home bbs files messages ]

Just a sample of the Echomail archive

Cooperative anarchy at its finest, still active today. Darkrealms is the Zone 1 Hub.

   OS2      Fidonet International OS/2 Conference      3,371 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 594 of 3,371   
   Lars Erdmann to All   
   Re: Fixpak file mapping - how to get a g   
   24 Nov 11 17:33:22   
   
   hCOkT^4e>H0AmKk^CT1A^BL@[n6iTbKebAoFi[P8\f[>hm6m7kPS2jZX>Tm   
   mp.os.os2.misc:3721   
   From: "Lars Erdmann"    
      
   The .TDF and .TFF always go into directory \os2\system\trace.   
   On the other hand and as you have observed they will for some reason or the    
   other   
   also show up in other directories (as duplicates I suppose).   
      
   The .TDF, .TFF are a black hole. You will need to make sure they match   
   the corresponding DLLs (or OS2KRNL) for example. The .TDF files specify at    
   what machine instruction   
   to set a dynamic tracepoint and what register values/memory location to add    
   to the tracing buffer   
   once the dynamic tracepoint is hit.   
   It should be obvious to you that there needs to be an EXACT match   
   between .TDF and corresponding DLL(s) / OS2KRNL for this to work.   
      
   For one reason or the other, an FP or CP install will not necessarily    
   install the .TDF,.TFF files when it updates   
   the DLLs. You might want to consider finding these files in the CP,unpacking    
   them by   
   hand and moving them to the \os2\system\trace directory. You might need to    
   distinguish between W4, SMP and UNI   
   kernel for those .TDFs that set dynamic tracepoints in OS2KRNL.   
   Even worse, some updates like for example (I think) JFS might come with    
   updated .TDF files so you will   
   have to make sure you also catch these.   
      
      
   What is CP2 with FP6 ? And is there no newer CP for you ?   
   You should consider upgrading to eCS. That will give you CP4 and CP5.   
   There even is a CP6 but it is unofficial, you would be on your own ...   
      
      
   Lars   
      
      
      
   "Dariusz Piatkowski"  schrieb im Newsbeitrag    
   news:Qg5I6Bo2seGy-pn2-dfemiKMkQyD6@neurobox...   
   > Hi Folks!   
   >   
   > Ever since my upgrade to the '21st century computing' (lol - a move from    
   > an AMD   
   > Barton to Phenom X4 setup) I have been experiencing an odd freeze here and   
   > there. These are not hard locks, rather the WPS stops responding, mouse    
   > cursor   
   > is still OK, CPU monitor (v 1.17) chugs along, but the WPS interface seems    
   > to   
   > stop working. This is always brought on by something a given application    
   > has   
   > attempted to do, once such app is AVXCAT, which I would love to use. The    
   > 2nd on   
   > is the latest PMView install program (WPI based). Both will soft-lock the    
   > WPS,   
   > only CAD helps.   
   >   
   > One thing I have been able to narrow this down to is the utilization of    
   > the 4   
   > cores...if I shut-off the 3 remaining cores (using setproc utility) I am    
   > able to   
   > breeze through these processes.   
   >   
   > So...what's the config I'm using? Warp4.52, CP2 with FP6 applied, using    
   > the SMP   
   > doscall, os2krnl, etc, PSD=OS2APIC.PSD /P=4 /APIC, although I have tried a    
   > bunch   
   > of different /PIC and /PREC options, none of which had made a difference.   
   >   
   > What to do next? Well, using a nifty fixpak decode utility I decided to go   
   > through FP6 again and make sure that the correct SMP specific code was    
   > deployed   
   > on my machine. I am making the assumption that the "5639A5550 - CP2 for    
   > Warp   
   > Server for e-business (SMP)" is the final setup I want to duplicate to    
   > support   
   > my SMP switch. Since I do not have the correct SYSLEVEL file for that OS/2   
   > revision and therefore can not simply re-apply FP6 against that SYSLEVEL    
   > I'm   
   > taking a somewhat of a longer path here.   
   >   
   > I had previously applied FP6 when still using the Barton cpu, which was a   
   > non-SMP system. Specifically, I find that the FP6 OS2.2 subdirectory in    
   > the FIX   
   > directory tree decodes to the following deployments (just a sample    
   > record):   
   >   
   > KRNLRDEV.TDF @SMPDIR      KRNLRDEV.SD_ 13 10/17/06 08:42    4307 8598E6BB   
   > Packed, Add                         \OS2\INSTALL\SMP   
   > KRNLRDEV.TDF @UNIDIR      KRNLRDEV.UD_ 13 10/17/06 08:31    4232 819A1896   
   > Packed, Add                         \OS2\INSTALL\UNI   
   > KRNLRFS.TDF  @SMPDIR      KRNLRFS.SD_  13 10/17/06 08:42    4927 1B715B3D   
   > Packed, Add                         \OS2\INSTALL\SMP   
   > KRNLRFS.TDF  @UNIDIR      KRNLRFS.UD_  13 10/17/06 08:31    4874 A40C5818   
   > Packed, Add                         \OS2\INSTALL\UNI   
   >   
   > ...so the TDF files are being tossed into the SMP or UNI subdirectories of    
   > the   
   > \os2\install path, and thus my question, is this the FINAL resting place    
   > for   
   > these? Previously, I had the UNI processor stuff going to    
   > G:\os2\system\TRACE,   
   > but is the SMP configuration different?   
   >   
   > I ask because as I'm attempting to debug the hanging applications I'm    
   > trying to   
   > get a good working DUMP/DEBUG/TRACE environment...   
   >   
   > Thanks!   
   > -Dariusz   
   >   
   >    
      
      
   --- Internet Rex 2.31   
    * Origin: Arcor (1:261/20.999)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca