Messages in this thread | | | Date | Sat, 4 Apr 1998 13:44:07 -0800 (PST) | From | George Bonser <> | Subject | Re: 2.1.92, the woes continue |
| |
I can compile 2.1.92 all day long on the 486 with 16MB ram with only 1MB free at the start of the compile if I am running 2.1.89 or 2.0.32. I noticed the random crashing in .91
If I remove all processes that are not needed to run (kill the name server, gpm, ypbind, apache, etc) it will compile through while running 2.1.91 or .92. Is sure looks to me like the kernel is loosing track of the correct status of pages and when it gets into a "swap frenzy" will page stuff out and new stuff in while the code is being executed. It has all apperances of something overwriting code that is executing.
What leads me to say this is that the exact nature of the compile failures is inconsistant. It fails at different places and in different ways. Usually a Sig 11, once a 6, always while compiling a different part of the kernel. The only thing that is consistant is that it will not compile a complete kernel. The behavior came abruptly with 2.1.91
On Sat, 4 Apr 1998, Riley Williams wrote:
> > Running again went right through. This is the oddest thing I have seen. > > I am wondering if stuff in getting paged in on top of stuff that is > > running. I do not yet know enough about the internals to start wading > > through this but something is very wrong. I have no trouble with a fully > > loaded system running .89 (none of THESE troubles, I mean). > > Perhaps somebody could try compiling it whilst running something like > kernel 2.0.33 so as to ensure that it's not caused by a bug in the > kernel you're running... > > Best wishes from Riley. > > > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.rutgers.edu > >
George Bonser Just be thankful that Microsoft does not manufacture pharmaceuticals. http://www.debian.org Debian/GNU Linux ... the maintainable operating system.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |