fix fp-error test on ppc64el
Bug #1846216 reported by
Colin Ian King
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| stress-ng (Ubuntu) |
Fix Released
|
Medium
|
Colin Ian King | ||
Bug Description
running stress-ng --fp-error on jessie ppc64el will trigger errors because of comparisons of NaN and Inf constants against fp Nan and Infintity fp values. Add extra sanity checking to ensure we compare these correctly even when the constants comparisons are not exact.
| Changed in stress-ng (Ubuntu): | |
| importance: | Undecided → Medium |
| assignee: | nobody → Colin Ian King (colin-king) |
| status: | New → In Progress |
Revision history for this message
| Launchpad Janitor (janitor) wrote | #1 |
| Changed in stress-ng (Ubuntu): | |
| status: | In Progress → Fix Released |
To post a comment you must log in.
This bug was fixed in the package stress-ng - 0.10.07-1
---------------
stress-ng (0.10.07-1) unstable; urgency=medium
* Makefile: bump version
* Remove smatchify script, we now use other tools to static
analyse stress-ng
* core-helper: add sanity check if ttyname returns NULL to avoid
null pointer dereferencing
* stress-fp-error: fix checks for NaN and Inf (LP: #1846216)
* Don't show utf8 chars on dump consoles (LP: #1846196)
* stress-clone: make child process larger to make it more
OOM-able (LP: #1845948)
* core-out-of-memory: open and close proc interface after each write
* Add autobuild detection of getsid() for better portability
* stress-ng.h: only use x86 target clones for x86 arches
* stress-rdrand: unroll loop x 4 more times for more throughput
* inline the __keep_stressing function, reduce call overhead
for more throughput
* stress-stack: make alternative signal stack privately
mapped (LP: #1845464)
-- Colin King <email address hidden> Tue, 1 Oct 2019 16:39:52 +0100