While attempting an NFS install using anaconda-11.4.1.15-1 against http://porkchop.redhat.com/rawhide/nightly/rawhide-20080721/ppc/os I get the following error from loader: loader received SIGABRT! Backtrace: /sbin/loader(loaderSegvHandler+0xa0)[0x10008100] [0x100364] [0xffd2a578] /lib/libc.so.6(abort+0x260)[0xf8d1590] /lib/libc.so.6[0xf91ba0c] /lib/libc.so.6(cfree+0x48)[0xf91df08] /sbin/loader(mountNfsImage+0x504)[0x100196d4] /sbin/loader(main+0x9c4)[0x100094c4] /lib/libc.so.6[0xf8b5ac4] /lib/libc.so.6[0xf8b5c80] install exited abnormally [1/1] sending termination signals...done sending kill signals...done disabling swap... unmounting filesystems... you may safely reboot your system I should note that loader seemed confused with the mount path and at one point appended "images/stage.img" to the directory. After manually removing that text, it accepted the value and proceeded to the segfault above.
[Jul 23 16:25:01] <clumens> | if you want to be really helpful, you can mention that nfsinstall.c:166 is the troublesome line [Jul 23 16:25:10] <clumens> | which looks like this: free(((struct nfsInstallData *) loaderData->stage2Data)->directory); [Jul 23 16:25:23] <clumens> | that way i don't have to eu-addr2line again :)
I don't think it's possible to hit this segfault anymore, so please try anaconda-11.4.1.18-1 and let me know. I tried a couple different cmdline parameters here and wasn't able to trigger the segfault though I was still going through the same code path.
Tested (this time with kickstart) using anaconda-11.4.1.18-1 against http://porkchop.redhat.com/rawhide/nightly/rawhide-20080724/ppc/os loader received SIGABRT! Backtrace: /sbin/loader(loaderSegvHandler+0xa0)[0x100081e0] [0x100364] [0xfff2e568] /lib/libc.so.6(abort+0x260)[0xf8d1590] /sbin/loader(setKickstartNfs+0x360)[0x100190f0] /sbin/loader(runKickstart+0x9c)[0x100117cc] /sbin/loader(main+0x17a4)[0x1000a384] /lib/libc.so.6[0xf8b5ac4] /lib/libc.so.6[0xf8b5c80] install exited abnormally [1/1] We can close this one out and do a NEW (or DUP) bz if this is a different issue.
Ah, this is because I actually need to check the return value of asprintf. This will be fixed in the next build of anaconda.