Description of problem: While checking why pax-utils fails to build on AArch64 I noticed that we have ld-linux-aarch64.so.1 in both /lib64 and /lib while on x86-64 we only have /lib64/ld-linux-x86-64.so.2 file. Version-Release number of selected component (if applicable): 2.22.90-2 How reproducible: always Steps to Reproduce: 1. ls -l /lib*/ld-*so* Actual results: 14:01 hrw@pinkiepie-rawhide$ ll /lib*/ld-*so* -rwxr-xr-x. 1 root root 230600 08-17 18:04 /lib64/ld-2.22.90.so lrwxrwxrwx. 1 root root 13 08-17 17:59 /lib64/ld-linux-aarch64.so.1 -> ld-2.22.90.so lrwxrwxrwx. 1 root root 22 08-17 17:59 /lib/ld-linux-aarch64.so.1 -> ../lib64/ld-2.22.90.so Expected results: 14:01 hrw@pinkiepie-rawhide$ ll /lib*/ld-*so* -rwxr-xr-x. 1 root root 230600 08-17 18:04 /lib64/ld-2.22.90.so lrwxrwxrwx. 1 root root 13 08-17 17:59 /lib64/ld-linux-aarch64.so.1 -> ld-2.22.90.so Additional info: 14:03 <@pbrobinson> I see it on F-21 though too
Just to note: pax-utils 1.1.4 builds fine.
This bug appears to have been reported against 'rawhide' during the Fedora 24 development cycle. Changing version to '24'. More information and reason for this action is here: https://fedoraproject.org/wiki/Fedora_Program_Management/HouseKeeping/Fedora24#Rawhide_Rebase
Carlos any need for both?
Still an issue in F-25 $ ls -l /lib*/ld-*so* -rwxr-xr-x. 1 root root 230160 Aug 18 17:10 /lib64/ld-2.24.so lrwxrwxrwx. 1 root root 10 Aug 18 17:05 /lib64/ld-linux-aarch64.so.1 -> ld-2.24.so lrwxrwxrwx. 1 root root 19 Aug 18 17:05 /lib/ld-linux-aarch64.so.1 -> ../lib64/ld-2.24.so
We do not need the /lib/ link to ld.so. I don't know the history of it's presence, but I would think that after the mass rebuilds we've done it should not be needed any more by the distribution. Worse is that it will likely conflict with any future ILP32 ports currently being developed so we should remove it.
(In reply to Carlos O'Donell from comment #5) > We do not need the /lib/ link to ld.so. I don't know the history of it's > presence, but I would think that after the mass rebuilds we've done it > should not be needed any more by the distribution. Worse is that it will > likely conflict with any future ILP32 ports currently being developed so we > should remove it. Status? Can this go in F-27, there's been two mass rebuilds since.
This issue needs to go upstream to Arm and get solve there since there isn't a clear indication if we need the symlink for ABI compliance or not. I'm going to track this upstream with this bug: https://sourceware.org/bugzilla/show_bug.cgi?id=25129 I'm marking this issue as CLOSED/UPSTREAM since this is being tracked upstream.
Upstream confirmed this is not a bug and is required by the AArch64 ABI.