I didn't want the endian issue to linger for too long, so I went ahead
and applied the s/32/64/ portion of Pete's chunkd endian fix proposed
patch (commit 468e7e393a6e12a410e1ab788c0beeb751c0225e).
As mentioned, I created type-safe static inline wrappers for the GLib
endian conversion functions. These were added to cld in
include/cld-private.h, and to chunkd in include/chunk-private.h.
The Fedora ppc/ppc64 build definitely proceeds, getting farther but not
100% of the way. Here's the koji scratch build I just performed:
http://koji.fedoraproject.org/koji/taskinfo?taskID=1595728
build log that shows ppc build failure:
http://koji.fedoraproject.org/koji/getfile?taskID=1595729&name=build.log
which shows
[...]
test failed on line 37
FAIL: large-object
lotsa-objects PUT: 2104.63 ops/sec
lotsa-objects syncing...
test failed on line 58
FAIL: lotsa-objects
build log that shows ppc64 build failure:
http://koji.fedoraproject.org/koji/getfile?taskID=1595731&name=build.log
which shows
[...]
large-object PUT: 111.06 MB/sec
chunkd[7298]: client ::ffff:127.0.0.1 error NoSuchKey
chunkd[7298]: client ::ffff:127.0.0.1 error InvalidArgument
chunkd[7298]: client ::ffff:127.0.0.1 connected
test failed on line 118
FAIL: large-object
chunkd[7298]: client ::ffff:127.0.0.1 connected
chunkd[7298]: client ::ffff:127.0.0.1 error NoSuchKey
test failed on line 41
FAIL: lotsa-objects
But hey... it gets farther than my previous chunkd builds on ppc, and
Pete definitely found an endian conversion bug.
Jeff
--
To unsubscribe from this list: send the line "unsubscribe hail-devel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
