Description of problem: After installing glibc-2.3.2-4.80.i686.rpm binaries do not use the libraries in the directory /lib/i686 Our binaries use the capabilities provided in the /lib/i686/libpthread-0.10.so and does not work with /lib/libpthread.so.0 Version-Release number of selected component (if applicable): glibc-2.3.2-4.80 How reproducible: Before installing glibc-2.3.2-4.80.i686.rpm bash-2.05$ ldd a.out libpthread.so.0 => /lib/i686/libpthread.so.0 (0x40021000) libc.so.6 => /lib/i686/libc.so.6 (0x40036000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) After installing glibc-2.3.2-4.80.i686.rpm -bash-2.05b$ ldd a.out libpthread.so.0 => /lib/libpthread.so.0 (0x40021000) libc.so.6 => /lib/libc.so.6 (0x40074000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) Steps to Reproduce: 1.ldd a.out 2.install glibc-2.3.2-4.80.i686.rpm 3.ldd a.out Actual results: libpthread is NOT used from /lib/i686 Expected results: libpthread is used from /lib/i686 Additional info: We depend on the i686 version of the libpthread library
A bug in platform handling was fixed in glibc, unfortunately kernel in 8.0 is broken and doesn't supply proper value for AT_PLATFORM in ELF aux vector. Please try ftp://people.redhat.com/jakub/glibc/errata/8.0/ which has the glibc fix backed out.
Hi Jakub, I have downloaded the new patches from ftp://people.redhat.com/jakub/glibc/errata/8.0/ but unfortunately the same behavior Sorry, I still need you help
Wow.. I consider this a pretty big problem for those who rely on the performance benefits of a properly targeted glibc. Jakub - any idea if this warrants a newly issued errata package (considering the amount of havoc the last update seems to have created...)?
Oops, sorry. There were more things needed to make /lib/i686 work on a buggy kernel. ftp://people.redhat.com/jakub/glibc/errata/8.0/*4.80.3* already uses /lib/i686 on the current 8.0 errata kernel, and unlike 2.2.93-5 even searches /lib/i686 with LD_LIBRARY_PATH=/lib
Hi Jakub, I have some good news! Your latest version works as expected. We did test it with our software and it works. Thanks for your help. When will this fix be available through errata?
Sorry I did not want to close it!
Has there been a kernel bug filed to track the problems that have caused "unnecessary" hacks to get glibc to properly work with the Red Hat kernel?
RHL9 (and RHL8 with the latest updates of glibc and kernel) shouldn't have any problems.