[LU-5310] build error in Centos 7

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.6.0
• Affects Version/s: Lustre 2.7.0
• Severity: 3
• Assignee: Bob Glossman (Inactive)
• Reporter: Bob Glossman (Inactive)

Description

server build now breaks in Centos 7/el7. client-only build still works OK.
example errors:

  CC [M]  /home/bogl/lustre-release/lustre/mdd/mdd_dir.o
/home/bogl/lustre-release/lustre/mdd/mdd_dir.c: In function ‘mdd_migrate_create’:
/home/bogl/lustre-release/lustre/mdd/mdd_dir.c:3396:26: error: argument to ‘sizeof’ in ‘memset’ call is the same expression as the destination; did you mean to dereference it? [-Werror=sizeof-pointer-memaccess]
  memset(mgr_ea, 0, sizeof(mgr_ea));
                          ^
cc1: all warnings being treated as errors
make[6]: *** [/home/bogl/lustre-release/lustre/mdd/mdd_dir.o] Error 1
make[5]: *** [/home/bogl/lustre-release/lustre/mdd] Error 2
make[4]: *** [/home/bogl/lustre-release/lustre] Error 2
make[3]: *** [_module_/home/bogl/lustre-release] Error 2
make[3]: Leaving directory `/home/bogl/rb/BUILD/kernel-3.10.0_123.4.2.l0708'
make[2]: *** [modules] Error 2
make[2]: Leaving directory `/home/bogl/lustre-release'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/bogl/lustre-release'
make: *** [all] Error 2

I believe this is due to the recent landing in master of commit de2d5808bd2987f76d2486272e1a9c192ba277d4; LU-5223 lmv: build master LMV EA dynamically build via readdir

It appears the gcc v4.8.2 in el7 is a bit pickier about some things than earlier gcc versions.

Issue Links

• is related to: LU-5223 generate DNE stripe directory LMV EA dynamically via readdir
• is related to: LU-5022 support for 3.10 rhel7 linux kernel

Comments

[Bob Glossman (Inactive)]

Have verified the build failure isn't specific to el7. It will happen with any gcc v4.8.x. I've seen it now with both v4.8.2 and v4.8.3. The fix is good for all cases so far.

[Bob Glossman (Inactive)]

I think the 1 line correction I commented on above & pushed as a patch is functionally correct as well as repairing the build, but I will appreciate good reviews on it to make sure I didn't mess up.

[Bob Glossman (Inactive)]

http://review.whamcloud.com/11035

[James A Simmons]

At least it is a nice simple patch.

[Brian Murrell (Inactive)]

It appears the gcc v4.8.2 in el7 is a bit pickier about some things than earlier gcc versions.

Isn't it always? Ultimately though, this is good though.

[Bob Glossman (Inactive)]

fwiw, changing the offending source line to:

memset(mgr_ea, 0, sizeof(*mgr_ea));

fixes the build.