From: Paolo Victor (paolovictor_at_gmail_dot_com)
Date: Thu Mar 17 2005 - 09:41:44 PST
Paul, Here's the output you asked for: > grep d_path /boot/System.map-2.4.26-1-386 c0142592 T __d_path c0173a5f T acpi_ns_valid_path_separator c0233987 R __kstrtab___d_path c023f328 R __ksymtab___d_path I believe that the kernel sources are not configured. How can you help me with this? Thanks, Paolo On Wed, 16 Mar 2005 13:41:22 -0800, Paul H. Hargrove <PHHargrove_at_lbl_dot_gov> wrote: > Paolo, > > Your sysadmin is both right and wrong. The kernel sources you used > are probably the correct source for building your kernel, but they are > not fully configured, or the version found in include/linux/version.h > (2.4.26) would have matched. Based on the configuration settings a > given kernel source can generate many (millions) of different kernels. > The differences among most of those potential kernels is small, but even > the smallest changes will cause the kernel's functions to appear at > different addresses in different builds. These differences are handled > using a symbol table. However, there are some difference (such as > configuring SMP support on or off) that are more significant because > they change the fields in some kernel data structures. To operate > correcly any loadable kernel module must agree with the running kernel > on the composition of any data structures it uses. This agreement can > optionally be checked at module load time using versioning on the > symbols in the symbol table. In the example error you provided 0e2ea584 > is a sort of checksum of the types of the arguments to the function > __d_path(). > > There are two possibilities now. One is that you'll need to get a > configured kernel source, which we may be able to help you with. The > other possibility is that the blcr kernel modules are incorrectly > assuming symbol versioning is being used when it is not enabled in your > kernel. To determine which is the case, please provide the output of > grep d_path /boot/System.map-2.4.26-1-386 > > -Paul > > Paolo Victor wrote: > > > Hello, > > > > I'm trying to install BLCR on my machine. I ran configure, make and > > make install, the libs and modules were created and put in the > > usr/local dir, but when I try to install the vmadump and blcr modules, > > some errors occurr, being all of them like this: > > > >> /usr/local/lib/blcr/2.4.26/vmadump_blcr.o: > >> /usr/local/lib/blcr/2.4.26/vmadump_blcr.o: unresolved symbol > >> __d_path_R0e2ea584 > > > > > > So, what could be wrong? Did I try to use a wrong version of the > > kernel sources? I suppose the problem lies on the kernel sources > > I'm using because when I run configure, I get this warning: > > > >> ******************************************************************* > >> ***** WARNING WARNING WARNING WARNING WARNING WARNING WARNING ***** > >> ******************************************************************* > >> * The kernel source does not match currently the running kernel. * > >> * Compilation will produce modules unsuitable for the currently * > >> * running kernel, which may not be what you intended. * > >> ******************************************************************* > >> ***** WARNING WARNING WARNING WARNING WARNING WARNING WARNING ***** > >> ******************************************************************* > >> ====================================================================== > >> Please review the following configuration information: > >> Kernel source directory = /usr/src/kernel-source-2.4.26 > >> Kernel build directory = /usr/src/kernel-source-2.4.26 > >> Kernel symbol table = /boot/System.map-2.4.26-1-386 > >> Kernel version probed from kernel build = 2.4.26 > >> Kernel running currently = 2.4.26-1-386 > >> ====================================================================== > > > > > > But when I asked the system admin if the versions where different, > > he said that the "-1-386" is just a kind of label and that the > > versions are the same. Could you help me with this? > > > > Thanks in advance for your help, > > Paolo Victor > > -- > 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 > >