[#FL-2711] lvm2 fails to compile with new gcc - FITS

Foresight Linux

lvm2 fails to compile with new gcc

Created: 18/Feb/11 05:10 PM Updated: 12/Jan/13 09:26 PM

Component/s:

None

Affects Version/s:

2.1.2

Fix Version/s:

None

Security Level:

Public (Everyone can see this issue)

Time Tracking:

Not Specified

File Attachments:

lvm2-2.02.72-conflict.patch (3 kB)
2.

lvm2.recipe (5 kB)

Description

« Hide

The new gcc makes lvm2 fail to compile with:

gcc -O2 -g -D_FORTIFY_SOURCE=2 -fstack-protector -O2 -g -D_FORTIFY_SOURCE=2 -fstack-protector -fPIC -Wall -Wundef -Wshadow -Wcast-align -Wwrite-strings -Wmissing-prototypes -Wmissing-declarations -Wnested-externs -Winline -Wmissing-noreturn -Wformat-security -Wredundant-decls -O2 -O2 -g -D_FORTIFY_SOURCE=2 -fstack-protector -fPIC -Wall -Wundef -Wshadow -Wcast-align -Wwrite-strings -Wmissing-prototypes -Wmissing-declarations -Wnested-externs -Winline -Wmissing-noreturn -Wformat-security -Wredundant-decls -O2 -g -O1  -g -O1  -Wl,--export-dynamic -L./libdm -L./lib -L./daemons/dmeventd -g -O1  -Wl,--export-dynamic -L../libdm -L../lib -L../daemons/dmeventd -static -L../libdm/ioctl \
              -o dmsetup.static dmsetup.o -ldevmapper   -lm  
/usr/bin/ld: dynamic STT_GNU_IFUNC symbol `strcmp' with pointer equality in `/usr/lib64/gcc/x86_64-unknown-linux/4.4.4/../../../../lib64/libc.a(strcmp.o)' can not be used when making an executable; recompile with -fPIE and relink with -pie
collect2: ld returned 1 exit status

of course if I disable the static building flag, it succeeds. I don't really understand enough to resolve this issue.

All

Comments

Work Log

Change History

Sort Order:

[ Permalink | « Hide ]

Adam Chasen added a comment - 19/Feb/11 12:44 AM

out of date binutils?
http://forums.gentoo.org/viewtopic-t-840456-start-0.html
http://bugs.gentoo.org/332905

[ Permalink | « Hide ]

Adam Chasen added a comment - 19/Feb/11 12:52 AM

attached working recipe (also with clvmd on)

[ Permalink | « Hide ]

Adam Chasen added a comment - 23/Feb/11 07:50 AM

I have this working

[ Permalink | « Hide ]

Adam Chasen added a comment - 23/Feb/11 10:34 AM

lvm2:source=/foresight.rpath.org@fl:2-devel//2-devel//rmake-repos.eng.rpath.com@local:2-devel/2.02.84-0.4.1[is: x86(~cmov,~i486,~i586,~i686,~mmx,~nx,~sse,~sse2)]{x86} ->
device-mapper=/foresight.rpath.org@fl:2-devel//2-devel/2.02.84-0.5-1[is: x86]
lvm2=/foresight.rpath.org@fl:2-devel//2-devel/2.02.84-0.5-1[is: x86]

lvm2:source=/foresight.rpath.org@fl:2-devel//2-devel//rmake-repos.eng.rpath.com@local:2-devel/2.02.84-0.4.1[is: x86_64(~3dnow,~3dnowext,~nx)]{x86_64} ->
device-mapper=/foresight.rpath.org@fl:2-devel//2-devel/2.02.84-0.5-1[is: x86_64]
lvm2=/foresight.rpath.org@fl:2-devel//2-devel/2.02.84-0.5-1[is: x86_64]
lvm2:source=/foresight.rpath.org@fl:2-devel//2-devel/2.02.84-0.5[]

+ Updating /tmp/lvm2 after commit

[ Permalink | « Hide ]

António Meireles [aka doniphon] added a comment - 12/Sep/11 04:59 AM

achasen,

the patch you have actually works. the trick is that since it touches configure.in one just needs to call autoreconf -fi before calling configure (as configure files need to be regenerated)

[ Permalink | « Hide ]

António Meireles [aka doniphon] added a comment - 12/Sep/11 06:12 AM

Adam,

Can you build/test locally lvm2:source=/foresight.rpath.org@fl:2-devel/2.02.84-3 ?
(it needs for build/run http://lists.rpath.org/pipermail/foresight-commits/2011-September/050206.html and http://lists.rpath.org/pipermail/foresight-commits/2011-September/050210.html which will land in 2-devel groups in next few hours)