt -U 113:105

But when I run the command " ps -e -o "user,CMD", I get the output as dumper still running on root. Please find the below log_1:

USER     CMD

0        procnto-smp-instr

110      pipe

111      slogger2

0        dumper

113      random

However with the secpolgenerate enabled I see that dumper running on non-root. Please find the below log_2:

USER     CMD

0        procnto-smp-instr

0        secpolgenerate

110      pipe

111      slogger2

112      dumper

113      random

Note:  The log_1 uses the same policy generated by the secpolgenerate where we see the dumper running on non-root in log_2.

post119768: Re: 'dumper' running on non-root does not work without secpogenerate

Roger Maclean — Thu, 27 Jun 2019 14:14:19 GMT

When using secpolgenerate are you setting the environment variable LD_PRELOAD to /proc/boot/secpol-preload.so?  This is currently required since it provides the only way of secpolgenerate from learning what abilities a process tries to configure using calls to procmgr_ability.  If you don't have this, the policies it generates will lock the abilities which will result in the calls failing when used without secpolgenerate.

If you look at the rules in the policy for dumper_t you should find a number of rules show up as unlocked.

On 2019-06-27, 9:59 AM, "Gauresh Badve" <community-noreply@qnx.com> wrote:

    Let me explain my scenario:

    Version : QNX 7.0.0



    I am using secpolgenerate utility to generate security policy for the dumper. The secpolgenerate observes all the behavior of the dumper and creates a policy file at /dev/secpolgenerate/policy.



    The command I used for secpolgenerate:

    'secpolgenerate -u -t 100'



    I start the dumper using command:

    on -T dumper_t /proc/boot/dumper -U 112 ( Here 112 is the user id on which I want the dumper to run and dumper_t is the type defined for dumper in the security policy file )



    The "pidin ar" command shows the dumper running with user id - 112 and also the policy file generated in location /dev/secpolgenerate/policy shows all the abilities the type dumper_t require to run the dumper.



    The same generated policy( at /dev/secpolgenerate/policy) is added into the security policy document. The new security policy document is compiled using secpolcompile utility and the binary of the security policy is added into /proc/boot/secpol.bin.



    This time I do not run the command 'secpolgenerate -u -t 100' and run the dumper "on -T dumper_t /proc/boot/dumper -U 112".



    Now the "pidin ar" command does not show the dumper running.



    I believe the secpolgenerate utility generate all the abilities require by the dumper to run. But with the same abilities I don't see the dumper running on non-root i.e. /proc/boot/dumper -U 112 when we disable the command 'secpolgenerate -u -t 100'.









    _______________________________________________



    General

    http://community.qnx.com/sf/go/post119767

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

post119767: 'dumper' running on non-root does not work without secpogenerate

Gauresh Badve(deleted) — Thu, 27 Jun 2019 13:59:51 GMT

Let me explain my scenario:

Version : QNX 7.0.0

I am using secpolgenerate utility to generate security policy for the dumper. The secpolgenerate observes all the behavior of the dumper and creates a policy file at /dev/secpolgenerate/policy.

The command I used for secpolgenerate:

'secpolgenerate -u -t 100'

I start the dumper using command:

on -T dumper_t /proc/boot/dumper -U 112 ( Here 112 is the user id on which I want the dumper to run and dumper_t is the type defined for dumper in the security policy file )

The "pidin ar" command shows the dumper running with user id - 112 and also the policy file generated in location /dev/secpolgenerate/policy shows all the abilities the type dumper_t require to run the dumper.

The same generated policy( at /dev/secpolgenerate/policy) is added into the security policy document. The new security policy document is compiled using secpolcompile utility and the binary of the security policy is added into /proc/boot/secpol.bin.

This time I do not run the command 'secpolgenerate -u -t 100' and run the dumper "on -T dumper_t /proc/boot/dumper -U 112".

Now the "pidin ar" command does not show the dumper running.

I believe the secpolgenerate utility generate all the abilities require by the dumper to run. But with the same abilities I don't see the dumper running on non-root i.e. /proc/boot/dumper -U 112 when we disable the command 'secpolgenerate -u -t 100'.

post119765: Re: qtalk transfer file

Albrecht Uhlmann — Thu, 20 Jun 2019 09:19:05 GMT

${QNX_TARGET}/${CPU}/usr/bin/qtalk

It is always there.

It might not be on your target system, you need to simply copy it there, or include it in the IFS/rootfs build script

-Al

post119764: Re: A/B (Dual Copy) Booting in QNX.

Albrecht Uhlmann — Thu, 20 Jun 2019 09:16:11 GMT

Not sure what you actually want to achieve, could you be more specific?

Do you want to have primary IFS for normal operations, and boot from secondary IFS in case of some abnormal condition?

Which QNX version are you using?

Which hardware platform are you using?

-Al

post119763: Re: is any doc guide to write CAN bus driver MCP2515 for QNX

Albrecht Uhlmann — Thu, 20 Jun 2019 09:14:34 GMT

I think writing CAN driver for this QNX release means designing and implementing a resource manager from scratch.

With newer versions, QNX provides libio-can.a library and a standardized interface to CAN, meaning you would *only* need to implement the hardware layer to serve the actual chip. But I am not sure if this framework is available for 6.4.1.

-Al

post119762: Re: Max disk size supported by QNX 6.5.0

Albrecht Uhlmann — Thu, 20 Jun 2019 09:11:49 GMT

I think this question can only be answered by QSS engineers because the disk drivers are not public source.

post119757: Re: Max disk size supported by QNX 6.5.0

Jagat Bhusan(deleted) — Thu, 13 Jun 2019 13:24:50 GMT

I am still waiting for a response to the query .

post119756: Re: Max disk size supported by QNX 6.5.0

Jagat Bhusan(deleted) — Thu, 13 Jun 2019 13:23:31 GMT

> Hi,

>

> Thanks for your immediate response.

> In the link provided by you , it is mentioned as :

>

> For QNX6 FS

> -----------------

> Physical disk sector

>     32-bit (2 TB), using the devb API.

>

> Does it mean that the max size of disk / USB is limited to 2TB ?

>

> For QNX4 FS

> -----------------

> Disk size

>     2^64 bytes; limited by the disk driver.

>

> Does it mean that the max size supported by qnx4 fs is 2^64 bytes.

> If so, what is the meaning of "limited by disk driver" ? How to know about the

>  max size supported by the driver ?

>

> Thanks

>

> Jagat

>

>

post119755: is any doc guide to write CAN bus driver MCP2515 for QNX

Chris Liang(deleted) — Wed, 12 Jun 2019 14:47:45 GMT

Hi all,

I am trying to develop MCP2515 CAN bus driver for QNX 6.4.1. The platform arch is PPC440.

I am wondering anyone can point me where I should start? or is similar to linux device driver development? Linux has kernel code and user code, I would assume it is different since qnx is micro kernel.

Please..

Thanks in advance.

post119748: A/B (Dual Copy) Booting in QNX.

Michael Delaney(deleted) — Sun, 09 Jun 2019 18:19:50 GMT

I'm looking to setup a A / B (Dual Copy) boot sequence with QNX. Are there any documents, or working examples, on how to set this up?

post119743: Re: qtalk transfer file

khiemspdt lq(deleted) — Fri, 07 Jun 2019 11:51:59 GMT

Thank you,

But currently I've checked in my system don't have qtalk.

Do you know where I can get qtalk binary or source code ?

Thanks.

post119742: Re: Max disk size supported by QNX 6.5.0

Jagat Bhusan(deleted) — Fri, 07 Jun 2019 08:32:52 GMT

Hi,

Thanks for your immediate response.

In the link provided by you , it is mentioned as :

For QNX6 FS

-----------------

Physical disk sector

32-bit (2 TB), using the devb API.

Does it mean that the max size of disk / USB is limited to 2TB ?

For QNX4 FS

-----------------

Disk size

2^64 bytes; limited by the disk driver.

