Re: Strange core dump in newfs_lfs(8).
Nick Hudson <[email protected]> Thu, 14 Dec 2006 15:48:17 +0000
| Newsgroups | gmane.os.netbsd.ports.hp700 |
|---|---|
| Message-ID | <[email protected]> |
On Monday 11 December 2006 19:20, Jochen Kunz wrote: > Hi. > > I tried to use LFS on one of my hp700 machines. newfs_lfs(8) dumped core > with a segmentation fault immediately. ktrace(1) showed that this must > happen quite early in startup. I recompiled newfs_lfs(8) with a debug > printf(3) as the first statement in main(): Segmentation fault (core > dumped), without any output. I renamed main() to main2(), added a new > main() that does nothing else as a debug printf(3) and then calling > main2(). This time I get output from the debug printf(3) in main() but > still as soon as main2() (the real main()) is called: Segmentation fault > (core dumped). I suspected some stack interaction and moved all > variables to the outside of main() to make them glabal, i.e. BSS instead > stack. Still the same. Segmentation fault as soon as main() is caled. > gdb(1) doesn't work and now I am lost. I don't have an idea what is > happenig here. Version is -current as of Friday and I had the same > problem with a build from late November. Increase your stack limit. Nick