 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : Picomite: Possible issue in error() for CSubs
| Author | Message | ||||
twofingers Guru  Joined: 02/06/2014 Location: GermanyPosts: 1682 |
@matherp Hi Peter, I just played a bit with my BitOrderReverse code and I think there is an issue with the error() for CSubs. When I use it with a CMM2 (compiled for CMM2) it works fine, but the CSub for my Picomite crashes when it hits the error() function (compiled with option -o0). //Bitorderreverse CSub #include "PicoCFunctions.h" void bor(long long *orig, long long *nbits, long long *reversed) { int i = 0; unsigned int nb = *nbits; unsigned long long r = 0; unsigned long long ONEBIT = 1; if (nb){ nb-=1; for(;i<=nb;i++) { if (*orig & (ONEBIT<<i)) r += ONEBIT<<(nb-i); } } else { //not working! --> crash on PicoMite! error("0 is invalid (valid is 1 to 64)\r\n"); return; } *reversed=r; } Sample code for CMM2 (works) and Picomite (crashes): Bitorderreverse.zip BTW: A fast (7us/378MHz) bitorderreverse CSub for 8 bits on Picomite, for those who want it: Bor_8.zip I think there ist also an issue with MMPrintString(). But I haven't checked that thoroughly. This is not an urgent issue! (It seems I'm the only one using CSubs.  )I hope you are well! Best regards Michael causality ≠correlation ≠coincidence |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 10628 |
Almost certainly the handling of constant strings by the compiler. You can try various compiler options or search the CMM2 CSUB thread for using the original way of defining constant strings (a la MM2) |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1682 |
This is what I use: arm-none-eabi-gcc -c -mcpu=cortex-m0 -mfloat-abi=soft -mthumb -Wall -Wno-main -ffunction-sections -O0 -fPIC -I. main.c -o main.o Hmm, I think I've already tried all the options (even "GetTempMemory"). There aren't that many. At least that should work: #include "PicoCFunctions.h" void perro(char *s) { error(s); } perro("Hello") end 'File perror.bas written 06-08-2022 12:25:46 v1.44 CSUB perro 00000000 B082B580 6078AF00 681B4B04 687B001A 47900018 46BD46C0 BD80B002 10000364 End CSUB Do you have a working example for error()? Best regards Michael causality ≠correlation ≠coincidence |
||||
disco4now Guru Joined: 18/12/2014 Location: AustraliaPosts: 1055 |
I get the same result on latest firmware,i.e. reboots if bits =0 and MMPrintString() doesn't print, PutConsole too I think, but works on 5.0704 I had some other tests which work on this version but not the latest. I need to step through the firmware versions and see where it became broken. What version are you using. -------------------------------------------------- Connected to: PicoMite 5.0704 -------------------------------------------------- BitOrderReverse Demo Input: No of Bits : 0 [15] Bor(a,nBits,b) Error : 0 is invalid (valid is 1 to 64) Regards Gerry F4 H7FotSF4xGT |
||||
| matherp Guru Joined: 11/12/2012 Location: United KingdomPosts: 10628 |
It seems the later compiler has changed the location of the call table. Try updating the header file to set the base address at 310 #define BaseAddress 0x10000310 Thanks to Gerry for pointing out it worked in 5.07.04. As always fuller information yields an easy fix Edited 2022-08-06 23:01 by matherp |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1682 |
5.07.05b15 Hi Gerry, thanks for assisting! Very good idea to use a different firmware!  Best regards Michael causality ≠correlation ≠coincidence |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1682 |
Hi Peter, I can confirm my example #include "PicoCFunctions.h" void perro(char *s) { error(s); } now works! Seems to be solved! Thanks Peter, thanks Gerry! EDIT Working BitOrderReverse for version 5.070515: Bor(a,nBits,b) 'File bor.bas written 06-08-2022 15:15:40 v1.44 CSUB bor 00000000 4657B5F0 464E46DE B5E04645 B085680F 2F009203 1E7BD04E 6803469A 46982100 68432220 469C2000 23004691 90002601 E0229101 40910031 0032468B 409A4641 400A4660 40014659 D013430A 21204652 42491AD2 18524693 0031D419 000D4091 00324659 0014408A 9A019900 416A1909 92019100 429F3301 2220D010 189A4252 464AD5D8 1AD20031 468B40D1 4659E7D5 1A52464A 40D10031 E7E1000D 98009B03 60189901 B0056059 46BBBCF0 46A946B2 BDF046A0 48034B02 4478681B E7F24798 10000368 0000000A 73692030 766E6920 64696C61 61762820 2064696C 31207369 206F7420 0D293436 4700000A End CSUB I'm afraid this means that in the future we might need different CSubs for different FW versions.  Edited 2022-08-06 23:22 by twofingers causality ≠correlation ≠coincidence |
||||
| disco4now Guru Joined: 18/12/2014 Location: AustraliaPosts: 1055 |
So looks like from PicoMite V5.07.05b13 onwards the new entry for the baseaddress is required. F4 H7FotSF4xGT |
||||
| twofingers Guru Joined: 02/06/2014 Location: GermanyPosts: 1682 |
Thanks Gerry for the investigation! causality ≠correlation ≠coincidence |
||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2025 |