Does it mean that the max size supported by qnx4 fs is 2^64 bytes.

If so, what is the meaning of "limited by disk driver" ? How to know about the max size supported by the driver ?

Thanks

Jagat

There is an expli

post119741: Re: Detecting unaligned access during runtime

Albrecht Uhlmann — Fri, 07 Jun 2019 08:13:10 GMT

I think that a thread doing unaligned access creates a fault at processor level that gets translated to a signal (SIGBUS I think, but not sure). You can catch that in your application, but not from outside.

To make this wark, you have to disable "alignment fault emulation" in the kernel, its a command line option to procnto. (-ad or something)

-Al

post119740: Re: Block Device Level Encryption

Muthu Vengataraman — Fri, 07 Jun 2019 08:12:54 GMT

Hi

Thanks for your response.

As far as my understanding the "fsencrypt" will be used to encrypt the files and folders which are part of the QNX system.

The usecase that I am trying is

Encrypt a folder/directory (with subdirectories and files inside) in a host operating system running LINUX.

This encrypted folder/directory should be decrypted in the target operating system which will be running QNX.

I am able to implement the above usecase on a LINUX HOST and LINUX TARGET system using dm-crypt + LUKS (cryptsetup). but not on QNX TARGET since cryptsetup is not there.

Thanks

V.Muthusubramanian

post119739: Re: qtalk transfer file

Albrecht Uhlmann — Fri, 07 Jun 2019 08:10:58 GMT

qtalk comes as a prebuilt binary from QNX with the standard installation of Momentics. You just need it to include in your bootimage/filesystem image.

-Al

post119738: Re: Max disk size supported by QNX 6.5.0

Albrecht Uhlmann — Fri, 07 Jun 2019 08:09:03 GMT

See http://www.qnx.com/developers/docs/6.5.0SP1.update/#./com.qnx.doc.neutrino_user_guide/limits.html. There is a specific topic on filesystem limits.

HTH

-Al

post119737: Re: Block Device Level Encryption

Albrecht Uhlmann — Fri, 07 Jun 2019 08:07:19 GMT

QNX 7 offers the Merkle filesystem on top of the regular QNX filesystem for integrity, and the QNX filesystem does have an encryption module that can be used as an option.

http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.utilities/topic/f/fsf-merkle.so.html

http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.utilities/topic/f/fsencrypt.html

Regards,

Albrecht

post119736: Max disk size supported by QNX 6.5.0

Jagat Bhusan(deleted) — Fri, 07 Jun 2019 06:18:36 GMT

Hi,

What is the maximum disk (hard disk & USB) size supported by QNX 6.5.0 when formatted with QNX file systems such as QNX 4 & QNX6 (fs-qnx6.so) ?

Regards

Jagat

post119735: Block Device Level Encryption

Muthu Vengataraman — Fri, 07 Jun 2019 05:04:54 GMT

In Linux we could use the dm-crypt + LUKS (cryptsetup) to encrypt and decrypt disks,partitions, logical volumes.

Is there any similar or equivalent utility available in QNX 7.0.0.

Thanks

V.Muthusubramanian

post119731: qtalk transfer file

khiemspdt lq(deleted) — Thu, 06 Jun 2019 14:35:38 GMT

Hi all,

I need to transfer file from PC to target via UART, but it need qtalk on the target.

Anyone could share me qtalk source code ? then I try to build at my environment.

Thanks.

post119726: Re: TARFS Resource Manager support

Steve Reid(deleted) — Tue, 04 Jun 2019 21:34:23 GMT

I'll take a look at it, but I don't have an AArch64 machine to try it on.

post119725: Re: TARFS Resource Manager support

Muthu Vengataraman — Tue, 04 Jun 2019 14:13:20 GMT

I was able to cross compile the Code for AARCH64 on QNX7.

but when I ran the commands

tarfs &

mount -T tarfs /home/root/sample.tar sample.tar.dir

the mount command didn't return any error.

But when I try to cd to sample.tar.dir

it gave Bad descriptor error.

Thanks

V.Muthusubramanian

post119724: Re: TARFS Resource Manager support

Muthu Vengataraman — Tue, 04 Jun 2019 10:40:15 GMT

Hi Steve

Yes. I downloaded the source code from the link that you provided.

I tried cross-compiling it for AARCH64 architecture on QNX7 and I get link errors for gzopen,gzclose APIS.

Thanks

V.Muthusubramanian

post119723: Re: TARFS Resource Manager support

Steve Reid(deleted) — Mon, 03 Jun 2019 14:29:44 GMT

Do you mean the one that's discussed in the QNX Neutrino Cookbook? If so, you can get the source from the Download area on our website (it's in the same directory as the PDF versions of the SDP documentation). Here's a link:

http://www.qnx.com/download/feature.html?programid=30134

post119716: TARFS Resource Manager support

Muthu Vengataraman — Thu, 30 May 2019 11:46:44 GMT

Hi

How do I test the TARFS resource manager on QNX7.0 running on a AARCH64 CPU.

Where can I get the source code of the same.

Thanks

V.Muthusubramanian

post119713: Detecting unaligned access during runtime

Sesh — Fri, 24 May 2019 22:51:14 GMT

We are using QNX 6.6 with ARMv7 cores. Does QNX have any options to detect unaligned memory access at runtime? With linux this can be achieved by

echo 3 > /proc/cpu/alignment to warn and fixup errors.

post119703: qcc stopped working.

Kevin Phillips(deleted) — Fri, 10 May 2019 22:57:37 GMT

I am a newbie to QNX development (taking over an old project).  Something has changed, and I don't know what, and I am no longer able to compile anything.

I am using the QNX SDP 6.4.1 on a Windows 10 PC.  Every compile fails with the following message:

'-p' is not recognized as an internal or external command,

operable program or batch file.

*********************************************************

Your software must be activated, run the "qnxactivate" application to activate.

For more information or to get a new license key contact

QNX Software Systems at activate@qnx.com

*********************************************************

license check failed

Even if I run qcc by itself on the command line, with no options, I get the same error.  I do have a valid license:

C:\Users\myusername>showlicense

The currently active license is: Commercial

The text corresponding to this license can be found in:

C:\QNX641\install\qnxsdp\6.4.1\licenseCommercial.txt

Any ideas on how to toubleshoot this?

Thanks!

post119691: How to use Eigen in QNX 6.4.1

Chris Liang(deleted) — Thu, 02 May 2019 02:39:19 GMT

Hi all

Anyone knows how to include Eigen?

my target is QNX 6.4.1, PPC

I am using momentics for corss-compiling on windows

my program is simple

#include <Eigen/Dense>

using Eigen::MatrixXd;

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

{

    MatrixXd m(2,2);

    return 0;

}

when I compile it, it shows following errors:

Compiling my_math.cpp

qcc -V4.2.4,gcc_ntoppcbe_ecpp-ne  -D_NO_CONSOLE_IO_ -DNDEBUG -O2 -IC:\src_Robotics\Robotics/Releases/Include -IC:\eigen_include\Eigen\src\Core -D_EMBEDDED_APPLICATION_  -Wno-unknown-pragmas -c vo_math.cpp

vo_math.cpp:26:23: error: Eigen/Dense: No such file or directory

cc: C:/QNX640/host/win32/x86/usr/lib/gcc/powerpc-unknown-nto-qnx6.4.0/4.2.4/cc1plus caught signal 1

make: *** [vo_math.o] Error 1

Thanks in advance

post119684: About QNX 7 include path

martin fong(deleted) — Sun, 28 Apr 2019 06:54:29 GMT

Dear all:

I'm using QNX 7 on ubuntu 18.04 OS. Using include path qnx700/target/qnx7/usr/include/c++/5.4.0 to complie Google's Gtest, compile success. But using include path qnx700/target/qnx7/usr/include/c++/v1, encountered the follow errors:

./libgtest_main.a(gtest-all.o): In function `testing::internal::Random::Generate(unsigned int)':

googletest/make/../src/gtest.cc:348: undefined reference to `std::__1::basic_ostream<char, std::__1::char_traits<char> >::operator<<(unsigned int)'

googletest/make/../src/gtest.cc:349: undefined reference to `std::__1::basic_ostream<char, std::__1::char_traits<char> >::operator<<(unsigned int)'

What is the diffeerence between of 5.4.0 and v1?

What path should be included for C++ project?

post119653: showmem showing weird memory information

Chris Liang(deleted) — Thu, 11 Apr 2019 15:08:21 GMT

Hi,

I want to check the program memory usage, data/bss, heap

I am using showmem -P to get the process memory usage

In my program, I create a large static buffer:

static char testBuf[10240]; // 10k bytes buffer

memset(testBuf, 0, sizeof(testBuf)); // set to zero

open this program in momentics, it shows

.bss 0x00002800 testBuf

total .bss  0x000152e0

this is correct, testBuf is static variable and it should belong to .bss section

When I ran this program in qnx 6.4.1, using showmem

it gives

Total             Code         Data        Heap         Stack     Other

4382720    1540096      81920    2752512       8192          0

Now, I remove the testBuf. after compiled,

total .bss  0x00012ae0,   so the difference is 0x2800 which makes sense.

However, when I ran the new program again, check the showmem

it shows:

Total             Code         Data        Heap         Stack     Other

4374528    1540096      81920    2744320       8192          0

I am expecting Data section reduced, however, it is not and the heap section reduced.

Does it means .bss includes in the Heap section in showmem??

Thanks

post119623: Re: MMC util to read extcsd from emmc devices

Albrecht Uhlmann — Fri, 29 Mar 2019 12:16:03 GMT

The SD/MMC drivers usually come in source form in the various BSPs. Maybe you can look for some special devctl calls that would retrieve this information.

Regards,

Al

post119619: MMC util to read extcsd from emmc devices

janardhan reddy(deleted) — Thu, 28 Mar 2019 11:19:18 GMT

HI All,

I want read the extcsd register values to know about much info about emmc.

In linux we have MMC util to read the information like csd cid and extcsd and other information.

Is there any utility developed for QNX to read all the details about emmc device.

Can any one share the source code or utility to read it.

any help will be great for me...

Thanks in advance

Thanks & Regards

Janardhan Reddy

post119618: Re: O_DIRECT giving error while compiling the source code

janardhan reddy(deleted) — Thu, 28 Mar 2019 11:15:15 GMT

HI Albrecht Uhlmann,

Thanks for your help...

I am able to built code..

Thanks & Regards

Janardhan Reddy

post119600: Re: O_DIRECT giving error while compiling the source code

Albrecht Uhlmann — Fri, 22 Mar 2019 13:33:44 GMT

I think O_DIRECT is not supported under QNX. Since it is a performance optimization, you could simply do something like

#ifnef O_DIRECT

#define O_DIRECT 0

#endif

Regards,

Albrecht

post119599: Testdisk support for qnx660

janardhan reddy(deleted) — Thu, 21 Mar 2019 11:53:45 GMT

HI QNX,

I am trying to compile netbsd testdisk package which is part of qnx660/pkgsrc to run on my QNX660 platform.

But i am getting build errors and unable to compile successfully.

Is testdisk support for qnx??

can anyone share the source code path if it support on QNX.

Thanks in advance

Thanks & Regards

Janardhan Reddy

post119598: O_DIRECT giving error while compiling the source code

janardhan reddy(deleted) — Thu, 21 Mar 2019 11:50:02 GMT

HI QNX,

I am compiling one of the source code which having OPEN( ) function call having O_DIRECT to open the file.

But i am getting error while compiling that O_DIRECT is not declared.

Is there any alternate to use to open the file which does same functionality as O_DIRECT.

Thanks in advance

Thanks & Regards

Janardhan Reddy

post119587: Re: user managing and/or passwd problem !

Albrecht Uhlmann — Mon, 18 Mar 2019 10:00:21 GMT

The image file system is read-only, you can only place pre-prepared versions of /etc/group, /etc/passwd and /etc/shadow there, these files are not changable at runtime, so passwd utility fails.

For quick-tests, you can startup a RAMdrive and mount it. For persistent storage, you need a writable file system and mount it in appropriate place.

Regards,

Al

post119585: user managing and/or passwd problem !

Mh Mansouri(deleted) — Sat, 16 Mar 2019 07:33:53 GMT

Good day,

I have a problem similar to a post titled “Passwd Trouble - post56074”. I worked with a BSP that is based on “image file-system”. The original BSP did not include some account managing files in “/etc” as “group”, “shadow” and “.pwlock”. I have edited and/or added those missing files into the folder and compiled the BSP. Although “passwd new_user” could be invoked, I face with “cannot create password entry” message in each try. Moreover, despite of any “passwd” calling for changing the “root” password, the next “login” as “root” does not request the password as shown below:

# passwd

changing password for root

New password:

Retype new password:

# login

login: root

# passwd jack

User id # (100) 100

Group id # (100) 100

Real name () jackson

Home directory (/usr/jack) /ssd/jack

Login shell (/bin/sh)

New password:

Retype new password:

cannot create password entry

Regards,

post119565: Re: Error in Final Launch Sequence (Unknown Response Error Attempting to Exec Shared Lib)

Ian Clough(deleted) — Wed, 06 Mar 2019 19:16:15 GMT

Getting the same error, has anyone gotten anywhere with this?

post119558: Re: IOzone run on QNX

janardhan reddy(deleted) — Tue, 05 Mar 2019 07:52:57 GMT

HI Will,

I have resolved this fsync error by giving -l c option in make file.

Now i am able to run iozone successfully.

Thanks for your great support to execute iozone.

Thanks & Regards

Janardhan Reddy

post119557: Re: IOzone run on QNX

janardhan reddy(deleted) — Mon, 04 Mar 2019 11:32:32 GMT

Hi Will,

Thanks for your support.

Able to build the iozone and running on QNX660 arm processor.

After run iozone process is getting interrupted.

ERROR log:

# ./iozone -a

        Iozone: Performance Test of File I/O

                Version $Revision: 3.408 $

                Compiled for 32 bit mode.

                Build: qnx

        Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins

                     Al Slater, Scott Rhine, Mike Wisner, Ken Goss

                     Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,

                     Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,

                     Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy, Dave Boone,

                     Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root,

                     Fabrice Bacchella, Zhenghua Xue, Qin Li, Darren Sawyer.

                     Ben England.

        Run began: Wed Dec 31 20:56:56 1969

        Auto Mode

        Command line used: ./iozone -a

        Output is in Kbytes/sec

        Time Resolution = 0.001000 seconds.

        Processor cache size set to 1024 Kbytes.

        Processor cache line size set to 32 bytes.

        File stride size set to 17 * record size.

                                                            random  random    bkwd   record   stride

              KB  reclen   write rewrite    read    reread    read   write    read  rewrite     read   fwrite frewrite   fread  freread

              64       4fsync: Function not implemented

iozone: interrupted

exiting iozone

Can you help on this what is causing this error.

Thanks & Regards

Janardhan reddy

post119554: Re: IOzone run on QNX

Will Miles — Fri, 01 Mar 2019 15:00:13 GMT

Hi,

librt and libpthread aren't needed on QNX -- those functions are included in the base libc. You can just remove them from the linker lines in the makefile.

