[linrad] Re: Linrad 02.05 part II, svgalib version.
I do not undertand why this happens, because I chose N to the question: use mlockall to prevent swapping ? (Y/N) - N This is a trivial bug. There should have been a test if(ui.memloc==0) Ok. The xz() function has been a very efective way of finding where the crash occured. Thanks for the tip! In 02.06 and later renamed to if(ui.no_memlock == FALSE) I have started to use TRUE/FALSE for variables used as boolean variables and sometimes the variable name has to change to reflect whether a 0 or a 1 matches what the name suggests. (I do not want to change parameter values) Understood. Early like this the mlocall call seldomly fails, that is why I never detected it. As a temporary solution, just remove the line lir_lock_mem(); Yes, mlockall fails with consistence here within linrad, but if I do this simple program never fails here. #include sys/mman.h #include stdio.h void main(void){ int i; i=mlockall(MCL_CURRENT); printf(returned value is %d \n,i); } It always return a 0 value and never crashes Who knows!! :-) 73, Ramiro. There will be a greatly improved Lir-02.06 tomorrow:-) We will wait for the new version Thank you very much!!! # This message is sent to you because you are subscribed to the mailing list linrad@antennspecialisten.se. To unsubscribe, E-mail to: [EMAIL PROTECTED] To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED] To switch to the INDEX mode, E-mail to [EMAIL PROTECTED] Send administrative queries to [EMAIL PROTECTED]
[linrad] Re: Linrad 02.05 part II, svgalib version.
[EMAIL PROTECTED] wrote: Hi Ramiro, The way I locate this kind of errors is th change DUMPFILE, now in vernr.h, to 1. Hello Leif, I have done the debugging you suggested. After some compilation and xz iterations, I have found that the funcion that completly crashes my computer is mlockall(), called from function lir_lock_mem() in lsys.c. lir_lock_mem() is called from main() in lmain.c The call to mlockall() never returns and crashes the computer. I do not undertand why this happens, because I chose N to the question: use mlockall to prevent swapping ? (Y/N) - N Svgalib mouse settings are ok cause mousetest works fine. Tell me if you want me to do more tests. Ramiro. EA1ABZ. # This message is sent to you because you are subscribed to the mailing list linrad@antennspecialisten.se. To unsubscribe, E-mail to: [EMAIL PROTECTED] To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED] To switch to the INDEX mode, E-mail to [EMAIL PROTECTED] Send administrative queries to [EMAIL PROTECTED]
[linrad] Re: Linrad 02.05 part II, svgalib version.
Hi Leif, Forgot to post the debugging xz: int lir_lock_mem(void) { int i,k; xz(lir_lock_mem_1); i= mlockall(MCL_CURRENT); xz(lir_lock_mem_2); if(i==0)return 0; k=errno; munlockall(); errno=k; return i; } More information about the mlockall issue. This dirty code modification removes the first crash and linrad screen appears and works well during some time, until it crashes again when I press X. The crash blocks all pure text consoles, but I can get GNOME with ctrl-alt-F7. I run ps ax and can not locate any linrad process running in order to kill them. int lir_lock_mem(void) { int i,k; //i= mlockall(MCL_CURRENT); i=0; if(i==0)return 0; k=errno; munlockall(); errno=k; return i; } I do not understand anything... :-( Regards. Ramiro. EA1ABZ # This message is sent to you because you are subscribed to the mailing list linrad@antennspecialisten.se. To unsubscribe, E-mail to: [EMAIL PROTECTED] To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED] To switch to the INDEX mode, E-mail to [EMAIL PROTECTED] Send administrative queries to [EMAIL PROTECTED]
[linrad] Re: Linrad 02.05 part II, svgalib version.
Hi Ramiro, After some compilation and xz iterations, I have found that the funcion that completly crashes my computer is mlockall(), called from function lir_lock_mem() in lsys.c. lir_lock_mem() is called from main() in lmain.c The call to mlockall() never returns and crashes the computer. I do not undertand why this happens, because I chose N to the question: use mlockall to prevent swapping ? (Y/N) - N This is a trivial bug. There should have been a test if(ui.memloc==0) In 02.06 and later renamed to if(ui.no_memlock == FALSE) I have started to use TRUE/FALSE for variables used as boolean variables and sometimes the variable name has to change to reflect whether a 0 or a 1 matches what the name suggests. (I do not want to change parameter values) Early like this the mlocall call seldomly fails, that is why I never detected it. As a temporary solution, just remove the line lir_lock_mem(); There will be a greatly improved Lir-02.06 tomorrow:-) 73 Leif # This message is sent to you because you are subscribed to the mailing list linrad@antennspecialisten.se. To unsubscribe, E-mail to: [EMAIL PROTECTED] To switch to the DIGEST mode, E-mail to [EMAIL PROTECTED] To switch to the INDEX mode, E-mail to [EMAIL PROTECTED] Send administrative queries to [EMAIL PROTECTED]