artulab
projects / openbsd / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5b5a7f2) | patch
Make the floating-point computations a sligthly teeny bit more complex (but
authormiod <miod@openbsd.org>
Tue, 25 Oct 2022 19:55:31 +0000 (19:55 +0000)
committermiod <miod@openbsd.org>
Tue, 25 Oct 2022 19:55:31 +0000 (19:55 +0000)
commit80ef46263463eeb1861e5d6aabd0cb70bc5bef53
tree14bc6ba97a63e09150db45d21f5484c93b8aa26etree | snapshot
parent5b5a7f2d2890dd8c6563b7a877e3ed4adc26187dcommit | diff
Make the floating-point computations a sligthly teeny bit more complex (but
real), to prevent llvm panzers from optimizing too aggressively and generating
code which does not touch the slightest floating point register on some
platforms.

ok otto@
regress/sys/kern/signal/fpsig/fpsig.c diff | blob | history
OpenBSD src
by Ahmet Artu Yildirim