-Will

post119553: Re: IOzone run on QNX

janardhan reddy(deleted) — Fri, 01 Mar 2019 05:31:13 GMT

HI Will,

Thanks for your great help,

now i am able to compile the iozone but while linking getting the error.

error log:

Building iozone for qnx

qcc -c -O3 -D_FILE_OFFSET_BITS=64 -DHAVE_ANSIC_C -DASYNC_IO -DNO_MADVISE \

-DNAME='"qnx"' -DQNX_ARM \

-Dqnx -D_LARGEFILE64_SOURCE  iozone.c \

-o iozone_qnx-arm.o

qcc -c -O3 -D_FILE_OFFSET_BITS=64 -DHAVE_ANSIC_C -DASYNC_IO \

-Dqnx  libbif.c -o libif.o

qcc -c -O3 -Dqnx -D_FILE_OFFSET_BITS=64 -DHAVE_ANSIC_C -DASYNC_IO \

-D_LARGEFILE64_SOURCE  libasync.c  -o libasync.o

qcc -O3  iozone_qnx-arm.o libif.o libasync.o -lrt -lpthread -o iozone

/opt/qnx660/host/linux/x86/usr/bin/i486-pc-nto-qnx6.6.0-ld: cannot find -lrt

/opt/qnx660/host/linux/x86/usr/bin/i486-pc-nto-qnx6.6.0-ld: cannot find -lpthread

cc: /opt/qnx660/host/linux/x86/usr/bin/i486-pc-nto-qnx6.6.0-ld error 1

make: *** [qnx] Error 1

Thanks in advance

thanks & Regards

Janardhan Reddy

post119549: Re: IOzone run on QNX

Will Miles — Wed, 27 Feb 2019 21:51:04 GMT

Hi,

I picked up the iozone sources from iozone.org. From there, I basically did:

-> makefile: added a qnx target. I used -D_FILE_OFFSET_BITS=64 -D_HAVE_ANCIC_C -DASYNC_IO -DNO_MADVISE

-> iozone.c: Add some defined(__QNX__) to get the right headers and signal handling bits; also renamed long long delay to tp_delay to avoid linker issues with delay()

-> libasync.c: fix includes on QNX, we need <stdint.h> and <errno.h> instead of <sys/errno.h>

-> libif.c: fix header includes with some defined(__QNX__)

NB that I'm also building with qcc in the usual way, eg.

CC=qcc make qnx

Hope this helps some,

-Will

post119541: Re: QNX-IFS is not loading on IMX7-Sabare board

janardhan reddy(deleted) — Tue, 26 Feb 2019 05:34:19 GMT

HI,

We are able to load the QNX-IFS and platform is booting fine.

Thanks & Regards

Janardhan Reddy

post119540: Re: IOzone run on QNX

janardhan reddy(deleted) — Mon, 25 Feb 2019 11:17:29 GMT

Hi Will,

Thanks for your response.

I am looking into the latest iozone source code from netbsd.

I added some places in iozone.c with (__QNX__), while building facing lotof build errors.

As you mentioned the post because of licensing you are unable to share the source code,can you please suggest the files which we need to modify/tweek to port on QNX.

it will be a great help to me

Thanks & Regards

Janardhan reddy

post119534: Re: IOzone run on QNX

Will Miles — Wed, 20 Feb 2019 15:05:35 GMT

Hi,

I can't resist a challenge, so I took a quick look.   Are you sure you were looking at the right program?  iozone itself seems to depend on nothing but libc (and on QNX libsocket).   The tarball includes some scripts and pieces to process the output with gnuplot, but you don't need them to run the benchmark; and if you do want those graphs, you can run the processing on some other machine after collecting the data.

I found that porting iozone was pretty straightforward - basically just copying an existing OS profile with the desired features, adding || defined(__QNX__) in a few places, and the usual header idiosyncrasy cleanup.   Unfortunately, the license for iozone prohibits publication of derived works, so I cannot post my sources - sorry.

I would also note that iozone benchmarks using standard clock interfaces, which on QNX 6.x are quite coarse (typical resolution ~1 ms, although if you increase the clock interrupt rate via ClockPeriod() you could probably push it down to 100us with only a moderate overhead).   This limits the accuracy of the measurements somewhat, particularly for small transactions which may complete faster than 1 tick.   Adjusting to use ClockCycles() or the like for more accurate wall-time measurements would be a more complex task.  (We, er, have an LD_PRELOAD library we use to intercept the underlying libc time calls and augment them with CPU cycle counter data, but again for licensing reasons I can't share this code, sorry.)   I don't know if this has changed in QNX 7.

-Will

post119532: Re: IOzone run on QNX

Albrecht Uhlmann — Wed, 20 Feb 2019 08:38:48 GMT

We tried to port IOZone to QNX some time back but appeared to us not a straight-forward port, I think there were lots of library depencies that would need to be satisfied.

So if an internet search doesn't yield a result, I guess there is no maintained QNX port available. Maybe inside QNX for disk testing.

Regards,

Albrecht

post119531: IOzone run on QNX

janardhan reddy(deleted) — Wed, 20 Feb 2019 05:36:42 GMT

HI QNX,

I would like to run IOZONE on QNX to get performance results.

Can i get the iozone source code/ link to source code to run on QNX6.6.0

Thanks in advance

Thanks & Regards

Janardhan Reddy

post119498: Re: question for procnto-booke and procnto-booke-instr kernels

Chris Liang(deleted) — Fri, 08 Feb 2019 16:43:00 GMT

> In that case we would need more information.

>

> - can you post the build script that you use to create the boot image?

> - are you sure all settings/env vars etc. are the same, only procnto replaced

> with instrumented version?

> - How are you launching you processes?

> - How do they terminate? (external signal, from their own)?

>

> It might be possible that the instrumented procnto behaves differently

> regarding process termination, maybe someone from QNX can comment on this

> after reviewing your data.

>

> Regards,

> -Al

To be honest, I dont think the env variable or settings are same, cuz i just tried to load with procnto. it still fails some of my processes.

Thanks

post119496: Re: question for procnto-booke and procnto-booke-instr kernels

Albrecht Uhlmann — Fri, 08 Feb 2019 15:43:51 GMT

In that case we would need more information.

- can you post the build script that you use to create the boot image?

- are you sure all settings/env vars etc. are the same, only procnto replaced with instrumented version?

- How are you launching you processes?

- How do they terminate? (external signal, from their own)?

It might be possible that the instrumented procnto behaves differently regarding process termination, maybe someone from QNX can comment on this after reviewing your data.

Regards,

-Al

post119495: Re: question for procnto-booke and procnto-booke-instr kernels

Chris Liang(deleted) — Fri, 08 Feb 2019 15:15:07 GMT

> To my understanding, no modifications are necessary. Just replace boot image

> and reboot target.

>

> There could be restrictions the other way round: If you place calls to

> TRACE_EVENT in your code these functions will fail if executed on a non-

> instrumented kernel.

>

> -Al

Hi Albrecht,

Thanks replying.

Then, it is weird. After I replace with instr kernel image, some of my processes become zombies. any idea?

post119493: Re: question for procnto-booke and procnto-booke-instr kernels

Albrecht Uhlmann — Fri, 08 Feb 2019 12:33:04 GMT

To my understanding, no modifications are necessary. Just replace boot image and reboot target.

There could be restrictions the other way round: If you place calls to TRACE_EVENT in your code these functions will fail if executed on a non-instrumented kernel.

-Al

post119492: question for procnto-booke and procnto-booke-instr kernels

Chris Liang(deleted) — Thu, 07 Feb 2019 23:39:49 GMT

Hi all,

a quick question.

If I wrote program running on procnto-booke kernel, does the program also be able to run on procnto-booke-instr kernel if I changed BSP image from procnto-booke to procnto-booke-instr?

Thanks

post119467: GCOV

Arun Vijay Selvaraj(deleted) — Mon, 04 Feb 2019 11:14:27 GMT

Hello QNX,

I am in need of your help regarding simplistic ''HELLO WORLD'' program using GCOV with the compile-option “std=gnu++98” and are able to compile and link code with QNX660 qcc. In code::blocks I try to compile the target 'debgcov_qnx660_qcc_x86' but, I am getting an error as undefined reference to `__gcov_init' and undefined reference to `__gcov_merge add'. So, I kindly need your help to short out these errors using Gcov in code::blocks.

