From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Thu Aug 30 2007 - 08:03:34 PDT
On Thu, 30 Aug 2007, John Hodrien wrote: > On Thu, 30 Aug 2007, Paul H. Hargrove wrote: > > > I had already added code (only barely tested) to perform the relocation of > > our symbols captured from System.map, but did not fix the validation step. > > Since relocations need to obey the compiled-in PHYSICAL_ALIGN restriction, it > > should be easy to validate that the symbols differ from System.map by only a > > legal offset. > > > > John, > > Please try building from a BLCR source tarball with the attached patch > > applied (and not using kallsyms) and let me know if it works. > > Sure. The patch was nearly right bar a typo, as it should be > CONFIG_PHYSICAL_ALIGN I assume, rather than PHYSICAL_ALIGN. I'm also guessing > at that being a bitwise and. That then becomes: > > #if defined(CONFIG_RELOCATABLE) && defined(CONFIG_PHYSICAL_ALIGN) > unsigned long unaligned = offset1 & (CONFIG_PHYSICAL_ALIGN - 1); > > Which seems to work nicely. Is that fix correct? > > jh John, Yes, your version is what I had intended. Thanks much for the proof reading. By "seems to work nicely" do you mean that with this patch (suitable modified) you can load the module *and* have successfully run tests (e.g. "make check")? If you can confirm that you pass the tests, I'll roll the patch into 0.6.0. -Paul -- Paul H. Hargrove PHHargrove_at_lbl_dot_gov Future Technologies Group HPC Research Department Tel: +1-510-495-2352 Lawrence Berkeley National Laboratory Fax: +1-510-486-6900