[Bug 211152] benchmarks/iozone: Build fails on typedef redefinition with different types ('long long' vs '__off64_t' (aka 'long'))

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211152

--- Comment #17 from Mark Millard <markmi at dsl-only.net> ---
(In reply to w.schwarzenfeld from comment #16)

I do not have all that much time today or tomorrow for this but I took a quick
look today.

The new iozone.c updates address most of my concerns. As far as FreeBSD amd64
and armv6 goes it appears to compile only with a complaint about
CACHE_LINE_SIZE being redefined. I expect that powerpc, powerpc64, and the rest
would likely get the same results. (There is a separate, earlier submittal
(210857) reporting the CACHE_LINE_SIZE override.)

The only part that might not fit as an upstream update to the iozone.c source
code might be the sscanf use of (long long *) casts if off64_t can ever have a
different memory layout from long long. (Not a likely issue, even outside
FreeBSD?)

So an updated /usr/ports/benchmarks/iozone/files/patch-iozone.c that has both
its old changes to iozone.c and your new changes to iozone.c covered would be
an improvement to the FreeBSD benchmarks/iozone port.

-- 
You are receiving this mail because:
You are the assignee for the bug.