Thank you in advance and looking forward to your positive response

Arun Vijay

post119463: Re: problem with fslSATA detecting ssd mSATA

hamed davaneh(deleted) — Wed, 30 Jan 2019 11:04:22 GMT

Hello Mr. Varlet,

I've tried your binary file but it gave me exact same result as before. I compared two binary files, apart from some info related to host machine and build date they are all the same.

Thank you for you response I'll try to contact MEN Germany.

post119462: Re: problem with fslSATA detecting ssd mSATA

Guillaume Varlet — Wed, 30 Jan 2019 09:01:31 GMT

attachement

post119461: Re: problem with fslSATA detecting ssd mSATA

Guillaume Varlet — Wed, 30 Jan 2019 09:00:55 GMT

Hello M. Davaneh.

Please try the attached driver.

If you have still problem, please contact MEN Support in Germany (support at men.de)

Guillaume Varlet,

Customer Support / SW developer

MEN Mikro Elektronik SAS

18, rue René Cassin

ZA de la Châtelaine

74240 Gaillard, France

www.men-france.fr

post119460: Re: problem with fslSATA detecting ssd mSATA

Guillaume Varlet — Wed, 30 Jan 2019 08:49:59 GMT

test

post119452: Re: QNX-IFS is not loading on IMX7-Sabare board

janardhan reddy(deleted) — Tue, 29 Jan 2019 03:32:30 GMT

> Hello QNX,

>

> I have downloaded MCIMX7 SABARE target BSP from below link

> Bsp link: http://community.qnx.com/sf/frs/do/viewRelease/projects.bsp/frs.

> sdp_6_6_bsp_for_the_nxp_mcimx7_s.sdp_6_6_bsp_for_the_nxp_mcimx7_s

>

> Built the bsp using QNX660 and created IPL(ipl-imx-sabre.bin) and QNX-IFS by

> following the user guide.

> Flashed the images on SD card and kept the card on target. after power on IPL

> is loading successfully and asked to select the load image from SD card.

> Given option to load from SD card, Image load is successful but image scan is

> getting failed always.

>

> How can i check what is causing to fail image scan.

>

> Could you please help me on this.

>

> Thanks in advance

>

> Thanks & Regards

> Janardhan Reddy

post119451: QNX-IFS is not loading on IMX7-Sabare board

janardhan reddy(deleted) — Tue, 29 Jan 2019 03:26:19 GMT

Hello QNX,

I have downloaded MCIMX7 SABARE target BSP from below link

Bsp link: http://community.qnx.com/sf/frs/do/viewRelease/projects.bsp/frs.sdp_6_6_bsp_for_the_nxp_mcimx7_s.sdp_6_6_bsp_for_the_nxp_mcimx7_s

Built the bsp using QNX660 and created IPL(ipl-imx-sabre.bin) and QNX-IFS by following the user guide.

Flashed the images on SD card and kept the card on target. after power on IPL is loading successfully and asked to select the load image from SD card.

Given option to load from SD card, Image load is successful but image scan is getting failed always.

How can i check what is causing to fail image scan.

Could you please help me on this.

Thanks in advance

Thanks & Regards

Janardhan Reddy

post119450: Re: qnx 6.6 appsmedia on beaglebone black

mike scott(deleted) — Mon, 28 Jan 2019 05:47:30 GMT

following 《QNX SDK for Apps and Media 1.1 Getting Started》， I am able to building a target image of beaglebonedemo.

Thanks for well documented.

Mike

post119440: Re: problem with fslSATA detecting ssd mSATA

Albrecht Uhlmann — Wed, 23 Jan 2019 14:18:24 GMT

We developed our own S-ATA driver for the P1013 long time ago. I remember we had to add some settings in startup: Add dynamic interrupts for the S-ATA controller. Maybe that's the reason.

Regards,

Al

post119433: problem with fslSATA detecting ssd mSATA

hamed davaneh(deleted) — Tue, 22 Jan 2019 13:15:39 GMT

I'm using QNX 6.5 on an NXP single-core P1013. It has an mSATA slot. in U-Boot it successfully detects the SATA:

A21=> sata device SATA0

SATA device 0: Model: G-BONG G300 32G Firm: 20170223 Ser#: AA000000000003083320

            Type: Hard Disk

            Capacity: 30208.0 MB = 29.5 GB (61865984 x 512)

... is now current device

A21=> sata part SATA0

Partition Map for SATA device 0  --   Partition Type: DOS

Partition     Start Sector     Num Sectors     Type

    1                 2048        61859840       6

But in the OS with fslSATA after about one minute I get error. I'm using fslSATA as noted in release note for P1010-RDB BSP :

devb-fslSata fslSata ioport=0xFF718000,irq=65

Error:

xpt_configure: fslSata SIM attach failure

post119418: qnx 6.6 appsmedia on beaglebone black

mike scott(deleted) — Thu, 17 Jan 2019 06:01:43 GMT

Hi QNX,

       qnx sdp 6.6, qnx sdk for apps and media 1.1 release,

       BSP_ti-am335x-beaglebone_br-660_be-660_SVN797070_JBN574.zip

       beaglebonedemo.ext.anm11.258.150325_2017

       About_the_QNX_SDK_for_Apps_and_Media.pdf

I am working on beaglebone black with qnx 6.6 + sdk for apps and media.

According to 《About_the_QNX_SDK_for_Apps_and_Media.pdf》 page 11, quote,

"You begin by downloading a BSP from the QNX website. Among other things, the BSP should include three buildfiles:

• a QNX SDP 6.6 BSP-specific buildfile

• a board-specific buildfile for Apps and Media (appsmedia-sample-board.build)

• a common buildfile for Apps and Media (appsmedia-sample-common.build)"

However, there is only one build file in the BSP_xxxx.zip , i.e., beaglebone.build, while missing appsmedia-sample-beaglebone.build and appsmedia-sample-common.build.

On the other hands, I have tried beaglebonedemo.ext.anm11.258.150325_2017, which actually provides full support of appsmedia package.

I am wondering whether QNX experts could provide appsmedia-sample-beaglebone.build and appsmedia-sample-common.build, or provide the building script for beaglebonedemo to facilitate my study.

Please advise, thanks in advance.

Mike

post119412: Re: Error in Final Launch Sequence (Unknown Response Error Attempting to Exec Shared Lib)

Hari Prasad(deleted) — Tue, 15 Jan 2019 00:34:11 GMT

@Amit: Did you figure out the solution? If so, please let me know the solution.

I am experiencing the same issue.

post119411: Re: DIM2 (MediaLB), i.MX6, clocks, porting Linux function

Albrecht Uhlmann — Thu, 10 Jan 2019 16:49:12 GMT

Hi Andrey,

I don't have this code at hand, haven't used MediaLB explicitly myself. I just explained the general procedure.

Regards,

Al

post119410: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Thu, 10 Jan 2019 15:53:50 GMT

> Hi Chris,

> the only reason I can think of is, when your application is writing binary

> data into the slogf().

> The API for slog maintains a flag, telling the server (slogger) that the data

> is readably text or just binary data.

> So when your Apps are writing binary data using slogf(), this garbage will be

> in the servers text buffers.

>

> In general, when sending binary (raw) data, you should use the slogb() API,

> when sending just readable text, use slogf(). A third API to send just

> integers as an array which later can be translated when the message is read

> and interpreted.

>

> Again, I think in your case, one of the formatted message contains no zero

> termination or garbage. I can't imagine, the slogger damages the messages on

> the server side.

>

> Do you have the source of the app, the garbage comes from?

Hi Peter,

Thank you again for all the comments and thoughts.

You are right, to send formatted messages, use slogf().

I am sending formatted ascii txt message in my case.

my code is:

sprintf(debugMsg, "%.3f, %.3f", distance, time);

slogf(_SLOG_SETCODE(_SLOGC_TEST, 0), _SLOG_DEBUG1, "<%s:%d> compute :%s", LOG_IDENT(id), __LINE__, debugMsg);

As you can see, this is formatted readable ASCII message. slogf is correct API to use. I am sending exactly same message repeat all over again. I saw the garbage messages after the program ran for 10 mins. the logging rate is in 200Hz.

