Menu
▾
▴
[sdcc-devel] [ sdcc-Bugs-1723128 ] Code compiles 2.6.0 but causes fatal error on 2.7.0
|
From: SourceForge.net <no...@so...> - 2007-05-28 05:18:17
|
Bugs item #1723128, was opened at 2007-05-21 23:10 Message generated for change (Comment added) made by grahamnz You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1723128&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: fixed Status: Closed Resolution: Fixed Priority: 5 Private: No Submitted By: GrahamNZ (grahamnz) Assigned to: Maarten Brock (maartenbrock) Summary: Code compiles 2.6.0 but causes fatal error on 2.7.0 Initial Comment: Hello SDCC support team! I have been encouraged to use a snapshot of the SDCC compiler due to a problem I was experiencing. The snapshot I have is 2.7.0 4806 (21 May 2007) and when compiling my old project (which compiles correctly on 2.6.0) I get the following in the message window: Invoking project defined compiler. C:\Program Files\SDCC\bin\SDCC.EXE -c --debug --use-stdout -V --model-small -IC:\Work\impedancemodule C:\Work\impedancemodule\main.c ... + C:\PROGRA~1\SDCC\bin\sdcpp.exe -nostdinc -Wall -std=c99 -I"C:\Work\impedancemodule" -obj-ext=.rel -DSDCC_MODEL_SMALL -DSDCC=270 -DSDCC_mcs51 -D__mcs51 -I"C:\Program Files\SDCC\bin\..\include\mcs51" -I"C:\Program Files\SDCC\bin\..\include" "C:\Work\imp dancemodule\main.c" + C:\PROGRA~1\SDCC\bin\asx8051.exe -plosgffc "main.asm" Invoking project defined compiler. C:\Program Files\SDCC\bin\SDCC.EXE -c --debug --use-stdout -V --model-small -IC:\Work\impedancemodule C:\Work\impedancemodule\measure.c ............. + C:\PROGRA~1\SDCC\bin\sdcpp.exe -nostdinc -Wall -std=c99 -I"C:\Work\impedancemodule" -obj-ext=.rel -DSDCC_MODEL_SMALL -DSDCC=270 -DSDCC_mcs51 -D__mcs51 -I"C:\Program Files\SDCC\bin\..\include\mcs51" -I"C:\Program Files\SDCC\bin\..\include" "C:\Work\imp dancemodule\measure.c" + C:\PROGRA~1\SDCC\bin\asx8051.exe -plosgffc "measure.asm" Invoking project defined compiler. C:\Program Files\SDCC\bin\SDCC.EXE -c --debug --use-stdout -V --model-small -IC:\Work\impedancemodule C:\Work\impedancemodule\comms.c Assertion failed: !strcmp (s, "c"), file /home/sdcc-builder/build/sdcc-build/orig/sdcc/src/mcs51/gen.c, line 1673 This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information. Caught signal 22: SIGABRT + C:\PROGRA~1\SDCC\bin\sdcpp.exe -nostdinc -Wall -std=c99 -I"C:\Work\impedancemodule" -obj-ext=.rel -DSDCC_MODEL_SMALL -DSDCC=270 -DSDCC_mcs51 -D__mcs51 -I"C:\Program Files\SDCC\bin\..\include\mcs51" -I"C:\Program Files\SDCC\bin\..\include" "C:\Work\imp dancemodule\comms.c" sdcpp.exe: fatal error: when writing output to : m Invoking project defined compiler. C:\Program Files\SDCC\bin\SDCC.EXE -c --debug --use-stdout -V --model-small -IC:\Work\impedancemodule C:\Work\impedancemodule\config.c .. + C:\PROGRA~1\SDCC\bin\sdcpp.exe -nostdinc -Wall -std=c99 -I"C:\Work\impedancemodule" -obj-ext=.rel -DSDCC_MODEL_SMALL -DSDCC=270 -DSDCC_mcs51 -D__mcs51 -I"C:\Program Files\SDCC\bin\..\include\mcs51" -I"C:\Program Files\SDCC\bin\..\include" "C:\Work\imp dancemodule\config.c" Caught signal 11: SIGSEGV Invoking project defined compiler. C:\Program Files\SDCC\bin\SDCC.EXE -c --debug --use-stdout -V --model-small -IC:\Work\impedancemodule C:\Work\impedancemodule\Flash.c + C:\PROGRA~1\SDCC\bin\sdcpp.exe -nostdinc -Wall -std=c99 -I"C:\Work\impedancemodule" -obj-ext=.rel -DSDCC_MODEL_SMALL -DSDCC=270 -DSDCC_mcs51 -D__mcs51 -I"C:\Program Files\SDCC\bin\..\include\mcs51" -I"C:\Program Files\SDCC\bin\..\include" "C:\Work\imp dancemodule\Flash.c" + C:\PROGRA~1\SDCC\bin\asx8051.exe -plosgffc "Flash.asm" Link in progress... C:\Program Files\SDCC\bin\SDCC.EXE --debug --use-stdout -V --xram-size 2048 --code-size 32768 --iram-size 256 --model-small -Wl -bBSEG=0024 -Wl -bXSEG=0015 -IC:\Work\impedancemodule -oC:\Work\impedancemodule\ImpMod.hex C:\Work\impedancemodule\main.rel C:\Work\impedancemodule\measure.rel C:\Work\impedancemodule\comms.rel C:\Work\impedancemodule\config.rel C:\Work\impedancemodule\Flash.rel . C:\Work\impedancemodule\comms.rel: cannot open. + C:\PROGRA~1\SDCC\bin\aslink.exe -nf "C:\Work\impedancemodule\ImpMod.lnk" The problems are the: - Assertion failed (on a file not present on my machine) - which is followed by 'runtime termination in an unusual way' - Caught signal 22: SIGABRT - Fatal error: writing output to: m - Caught signal 11: SIGSEGV As I mentioned previously this fileset compiled and produced functional code. None of the above messages provide any reference to a problem line or statement in my source. Rather they seem to indicate an internal issue. The messages also state 'Please contact the applications support team for more information'. Hence this report! Help and assistance appreciated. Regards, Graham ---------------------------------------------------------------------- >Comment By: GrahamNZ (grahamnz) Date: 2007-05-28 05:18 Message: Logged In: YES user_id=1797831 Originator: YES Hi Maarten, I have downloaded and installed 4815 and do not get the errors that I originally encountered. However, I did get the following: Link in progress... C:\Program Files\SDCC\bin\SDCC.EXE --debug --use-stdout -V --xram-size 2048 --code-size 32768 --iram-size 256 --model-small -Wl -bBSEG=0024 -Wl -bXSEG=0015 -IC:\Work\impedancemodule -oC:\Work\impedancemodule\ImpMod.hex C:\Work\impedancemodule\main.rel C:\Work\impedancemodule\measure.rel C:\Work\impedancemodule\comms.rel C:\Work\impedancemodule\config.rel C:\Work\impedancemodule\Flash.rel ... memory overlap near 0x0 for XABS0 memory overlap near 0x0 for XABS0 memory overlap near 0x0 for XABS0 memory overlap near 0x0 for XABS0 + C:\PROGRA~1\SDCC\bin\aslink.exe -nf "C:\Work\impedancemodule\ImpMod.lnk" and the code does not appear to run correctly. This I will have to investigate tomorrow. Comments welcome? Especially if you can confirm the memory area that this is referring to. Regards, Graham ---------------------------------------------------------------------- Comment By: Maarten Brock (maartenbrock) Date: 2007-05-25 06:09 Message: Logged In: YES user_id=888171 Originator: NO The other issue was easier to fix and thus both are fixed now in SDCC #4814. Graham, if you're still concerned about your code being out in the open, feel free to remove it from this tracker item. ---------------------------------------------------------------------- Comment By: Maarten Brock (maartenbrock) Date: 2007-05-24 22:30 Message: Logged In: YES user_id=888171 Originator: NO No, not all issues, only the ones in comms.c The SIGSEGV in config.c is still there. ---------------------------------------------------------------------- Comment By: GrahamNZ (grahamnz) Date: 2007-05-24 21:10 Message: Logged In: YES user_id=1797831 Originator: YES Hi Borut and Maarten, Does that mean all the issues I raised have been resolved? Let me know when the next snapshot is available and I will try it on my full source. Graham ---------------------------------------------------------------------- Comment By: Borut Razem (borutr) Date: 2007-05-24 21:06 Message: Logged In: YES user_id=568035 Originator: NO Maarten, please apply the patch. Borut ---------------------------------------------------------------------- Comment By: Maarten Brock (maartenbrock) Date: 2007-05-24 15:43 Message: Logged In: YES user_id=888171 Originator: NO I have attached a regression test and a fix for the assertion error. I need approval from the release manager to apply it (Borut...?) File Added: fix.patch ---------------------------------------------------------------------- Comment By: Maarten Brock (maartenbrock) Date: 2007-05-24 15:36 Message: Logged In: YES user_id=888171 Originator: NO File Added: bug1723128.c ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-05-24 11:18 Message: Logged In: NO grahamnz - your source (SDCC270Error.zip) is publicly available to download here (link at bottom of page), in case you are unaware. ---------------------------------------------------------------------- Comment By: GrahamNZ (grahamnz) Date: 2007-05-23 04:46 Message: Logged In: YES user_id=1797831 Originator: YES Hello Maarten, Thanks for upgrading this - I presume it has sparked a high level of concern due to the imminent release of 2.7.0. Today I have pruned our code to a level where it is possible to give to you for your research into this problem, while maintaining whatever is required for the error conditions to still be reported. Please note that it is extracted from a commercial product and so we do expect that it will only be used for your research and not destributed further. I have attempted to reduce the overall line count and files involved in the process, and this has resulted in the removal of two files and quite some code. This has not resulted in a 'tiny' test file but I hope that it will be of use all the same. The project consists of three files with their respective headers, plus a device header file. These files have been zipped and uploaded. While reducing the file sizes I noticed that the 'sdcpp.exe fatal error' would come and go. In fact compile the program with no changes and 1 in 10 or 1 in 20 would produce the error. Originally I thought this error had gone and was related to the line(s) just removed - it was with some relief that I determined it was quite random in nature. The other errors of 'caught signals' and 'assertion fails' seemed to be 'hard' faults in that they always appeared. The 'caught signal 11: SIGSEGV' error can be avoided (apparently) by removing the AUTOCAL_CFG structure and the reference to it. I do not know if this is because of a problem in the structure definition or because of a 'overflow' condition in the compiler/preprocessor. In all my manipulations to the 'comms.c' file I was unable to stop the 'assertion failed' or 'caught signal' messages from appearing. I look forward to hearing about the results of your investigation. Let me know when there is a new snapshot and I will try it out for you. If you require other details let me know. Regards, Graham File Added: SDCC270Error.zip ---------------------------------------------------------------------- Comment By: Maarten Brock (maartenbrock) Date: 2007-05-22 06:45 Message: Logged In: YES user_id=888171 Originator: NO Hi Graham, This looks like a real bug, so I reclassified it. Can you please attach the sources that produce these errors? Try to make it as small as possible if you can and include the necessary header files. Greets, Maarten ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1723128&group_id=599 |