[linrad] W3SZ system configuration and console switching..answer to Kohjon's posted questions LONG

[W3SZ <w3sz@xxxxxxxxxxx>]

Hi, Kohjin!

I will answer some of your questions about what I have seen here and
what my system configuration is.  I am sending this to the list as I
hope it will help others, too.  I will not offer any explanations about
why things do or don't work here, as I don't have the answers.  I think
you are far ahead of me in your understanding of Linux.  I have learned
enough to get Linrad working, but no more than that.  I am learning a
lot from you and the other posters to the net.  And, of course, Leif has
given me graduate level education (*plus*) in many of these areas, but I
am not smart enough to retain all of it for long ;(

There seem to be a LOT of differences in what problems occur on
different systems, so I will list first some of my system
characteristics:

Dell Pentium 4 1.4 GHz 

I have used several versions of RHL.  I am currently running 7.2,
release 2.4.7-10.  I have kept current with updates WHERE I think I need
them.  I have NOT installed all possible updates, as many of the files
are not necessary and will just bloat the system.  I have kept gcc and
the other important files updated consistently.

I have 256 MB RAM.

Video Card is ATI Radeon (generic driver).  Linux identifies this as
"Radeon QD".  It is an AGP card.

Sound cards are the Delta44 and a SoundBlaster (Ensoniq) PCI64 that OSS
identified as 'Ensoniq AudioPCI97' using the es1371 driver.

I am using OSS for sound system.

I have gcc version 2.96.

I generally run Linrad in Xwindows, but it runs fine in command mode,
too.

Now to your questions:



> First, remember that svgalib uses 'allocated virtual console 
> #8' on Linrad started from Xwindows while it uses the same 
> console on which Linrad started, not Xwindows but text base 
> (shell) console. Also remember that you can reach 'allocated 
> virtual console #8' by CNTL-ALT-F8. It means you can switch 
> between Linrad screen and Xwindows console by CNTL-ALT-F8 
> and/or CNTL-ALT-F7. Since I have been using Linrad from only 
> Xwindows I didn't recognize above difference until our recent 
> 'console switching' discussion.

This is very interesting.  This has not worked for me in the past.  When
I tried it I would get ^[[19~ appearing in the xterm window from which I
started Linrad.  But with Kohjin's comments I tried moving the mouse
cursor outside of that xterm window to the background area of the
xwindows screen.  When I do that and type cntl-alt-F8 it does indeed get
me back to Linrad with NO problem.  My screen is not corrupted, and I
have NEVER to my recollection had a problem with the screen being
corrupted after console switching.  The only thing I get when console
switching is an "overrun error" message in the bottom right of the main
spectrum display, which to me is expected since Linrad processing is
stopped (no audio) while I am in another console.  

So now that I can get back to Linrad from Xwindows, by placing the mouse
cursor outside the xterm display window before typing cntl-alt-F8, I
have no problems with console switching.  I am using version 048 of
Linrad.

> 
> I have never met ESC sequence problem so far to exit from 
> Linrad at all! I have never used a CTRL-c or whatever before. 
> A problem for me has been a damaged corrupted display at the 
> return from other window, I believe this is the case every 
> one has. The sound is alright but the display completely 
> corrupted, so I would hit 'X' key and 'B' to go back to the 
> previous mode. Now, this clears up the display but clears 
> AWAY a cursor which was ever focused on a certain frequency 
> too. (too bad if the station was rare expedition etc)

This was a problem here from version 041 to version 047, and was
eliminated when Leif added the statement:  sleep(0.02); line 1069 in
wcw.c     which took care of some memory problems.

It is not a problem here with the addition of this statement, and I have
confirmed that removing this statement causes the 'lockup on escape'
problem to return.  Case solved.

> 
> case 00-47 (Linrad launched from text base console):
> On returning from other console it happens sometime (not 
> every time) a hit of 'X' doesn't bring main menu then another 
> second hit of 'X' required. In this case, the screen is 
> heavily corrupted.
> 
> case 00-48 (Linrad launched from text base console):
> Same as above.

I have never seen this problem here.

> 
> case 00-47 (Linrad launched from Xwindows):
> Launched and switched to other text base windows for a while 
> then crashed with the following error. INTERNAL ERROR: 1109 (errno=62)
> routine:open_audio_input   file:setad.c
> call to ioctl failed. (SNDCTL_DSP_GETISPACE)

I have never seen this here.  But I did not go back to 047 to try and
duplicate.

> 
> It looks above phenomena depends on the period after departed 
> from the console of Linrad launched.

> 
> case 00-48 (Linrad launched from Xwindows):
> No crash/error for a long time in other windows and jobs.
> Still garbage display but can clean up by 'X' and 'B'.

Again, I have never seen this problem.

> Like I wrote above I have never met any fatal system crash by 
> console switching before. It looks (for me at least) 00-47 
> and 00-48 both are about same. This was confirmed running 
> both version repeatedly.

