Bug 1009145
Summary: | ldconfig warns of mismatch elf e_flags on armv7hl on upgrade | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Kyle McMartin <kmcmartin> | ||||||||
Component: | glibc | Assignee: | Carlos O'Donell <codonell> | ||||||||
Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||||
Severity: | unspecified | Docs Contact: | |||||||||
Priority: | unspecified | ||||||||||
Version: | 20 | CC: | blc, codonell, fweimer, jakub, kmcmartin, law, peterm, pfrankli, schwab, spoyarek | ||||||||
Target Milestone: | --- | ||||||||||
Target Release: | --- | ||||||||||
Hardware: | arm | ||||||||||
OS: | Unspecified | ||||||||||
Whiteboard: | |||||||||||
Fixed In Version: | glibc-2.18-11.fc20 | Doc Type: | Bug Fix | ||||||||
Doc Text: | Story Points: | --- | |||||||||
Clone Of: | Environment: | ||||||||||
Last Closed: | 2013-10-09 14:29:32 UTC | Type: | Bug | ||||||||
Regression: | --- | Mount Type: | --- | ||||||||
Documentation: | --- | CRM: | |||||||||
Verified Versions: | Category: | --- | |||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||
Embargoed: | |||||||||||
Attachments: |
|
Kyle, I would change the check to be a more conservative: dlib_ptr->flag & FLAG_ARM_LIBHF && ((flag & FLAG_ARM_LIBHF) || (flag & FLAG_ARM_LIBSF) || (flag & FLAG_ELF_LIBC6)) This mirrors the kind of check we have already done in other places to capture the transitional nature of this change. Then wrap it all in #ifdef __arm__ / #endif to avoid it being used on x86-64. Would you like me to check something like this into rawhide and f20? Cheers, Carlos. The most conservative check is actually: dlib_ptr->flag & FLAG_ARM_LIBHF && (flag & FLAG_ELF_LIBC6)) Which is what would probably be best e.g. was previously LIBHF and is now unmarked. c. This is fine with me. (In reply to Kyle McMartin from comment #3) > This is fine with me. Perfect, I'll pitch something into rawhide and f20 and give blc a poke when I have a build ready for testing. c. Scratch rawhide build with fix building here: http://koji.fedoraproject.org/koji/taskinfo?taskID=5984332 OK, rawhide buildroots broken, I'll build a rpm by hand. Created attachment 803014 [details]
rawhide glibc.spec with fix included.
Created attachment 803015 [details]
rawhide patch file for fix.
Rawhide scratch build complete: http://koji.fedoraproject.org/koji/taskinfo?taskID=5989387 f20 scratch build complete: http://koji.fedoraproject.org/koji/taskinfo?taskID=6006489 Kyle, Care to test the new f20 packages? I've tested this on my f20 alpha-ish trimslice by using yum to reinstall all packages providing /lib/*.so*. Before testing the f20 scratch build this produced warnings every time ldconfig ran. With the f20 scratch build no warnings were generated. Fixed in rawhide and f20. We'll push out a build shortly. glibc-2.18-11.fc20 has been submitted as an update for Fedora 20. https://admin.fedoraproject.org/updates/glibc-2.18-11.fc20 Package glibc-2.18-11.fc20: * should fix your issue, * was pushed to the Fedora 20 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing glibc-2.18-11.fc20' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2013-18239/glibc-2.18-11.fc20 then log in and leave karma (feedback). glibc-2.18-11.fc20 has been pushed to the Fedora 20 stable repository. If problems still persist, please make note of it in this bug report. clearing needinfo. |
Created attachment 798943 [details] ldconfig armhfp The version of binutils in F-20 was regressed from F-19 briefly, which resulted in our soft/hard float flags disappearing in libraries compiled in the F-20 mass rebuild. I've backported the patch into the F-20 binutils, but in order to avoid the ldconfig complaints on upgrades from F-19 to F-20, we either need to do something about the warning, or mass rebuild F-20 on armv7hl which isn't in the schedule. So, I've written a (admittedly, massively gross) patch for ldconfig which avoids the issue if the hfp flag has gone missing from e_flags. Carlos, I'd appreciate your input on this. --Kyle