Red Hat Bugzilla – Bug 64673
GlibC 2.2.5 causes errors on somes program's
Last modified: 2016-11-24 10:14:24 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (compatible; Konqueror/3.0.0-10; Linux)
Description of problem:
When I try to install imici 1.0 and ethercap, both programs causes an error
not founding glibc libraries. but we have glibc 2.2.5.
other problem is on vmware that does not working with rh73-
Version-Release number of selected component (if applicable): 2.2.5
Steps to Reproduce:
2.Install vmware 3
Actual Results: program results:
VMware PANIC: (ide0:0) NOT_IMPLEMENTED F(831):688
VMware PANIC: (VMX) AIO: NOT_IMPLEMENTED F(831):688
imici: relocation error: imici: symbol __on_exit, version GLIBC_2.0 not
defined in file libc.so.6 with link time reference
You should be able to fix the VMware problem by upgrading VMware.
The vmware problem is bug in vmware, they incorrectly assume nice(3) will
return non-zero on error (while SuS and other standards require it to return
actual priority, error is signalled by returning -1 and setting errno).
AFAIK VMWare knows about this bug and have fixed it already.
Concerning imici, it is not supposed to use __on_exit which is a glibc private
interface (always has been). It has to use on_exit instead. Ask imici to fix
You have not listed any errors for ethercap, so I cannot say anything about it.
ok, VMWARE 3.1 and ETTERCAP 0.6.5 is now working
Could the nice problem be solved with the glibc symbol versioning?
Wasn't the versioning introduced just to support the this kind of
situtations where the interface must be changed but old binaries require the old
interface to function. Does the versioning slow down programs or is there any
other reason not to use it?
To get vmware 2.x series running I had to create my own preloaded library that
wraps the nice function and always returns 0. Unfortunately the vmware is a suid
program and thus the wrapper library needs to be in /usr/lib and the whole
program must be run as root to make it work.
This was a bug in nice(2) implementation (it did not work as documented)
and for bugfixes glibc doesn't come up with new symbol version.