I had crashes on exit via escape from Linrad when run from X windows
ONLY, until sleep statement was added.

> 
> At 23:43 02/10/10 -0400, Roger W3SZ wrote:
> >Well, I generally run Linrad in Xwindows.  I can alt-F1, 
> alt-F2 etc and 
> >log in to those other command-mode consoles, and when I come back to 
> >Linrad all seems OK, but I do get an overrun error message 
> in the lower 
> >right hand corner of the linrad spectrum display when I do that.  No 
> >problems though.
> 
> Unbelievable, are you sure that the screen display not 
> corrupted and only an overrun error message? Please make sure 
> that you come back from 'allocated virtual console #8' to the 
> console you started Linrad in Xwindows.

Yes.  I have never seen the problem with corrupted display, with any
version of Linrad that I can recall.  Although my memory going back two
years is I am sure incomplete.  But I am sure it has not happened in any
recent versions, at least since 2001 EME contest.


> 
> >But if I alt-F7 while linrad is running I go back to the original x 
> >window from which I started linrad,
> 
> This meant you have launched Linrad from a shell in Xwindows 
> and watching the Linrad display in 'allocated virtual console 
> #8' and then you go back to the original Xwindows shell.

Yes.

>  
> >and I am then locked out of linrad
> >and must cntl-c to stop linrad and then restart linrad to 
> get back in 
> >to it.

But now due to Kohjin's question I have solved this.  I move the mouse
cursor out of the xterm window before typing cntl-alt-F8 and things work
fine.


> 
> Hmm....I was confusing but now I think I understand!
> As far as when you are out of the 'allocated virtual console 
> #8' CNTL-c is the only way to exit Linrad in it's launched 
> console. Why don't you use ESC key to exit while you are in 
> 'allocated virtual console #8' ?

I am not sure we understand each other here.  Esc will allow me to exit
Linrad while it is running.  Before the sleep statement was added, it
would if Linrad had been running for a long time cause Linux to lock up
so that ONLY a reboot would fix things.  This happened only on versions
041 to 047, not version 040 and before.  And it is fixed in version 048.
And it ONLY happened when Linrad was started from X windows, not when it
started from a command line console.  It was likely due to memory used
by Xwindows being corrupted via Linrad, I think.

The issue of switching between consoles is a different issue.


> 
> >So, once I have gone back to an x window console I cannot 
> get back to 
> >linrad.
> 
> You can use CNTL-ALT-F8 to go back again to Linrad?
> This is the only way from Xwindows console to Linrad screen. 
> Remember this is not the case Linrad launched from text base console.

See above.  Here at W3SZ, mouse must be positioned outside of xterm
window for this to work.

> 
> >I can also start linrad from a command mode console and 
> switch to other 
> >command mode consoles and to the xwindows console.  In this 
> case I CAN 
> >get back to linrad from both the command mode consoles and 
> the xwindows 
> >console and all seems to work fine.
> 
> Again, are there no corrupted display after switching?
> In other words, Linrad runs same as before (except overrun 
> error)? If so, we should take into account of video card...???!!!

NEVER any corrupted display in Linrad.  So the video card or its driver
or some other difference between our setups is the key here.

> 
> >I use ATI Radeon AGP video card and have run with RHL 6.x, 7.0, 7.2, 
> >and 7.3.
> 
> Different video card might produce different result.
> Mine is Matrox Millennium G400 32MB.

OK.

> 
> By the way Roger, have you followed up to the RHN update?
> There are huge necessary update for 7.3!
> All you have to do is to register and launch Update Agent which will
> do all for you.   It will take a while though.
> This update includes gcc-2.96-112 and library too.

I have just been running 7.2.  I update selectively.  Otherwise I think
the operating system gets very bloated and unnecessary problems may
occur.  Others with more experience than I may disagree.  I have gcc
2.96 (last update today), but I am not sure if it is 112.  gcc --version
just gives me 2.96 here.

I wonder if the fact that I've not got RHL 7.3 on this machine, but just
7.2 is why I am not having problems?  I thought I had put 7.3 on it, but
it is clearly running 7.2.

> 
> I have strange problems in RedHat 8.0 (gcc-3.2).
> Changing another RHL8.0 installed 10GB disk I will report 
> about it shortly.

Thank you VERY much for this information Kohjin, it has saved us a LOT
of frustration I am sure!

I hope the above is helpful.  I answered each of your points even when
my answers were redundant, because I think this makes things less
confusing.

By the way, I had Linrad running but was switched to the main Xwindows
console the entire time I was writing this email.  The screen saver even
kicked in during period of inactivity on the screen.  When I got ready
to send this I did cntl-alt-F8 and got right back to linrad with NO
problem and NO screen corruption.

73,

Roger Rehr
W3SZ   FN20ah
2 Merrymount Road
Reading, PA  19609-1718
http://www.qsl.net/w3sz

LINRADDARNIL