I am thinking it might related to the heap memory. Because I ran slogger with 'slogger -s 256k -l /dev/shmem/log -f 7', -s 256k specify in-memory buffer size used which is allocated on heap. If I increase to -s 15000k, I do not see the garbage at all.

Thanks.

post119409: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Thu, 10 Jan 2019 15:51:55 GMT

Hi Peter,

Thank you again for all the comments and thoughts.

You are right, to send formatted messages, use slogf().

I am sending formatted ascii txt message in my case.

my code is:

sprintf(debugMsg, "%.3f, %.3f", distance, time);

slogf(_SLOG_SETCODE(_SLOGC_TEST, 0), _SLOG_DEBUG1, "<%s:%d> compute :%s", LOG_IDENT(id), __LINE__, debugMsg);

As you can see, this is formatted readable ASCII message. slogf is correct API to use. I am sending exactly same message repeat all over again. I saw the garbage messages after the program ran for 10 mins. the logging rate is in 200Hz.

I am thinking it might related to the heap memory. Because I ran slogger with 'slogger -s 256k -l /dev/shmem/log -f 7', -s 256k specify in-memory buffer size used which is allocated on heap. If I increase to -s 15000k, I do not see the garbage at all.

Thanks.

> Hi Chris,

> the only reason I can think of is, when your application is writing binary

> data into the slogf().

> The API for slog maintains a flag, telling the server (slogger) that the data

> is readably text or just binary data.

> So when your Apps are writing binary data using slogf(), this garbage will be

> in the servers text buffers.

>

> In general, when sending binary (raw) data, you should use the slogb() API,

> when sending just readable text, use slogf(). A third API to send just

> integers as an array which later can be translated when the message is read

> and interpreted.

>

> Again, I think in your case, one of the formatted message contains no zero

> termination or garbage. I can't imagine, the slogger damages the messages on

> the server side.

>

> Do you have the source of the app, the garbage comes from?

post119408: Re: DIM2 (MediaLB), i.MX6, clocks, porting Linux function

Andrey Shvetsov(deleted) — Thu, 10 Jan 2019 15:35:15 GMT

Hi Albrecht,

I appreciate your quick respond.

Could you please share the code that just activates the clocks mlb150_clk and pll6 on the i.MX6Q?

Thanks,

Andrey

post119407: Re: DIM2 (MediaLB), i.MX6, clocks, porting Linux function

Albrecht Uhlmann — Thu, 10 Jan 2019 15:20:25 GMT

Hi Andrey,

regarding the mx6x.h header file, I think that this is continuous work: As soon as a new driver or component is added that accesses some registers not yet defined, it will be added and part of the next BSP release.

What we do in our own i.MX6 BSP is to hold and maintain a private version of mx6x.h where we can add definitions as needed. We keep our section at the very bottom of the file so that a merge with a newer version from QNX is still possible.

Regards,

Albrecht

post119406: DIM2 (MediaLB), i.MX6, clocks, porting Linux function

Andrey Shvetsov(deleted) — Thu, 10 Jan 2019 14:54:11 GMT

I'm porting the Linux DIM2 (IP for the MediaLB) driver to the QNX.

The BSP is BSP_freescale_imx6x-sabreARD_br-660_be-660_SVN737103_JBN5.zip

The hardware is the i.MX6Q SABRE for Automotive Infotainment.

I can access the DIM2 IP using its register wall and it responses according to

the DIM2 manuals.

What makes me unsure, if I look into the BSP, is missing functionality around the

PLL6_MLB in the QNX BSP. For example the definition of MX6X_ANATOP_PLL6_MLB in

mx6.h has no usage at all. See how it is done for the MX6X_ANATOP_PLL8_ENET.

I'm looking for the ideas how to initialize the MediaLB clocks in the QNX like

it is done in the Linux function fsl_mx6_enable() from the link

https://elixir.bootlin.com/linux/v4.20/source/drivers/staging/most/dim2/dim2.c#L935

I also want to know what could be the equivalent to the Linux patch

https://github.com/microchip-ais/meta-mchp-most-fsl/blob/master/recipes-kernel/linux/linux-imx-3.14.28/0001-add-medialb-clk.patch

that describes the clocks for the MediaLB.

Thanks,

Andrey.

post119405: Re: slogger writes non-sense messages to the log file

Peter Weber — Thu, 10 Jan 2019 08:55:01 GMT

Hi Chris,

the only reason I can think of is, when your application is writing binary data into the slogf().

The API for slog maintains a flag, telling the server (slogger) that the data is readably text or just binary data.

So when your Apps are writing binary data using slogf(), this garbage will be in the servers text buffers.

In general, when sending binary (raw) data, you should use the slogb() API, when sending just readable text, use slogf(). A third API to send just integers as an array which later can be translated when the message is read and interpreted.

Again, I think in your case, one of the formatted message contains no zero termination or garbage. I can't imagine, the slogger damages the messages on the server side.

Do you have the source of the app, the garbage comes from?

post119404: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Wed, 09 Jan 2019 17:07:21 GMT

Peter,

Thanks for replying..

I am using the slogf API to log the data.

You recommend to use slogb?

slogf is for formatted message. I do format the message.

post119403: Re: slogger writes non-sense messages to the log file

Peter Weber — Wed, 09 Jan 2019 16:54:02 GMT

seems an app is writing binary data using the wrong API. e.g. slogf() instead of slogb() ?

post119402: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Wed, 09 Jan 2019 16:35:26 GMT

Yes,

I used 'sloginfo -t -w' which print messages once they arrived. the default log file is /dev/slog

I also saved to /var/log/message as 'slogger -s 2048k -l /var/logs/message -f 5' -l specify the saved raw log file.

then, 'sloginfo -t /var/log/message' to print out the saved log message.

They all contains some garbage logs.

> and have you tried using sloginfo to read the log file as suggested in the

> docs?

> http://www.qnx.com/developers/docs/6.6.0.update/#com.qnx.doc.neutrino.

> utilities/topic/s/sloginfo.html

post119401: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Wed, 09 Jan 2019 16:23:58 GMT

Sorry, it was a typo.

It is only one '-l'

slogger -s 2048k -l /var/logs/message -f 5

> what does the -l -l ...stands for?

post119400: Re: slogger writes non-sense messages to the log file

Peter Weber — Wed, 09 Jan 2019 16:22:28 GMT

and have you tried using sloginfo to read the log file as suggested in the docs?

http://www.qnx.com/developers/docs/6.6.0.update/#com.qnx.doc.neutrino.utilities/topic/s/sloginfo.html

post119399: Re: slogger writes non-sense messages to the log file

Peter Weber — Wed, 09 Jan 2019 16:15:34 GMT

what does the -l -l ...stands for?

post119398: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Wed, 09 Jan 2019 16:06:23 GMT

Please, any one knows?

> Hi,

> I am using slogger to write debugging message to a /var/log/message.

>

> execute the command 'slogger -s 2048k -l -l /var/logs/message -f 5' at the

> early system up stage.

>

> However, when I looked at the logged messages from sloginfo -t -w.

>

> it contains very non-sense messages which human cannot understand.

> ie.

>

> GOOD LOGS:

> Nov 30 23:07:24.501    5 10000     0 <compute.cpp:285> static lever arm v, 0.008, -0.133, 0.070

> , 0.000, 0.000, 0.000, 0.008, -0.133, 0.070

>

> BAD LOGS: the month is changed to Aug for some reason....

> Aug 23 15:49:06.663    3 93795   1862 1635149164 1701064291 1886403122

> 859061792 538976288 538979360 1936619103 1915642992 1529961768 825112113

> 875113504 1936619103 1915647579 828188720 774910008 690757740 1634956383

> 1985687901 674246192 808986924 543454258 674246192 808659244 541339944

> 824788018 674244896 1936619103 1915647016 909192752 926099756 544435830

> 1600416115 1949201448 909192752 926099756 544497513 1953900592 774910005

> 807993392 2826093 10000 1543619244 1013213555 1952542307 1700754546 1635149164

>  1701064291 1886403122 859061792 538976288 538979360 1936619103 1915642992

> 1530027304 758328886 909453612 544435830 1601318518 1530027304 758132273

> 858990892

>

> I thought it might be the sloginfo utility issue. Then, I wrote a stand along

> utility to decode slogger message by open /dev/slog

>

> Now the bad logs is not numbers but it is still not making any senses

>

> GOOD LOGS:

> Jan 06 20:43:13.528  5 10000   0: <compute.cpp:231> -77.586, 5.348, 45.341, -0.005, -0.000, 0.

> 003, 162802.148

>

> BAD LOGS:

> Sep 18 07:27:48.757  7 290669  1782: a.valid to false, (162806.585, 74597.004)

> , head from (2) advanced to (0)

> Feb 01 05:31:12.838  7 536617  514:

> Jan 01 02:46:40.758  5 133901   1: \2hf<compute.cpp:386> numOfMsgFromIin = 162759

> Sep 18 07:27:48.757  7 290669  1782: a.valid to false, (162807.585, 74598.004)

> , head from (0) advanced to (1)

>

> I think it might be related to slogger utility.

>

> Any one knows the issues?

>

> Thanks in advance...

post119394: Re: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Tue, 08 Jan 2019 15:08:41 GMT

Did I post in right place? If not, please guide me where I should post it

Many thanks

post119393: Re: Unable to install SDK for apps and media 1.1

KENTA OGATA(deleted) — Tue, 08 Jan 2019 09:06:56 GMT

> It's a known issue when installing AnM1.1 on Windows 7 64-bit system.

> There are 2 possible workarounds:

>

> 1.  open the console window, set up 660 environment by running qnx660-env.bat,

>  then launch AnM 1.1 installer from that terminal window

>

> If that does not solve the issue, you can try this workaround:

>

> 2. copy .com.zerog.registry.xml file from C:\Program Files (x86)\Zero G

> Registry\ to: C:\Program Files\Zero G Registry\.  Then launch AnM1.1 installer

> .

>

> Hope that helps.

>

> Natalia

>

Hello All,

I am also facing the same issue, unable to install the "App & Media Package (qnx-sdk-for-apps-and-media-1.1-201503251644

.exe)" while I already install the " SDP 6.6 (qnx-sdp-6.6-201402230339.exe)"

I tried the workaround for windows 7, but I could not install it.

My system / environment details are like Below:

OS      : Win 10 64 Bit

SDP    : 6.6 Version (Activated)

App & Media : Ver 1.1

--

Regards

post119389: slogger writes non-sense messages to the log file

Chris Liang(deleted) — Mon, 07 Jan 2019 15:55:16 GMT

Hi,

I am using slogger to write debugging message to a /var/log/message.

execute the command 'slogger -s 2048k -l -l /var/logs/message -f 5' at the early system up stage.

However, when I looked at the logged messages from sloginfo -t -w.

it contains very non-sense messages which human cannot understand.

ie.

GOOD LOGS:

Nov 30 23:07:24.501    5 10000     0 <compute.cpp:285> static lever arm v, 0.008, -0.133, 0.070, 0.000, 0.000, 0.000, 0.008, -0.133, 0.070

BAD LOGS: the month is changed to Aug for some reason....

Aug 23 15:49:06.663    3 93795   1862 1635149164 1701064291 1886403122 859061792 538976288 538979360 1936619103 1915642992 1529961768 825112113 875113504 1936619103 1915647579 828188720 774910008 690757740 1634956383 1985687901 674246192 808986924 543454258 674246192 808659244 541339944 824788018 674244896 1936619103 1915647016 909192752 926099756 544435830 1600416115 1949201448 909192752 926099756 544497513 1953900592 774910005 807993392 2826093 10000 1543619244 1013213555 1952542307 1700754546 1635149164 1701064291 1886403122 859061792 538976288 538979360 1936619103 1915642992 1530027304 758328886 909453612 544435830 1601318518 1530027304 758132273 858990892

I thought it might be the sloginfo utility issue. Then, I wrote a stand along utility to decode slogger message by open /dev/slog

Now the bad logs is not numbers but it is still not making any senses

GOOD LOGS:

Jan 06 20:43:13.528  5 10000   0: <compute.cpp:231> -77.586, 5.348, 45.341, -0.005, -0.000, 0.003, 162802.148

BAD LOGS:

Sep 18 07:27:48.757  7 290669  1782: a.valid to false, (162806.585, 74597.004), head from (2) advanced to (0)

Feb 01 05:31:12.838  7 536617  514:

Jan 01 02:46:40.758  5 133901   1: \2hf<compute.cpp:386> numOfMsgFromIin = 162759

Sep 18 07:27:48.757  7 290669  1782: a.valid to false, (162807.585, 74598.004), head from (0) advanced to (1)

I think it might be related to slogger utility.

Any one knows the issues?

Thanks in advance...

post119388: Re: porting bcm43455 wifi driver(bcmdhd) to imx6 + qnx6.6.0

Albrecht Uhlmann — Mon, 07 Jan 2019 11:46:02 GMT

The QNX network stack offers a shim layer to port NetBSD network drivers over to QNX quite easily. If that works for WiFi drivers I cannot tell. Using a Linux/Android source base as a starting point will definitely cause more porting effort.

The online help also contains a Network DDK with a sample Ethernet driver. For the additional things for WiFi (e.g. interface to WPA-Supplicant) I do not know about samples.

Which interface of the chip are you using? PCIe oder SDIO or UART? In case of SDIO you would also need to look up the supplied driver code for the devb-sdmmc driver and extract the SDIO part from it.

Regards,

Albrecht

post119378: porting bcm43455 wifi driver(bcmdhd) to imx6 + qnx6.6.0

wang tony(deleted) — Thu, 03 Jan 2019 01:41:28 GMT

Hi,

   I trying to bring wifi module bcm43455 (similar to bcm4335/bcm4339) on qnx6.6.0 platform.

   All i have is bcmdhd driver for linux/android, how can i port it to qnx system ?

  Thanks!

post119369: Re: QNX 6.5.0 - TFTP

Albrecht Uhlmann — Wed, 19 Dec 2018 11:25:10 GMT

Hi Hugo,

in a standard installation, /usr/sbin/tftpd is present by default. You need to enable the service in /etc/inetd.conf by commenting in the line "tftpd", and then (re)start "inetd".

/etc/bootptab and /etc/services may need to be adjusted, too.

Details can be found in utilities reference.

-Al

post119368: QNX 6.5.0 - TFTP

Hugo TOUZE(deleted) — Wed, 19 Dec 2018 10:23:37 GMT

Hello,

I am new to the QNX world.

I currently have a disk image (.iso) of the QNX 6.5.0 system. Unfortunately the TFTP service is not installed on this system. Do you know if it is possible to install this service via a package? If so, where can I download this package?

Best regards, Hugo

post119353: Re: syslog() function cannot log the message over 200 bytes???

Chris Liang(deleted) — Wed, 12 Dec 2018 16:14:24 GMT

I see.

Is syslog-ng available for qnx 6.4 or higher?

Thank you again.

> Our source for syslog() is from NetBSD, with some customization. I don't think

> we make the source available.

>

> I don't know about the limits on logging frequency, so let's hope someone else

> chimes in.

post119352: Re: syslog() function cannot log the message over 200 bytes???

Steve Reid(deleted) — Wed, 12 Dec 2018 15:41:45 GMT

Our source for syslog() is from NetBSD, with some customization. I don't think we make the source available.

I don't know about the limits on logging frequency, so let's hope someone else chimes in.

post119351: Re: syslog() function cannot log the message over 200 bytes???

