|
Re: Seeing memory corruption, GC moves my objects around: msg#00332lisp.cmucl.devel
Martin Cracauer <cracauer@xxxxxxxx> writes: > Generally, any suggestions what I can do to figure out what is going > on here? Assuming you've eliminated hardware error as a possible cause (memtest86 is good for finding that, everyone says)... gdb is your friend, I suspect. If it's always the same location being corrupted, run a garbage collection with gdb attached and a watchpoint at that location (one of the nicest things about the x86 architecture is that it has hardware watchpoints) and then at least you have a call stack and can look at the data. Is it in fromspace/tospace/something else? Is the page the object lives on still marked as allocated, or has it been marked free? Can you use search_dynamic_space (or manually grope around in memory) to find the object that the slot is part of, and is it the same object as you think it is? The good news is that if you've got a reproducable test case, that's 80% of the work done - now you only have the other 80% left ... -dan -- http://www.cliki.net/ - Link farm for free CL-on-Unix resources
|
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Re: side-effect of (rename-package "LISP" "LISP") [in code/exports.lisp]: 00332, Gerd Moellmann |
|---|---|
| Next by Date: | symbol/package conflicts in code/export.lisp: 00332, Sean Champ |
| Previous by Thread: | Re: Seeing memory corruption, GC moves my objects aroundi: 00332, Martin Cracauer |
| Next by Thread: | side-effect of (rename-package "LISP" "LISP") [in code/exports.lisp]: 00332, Sean Champ |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |