[LU-6216] Compilation error libtool on ppc64 Created: 05/Feb/15 Updated: 03/Mar/15 Resolved: 03/Mar/15 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.7.0 |
| Fix Version/s: | Lustre 2.7.0, Lustre 2.8.0 |
| Type: | Bug | Priority: | Minor |
| Reporter: | Frank Heckes (Inactive) | Assignee: | Yang Sheng |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | ppc | ||
| Issue Links: |
|
||||
| Severity: | 3 | ||||
| Rank (Obsolete): | 17388 | ||||
| Description |
|
Upon building Lustre for ppc64 the following error breaks the build in Jenkins Job 'lustre-ppc': libtool: link: mpicc -g -Wall -fPIC -D_GNU_SOURCE -g -O2 -Werror -m64 -o mdsrate mdsrate.o -L/var/lib/jenkins/workspace/lustre-ppc/arch/ppc64/build_type/client/distro/el6/ib_stack/inkernel/BUILD/BUILD/lustre-2.6.93/lustre/utils -llustreapi ../../../libcfs/libcfs/libcfs.a -lkeyutils libtool: link: mpicc -g -Wall -fPIC -D_GNU_SOURCE -g -O2 -Werror -m64 -o cascading_rw cascading_rw.o lp_utils.o -L/var/lib/jenkins/workspace/lustre-ppc/arch/ppc64/build_type/client/distro/el6/ib_stack/inkernel/BUILD/BUILD/lustre-2.6.93/lustre/utils -llustreapi ../../../libcfs/libcfs/libcfs.a -lkeyutils cc1: warnings being treated as errors llapi_hsm_test.c: In function 'test101_progress': llapi_hsm_test.c:563: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test102_progress': llapi_hsm_test.c:599: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test103_progress': llapi_hsm_test.c:631: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test104_progress': llapi_hsm_test.c:669: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test105_progress': llapi_hsm_test.c:704: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test106_progress': llapi_hsm_test.c:736: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test107_progress': llapi_hsm_test.c:768: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test108_progress': llapi_hsm_test.c:803: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test109_progress': llapi_hsm_test.c:835: error: format '%llu' expects type 'long long unsigned int', but argument 8 has type '__u64' llapi_hsm_test.c: In function 'test110_progress': llapi_hsm_test.c:869: error: format '%llu' expects type 'long long unsigned int', but argument 9 has type '__u64' llapi_hsm_test.c: In function 'test111_progress': llapi_hsm_test.c:905: error: format '%llu' expects type 'long long unsigned int', but argument 9 has type '__u64' llapi_hsm_test.c: In function 'test112_progress': llapi_hsm_test.c:942: error: format '%llu' expects type 'long long unsigned int', but argument 9 has type '__u64' llapi_hsm_test.c:961: error: format '%llu' expects type 'long long unsigned int', but argument 9 has type '__u64' make[4]: *** [llapi_hsm_test.o] Error 1 Build log can be accessed via: |
| Comments |
| Comment by Jodi Levi (Inactive) [ 06/Feb/15 ] |
|
Yang Sheng, |
| Comment by Yang Sheng [ 12/Feb/15 ] |
|
We have ever build Lustre on ppc64 successful before. But maybe lost some thing in cleanup works. |
| Comment by Andreas Dilger [ 12/Feb/15 ] |
|
The llapi_hsm_test.c file was only just added in 2.6.92. It makes sense to change these printf lines to cast the arguments into (unsigned long long) to avoid the use of LPU64 in userspace. |
| Comment by Frank Zago (Inactive) [ 18/Feb/15 ] |
|
I'll take that one. I have some other changes in that file too. |
| Comment by Frank Zago (Inactive) [ 18/Feb/15 ] |
|
BTW, can the PPC build be accessible in the regular build page? I only see i386 and x86_64 builds. |
| Comment by James A Simmons [ 18/Feb/15 ] |
|
Totally agree. We need to add ppc64 to the mix since places like ORNL are picking up ppc64 machines. |
| Comment by Gerrit Updater [ 19/Feb/15 ] |
|
frank zago (fzago@cray.com) uploaded a new patch: http://review.whamcloud.com/13800 |
| Comment by Frank Heckes (Inactive) [ 19/Feb/15 ] |
|
PPC builds aren't part of the well known Jenkins jobs, but a separate one has been created : 'lustre-ppc' (https://build.hpdd.intel.com/job/lustre-ppc/). Drawback is that builds are on master and your change(s) won't be compiled on ppc64 till they're landed. |
| Comment by Andreas Dilger [ 20/Feb/15 ] |
|
Frank, it probably makes more sense to run the PPC builds on lustre-reviews rather than master, since it will essentially also test master (plus the current patch), but at least it has some chance of catching PPC regressions before they land. If this patch fixes the PPC builds, after it lands it would also be good to get the PPC builds to comment into Gerrit if there is a build failure. |
| Comment by Frank Zago (Inactive) [ 20/Feb/15 ] |
|
I'm all for it. A clang build-only on i386/x86_64 would be nice too. |
| Comment by Gerrit Updater [ 27/Feb/15 ] |
|
Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/13800/ |
| Comment by Peter Jones [ 03/Mar/15 ] |
|
Landed for 2.7 and 2.8 |