Chris Liang(deleted) — Wed, 12 Dec 2018 15:28:16 GMT

Hi Steve,

Thanks for replying.

It is possible to change the size of the length. Where I can find the source code.

BTW, do you have an idea how fast the syslogd can handle writing logs. right now, I am writing debugging message in 200Hz rate. it seems it is always stop writing for 30 seconds and then start writing again.

Is there a limitation on the logging frequency?

Thanks

post119350: Re: syslog() function cannot log the message over 200 bytes???

Steve Reid(deleted) — Wed, 12 Dec 2018 14:51:09 GMT

I've looked at the source code, and it seems that 200 bytes is a hard-coded limit.

> Hi all,

>

> I am using syslog() function to log a debugging/info/warning message in /var/

> log/message.

>

> However, I realize that it chop the message if it is over 200 bytes. Is there

> a length of the message I can set?

>

> Thanks in advance..

post119349: How can I get third-party open source software "ceres"

Li Liu(deleted) — Wed, 12 Dec 2018 04:18:26 GMT

I need Ceres for Matrixs computation, but my qnx-adas-2.0 do not have this lib and include.

I also travelled all software in qnx software center, and get no result.

Really appreciate your help, Thank you!

post119348: syslog() function cannot log the message over 200 bytes???

Chris Liang(deleted) — Tue, 11 Dec 2018 20:48:12 GMT

Hi all,

I am using syslog() function to log a debugging/info/warning message in /var/log/message.

However, I realize that it chop the message if it is over 200 bytes. Is there a length of the message I can set?

Thanks in advance..

post119315: Re: Error in Final Launch Sequence (Unknown Response Error Attempting to Exec Shared Lib)

Amit Sharma(deleted) — Tue, 04 Dec 2018 07:33:53 GMT

Hey anyone sort out this issue . i have been looking for solution from past one week..

I have installed qnx sdp on my ubuntu and qnx neutrino rtos on vmware .

post119312: GPIO interrupts on Intel Atom E3800 (Baytrail)

Vasilij Yanikeev(deleted) — Wed, 28 Nov 2018 10:14:29 GMT

Hello!

I am trying to use level triggered interrupts on my Intel Atom E3800 based board. A GPIO pin, wich I want to use is GPIO_S5[13]. So This GPIO is located in Intel Legacy Block (iLB) and registers, wich correspond to this GPIO, are mapped to memory (not to legacy space). To configure this GPIO I made the follows:

...

pci_attach( 0 );

//Find the PCI/ISA bus (intel legacy block)

pci_find_class(0x060100, 0, &busnum, &devfuncnum);

//Get IOBASE register, which contains base address for memory mapped GPIOs registers (GPSSUS) accoding to datasheet

pci_read_config32(busnum, devfuncnum, 0x4C, 1, &ioBaseRegister);

uint32_t ioControllerBaseAddr = ioBaseRegister & 0xFFFFC000;

//Get pad configuration register for GPIO_S5[13]

uint64_t s5_13_pad_config =  (ioControllerBaseAddr + 0x2000 + 0x140);

uint32_t * ptr = (uint32_t*) mmap_device_memory( 0, 4, PROT_READ|PROT_WRITE|PROT_NOCACHE, 0, s5_13_pad_config );

// (1<<26) - Set  Gb Tne to 1 ( Detect falling edge. For level interrupt mode it will enable active low level irq)

// (1<<27) - Set Direct Irq En (direct_irq_en)

// (1<<24) - Set Gd Level (gd_level): When this bit is set a level irq will be choose and not edge irq

*ptr |= (1<<26)|(1<<27)|(1<<24);

//Get pad value register for GPIO_S5[13]

uint64_t s5_13_pad_val_unmaped =  (ioControllerBaseAddr + 0x2000 + 0x148);

ptr = (uint32_t*) mmap_device_memory( 0, 4, PROT_READ|PROT_WRITE|PROT_NOCACHE, 0, s5_13_pad_val_unmaped );

//Enable input, disable output

*ptr |= (1<<1);

*ptr &= ~(1<<2);

//Connect gpios5[13] to IRQ0 (APIC 67)

uint64_t direct_irq0 =  (ioControllerBaseAddr + 0x2000 + 0x980);

ptr = (uint32_t*) mmap_device_memory( 0, 4, PROT_READ|PROT_WRITE|PROT_NOCACHE, 0, direct_irq0 );

*ptr &= ~(0x7F);

*ptr |= 13;

pci_detach( phdl )

...

In my startup-apic -vv output a see:

...

[13] vec: 67,   cpumask:00000001, ID: 2, busid:0, busintr:13, dest intin:13, type:0, flags:00000005

...

So when I am waiting for interrupt I do the follows:

struct sigevent int_event;

static int int_counter;

const struct sigevent *handler (void *area, int id) {

     int_counter++;

     return (&int_event);

}

void * interruptWaitThread( void*  arg )

{

ThreadCtl( _NTO_TCTL_IO, 0 );

struct sigevent event;

int_event.sigev_notify = SIGEV_INTR;

InterruptAttach( 13, &handler, NULL, 0, _NTO_INTR_FLAGS_TRK_MSK);

while (1)

{

InterruptWait (0, NULL);

printf (" Got interrupt %lu ( 0x%lX ) #%4ld\n", 13, 13, int_counter);

}

return( 0 );

}

But the interruptWaitThread have never unblocked from InterruptWait(). What am doing wrong?

Regards,

Vasily.

post119311: GPIO interrupts on Intel Atom E3800 (Baytrail)

Vasilij Yanikeev(deleted) — Wed, 28 Nov 2018 10:14:27 GMT

post119310: GPIO interrupts on Intel Atom E3800 (Baytrail)

Vasilij Yanikeev(deleted) — Wed, 28 Nov 2018 10:14:25 GMT

post119148: is there USB 3.0 driver for QNX 6?

Cong Pham(deleted) — Mon, 24 Sep 2018 09:39:48 GMT

Hi,

Does QNX 6 support usb 3.0 stack ? And if it does, where can I get driver files ?

post119109: QNX 6.5 OpenSSL Build Error

Murugaiyan Perumal(deleted) — Sun, 16 Sep 2018 16:02:13 GMT

Hi,

Am trying to build the openssl source for QNX 6.5/6.6 OS. I have tried to build after the instructions given in internet.

1. QNX 6.6 build environment variable is set.

2. Executed below command.

sh-3.1$  ./Configure QNX6 shared --prefix=./qnx660/release --openssldir=./qnx660/release

3. make depend exectuted. Got compilation error.

../../util/domd ../.. -MD gcc -- -fPIC -DOPENSSL_PIC -DOPENSSL_THREADS  -DDSO_DLFCN -DHAVE_DLFCN_H  -I.. -I../.. -I../modes -I../asn1 -I../evp -I../../include  -DOPENSSL_NO_DEPRECATED -DOPENSSL_NO_EC_NISTP_64_GCC_128 -DOPENSSL_NO_GMP -DOPENSSL_NO_JPAKE -DOPENSSL_NO_LIBUNBOUND -DOPENSSL_NO_MD2 -DOPENSSL_NO_RC5 -DOPENSSL_NO_RFC3779 -DOPENSSL_NO_SCTP -DOPENSSL_NO_SSL_TRACE -DOPENSSL_NO_SSL2 -DOPENSSL_NO_STORE -DOPENSSL_NO_UNIT_TEST -DOPENSSL_NO_WEAK_SSL_CIPHERS --  dso_dl.c dso_dlfcn.c dso_err.c dso_lib.c dso_null.c dso_openssl.c dso_win32.c dso_vms.c dso_beos.c

dso_dlfcn.c:84:12: fatal error: dlfcn.h: No such file or directory

#  include <dlfcn.h>

Also when I executed make all from command line. Seems like nothing has been built.

sh-3.1$ make

sh-3.1$ make all

sh-3.1$

Could anyone help me out to build openssl for qnx 6.5 os.

Thanks,

Murugaiyan