[linrad] Re: Linrad-02.05

[Leif Asbrink <sm5bsz.com; leif@xxxxxxxxxxxxxxxx>]

Hi Roger,

> THis is just some 'user feedback' from 1.37 and 2.05 ;)
> 
> Windows:
> Linrad-02.05 is working fine so far.  I haven't made it fail under windows.
> 
> Linux:
> LInrad 01.37 seems to work fine here at home. I haven't put it thru its  
> paces at Hilltop yet.
> LInrad-02.05 works fine under svgalib.  I haven't identified any problems  
> yet.
OK. Fine:-)

There are still many 'errors' in the code. You can detect some of
them by pressing ESC. I have to add a lot of code to make it work
everywhere - it is imposed on me by MS Windows which requires each
thread to be properly terminated. With svgalib it was easy. Just 
call the error routine and exit from there. The other threads
were then killed properly by Linux. Under Windows such a strategy
causes memory leakage and is unacceptable.
Linrad uses the same code for svgalib, X11 and Windows so the
code has to do handle ESC the complicated way in all versions...


> ----------------below is all Linux with Xwindows version of Linrad using  
> Delta44 as input----------------
> Linrad-02.05 is still flaky under Xwindows.  I tried it with mlock all on  
> and mlockall off and it was flaky with both.
> With mlockall off it can take 5 tries of ./xlinrad to get it started.
OK. 

Xwindows crashes if the CPU is overloaded. Do you use mmx
routines? Another thing is memory usage. Make sure that
the baseband storage time is minimum and set 5 seconds each for
fft1 and fft2. This will eliminate problems due to lack of ram.
Once Linrad runs smoothly you can increase storage sizes
and find where the limits are.


> With mlockall in either setting, it has crashed when:
> 
> The mouse is moved across the screen [no click necessary]
When I move the mouse over the panel on my Debian Etch
system, the cpu load goes close to 100% (there is a utility
that shows a graph of total CPU load under system tools). 
Clicking is no problem because disk I/O is not CPU intensive.
Moving or resizing or just holding the button pressed can
be very costly depending on what program you have in 
parallel with Linrad.

> A parameter window was changed in size by pulling it with the mouse.

> A parameter [e.g. number of averages in waterfall, waterfall baseliine,  
> waterfall gain, high resolution waterfall number of averages, etc] is  
> changed.
> 
> The size of the whole linrad window is changed by pulling it with a mouse.
> 
> Also, it sometimes crashes after a time few seconds to 2.5 minutes with no  
> mouse use, but sometimes has run more than 10 minutes with no problem
All of this indicates CPU overload. These operations require the CPU 
to recompute data from memory to provide an immediate change
on screen and in the loudspeaker output with the new parameter
values. 


> I can't pull the S meter away from the bottom left corner
Try more times. It is difficult to find the right position
with the arrow mouse. I do not know yet how to change to
have a cross for the mouse.


> ---------------------------------thats all I found so far.  I compiled  
> with my users_hwaredriver.c file, which could make a difference I guess.
No, I do not think it matters. The hardware test mode works
fine on the parallel port so I do not think X11 has similar
restrictions as MS Windows.

> I listed below some of the error messages I got when Linrad 02.05 running  
> as xLinrad crashed:
> 
> debian:/home/linrad/linrad-new/linrad-02-X11.05# ./xlinrad
> X Error of failed request:  BadRequest (invalid request code or no such
> operation)
>    Major opcode of failed request:  171 ()
>    Minor opcode of failed request:  0
>    Serial number of failed request:  829
>    Current serial number in output stream:  832
>
> debian:/home/linrad/linrad-new/linrad-02-X11.05# ./xlinrad
> X Error of failed request:  BadRequest (invalid request code or no such
> operation)
>    Major opcode of failed request:  0 ()
>    Serial number of failed request:  43
>    Current serial number in output stream:  63
> 
> debian:/home/linrad/linrad-new/linrad-02-X11.05# ./xlinrad
> X Error of failed request:  BadLength (poly request too large or internal
> Xlib length error)
>    Major opcode of failed request:  72 (X_PutImage)
>    Serial number of failed request:  43
>    Current serial number in output stream:  63
> 
> debian:/home/linrad/linrad-new/linrad-02-X11.05# ./xlinrad
> X Error of failed request:  BadLength (poly request too large or internal
> Xlib length error)
>    Major opcode of failed request:  87 (X_AllocColorPlanes)
>    Serial number of failed request:  590
>    Current serial number in output stream:  608
I do not know enough about X11, but I guess it crashes
because xlinrad continues to send commands even when
X11 signals it is busy. I have no idea how to find
out. Guess there is a way of getting the 'serial numbers'
from the output stream and do something sensible within 
Linrad until X11 is ready for more data.

73


Leif  /  SM5BSZ

#############################################################
This message is sent to you because you are subscribed to
  the mailing list <linrad@xxxxxxxxxxxxxxxxxxxxx>.
To unsubscribe, E-mail to: <linrad-off@xxxxxxxxxxxxxxxxxxxxx>
To switch to the DIGEST mode, E-mail to <linrad-digest@xxxxxxxxxxxxxxxxxxxxx>
To switch to the INDEX mode, E-mail to <linrad-index@xxxxxxxxxxxxxxxxxxxxx>
Send administrative queries to  <linrad-request@xxxxxxxxxxxxxxxxxxxxx>


LINRADDARNIL