On Fri, 2006-10-20 at 11:06 +1000, David Gibson wrote: > Please don't, having two versions of the ld wrapper is horrid - it > already parses the -m option to work out which sort of thing it's > building. Instead we should either change ld.hugetlbfs so it looks in > multiple places for the linker scripts, or (IMO simpler and better) > just put the linker scripts in "both" split and common locations, > using symlinks.
Ok, that makes sense. I like the idea of just searching multiple places. I'm not sure I understand your second, simpler, suggestion though. Could you explain a bit more please? > > 2) The ld.hugetlbfs script tries to return the highest "bit-ness" for the > > hardware, regardless of being in the 32-bit lib dir. This results in > > trying to use a non-existant ld script when linking our "easy" > > way. > > Hrm.. not exactly. ld.hugetlbfs attempts to mimic the behaviour of > the default linker when invoked without special options. On ppc64 > systems with 32-bit userspace (the normal case), this will mean a > 32-bit link by default, not the "highest bitness". I ran into this issue on an x86_64 Ubuntu box, which looked for the 64-bit scripts by default. -- Steve Fox IBM Linux Technology Center ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Libhugetlbfs-devel mailing list Libhugetlbfs-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel
