3 ms·
Slightly terrifying is the optimizer may very well replace the following while(n--) *m++ = c; With a call to the built in memset()
by Gibbon1 10y ago
Slightly terrifying is the optimizer may very well replace the following
while(n--)
*m++ = c;
With a call to the built in memset()
- Spivak 10y agoI'm fairly inexperienced on this matter. Can you give a real life example where this would be a bad thing?
- temac 10y agoIf memset is actually not async signal safe on some target, and a compiler targetting the same does that transformation, the end results are very unsound...
- Gibbon1 10y agoWell there is this bug: 'memcpy implementation optimized as a call to memcpy' https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56888 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56888 The terrifying thing is while the programmer might know that memset() isn't async-signal-safe and not use it in a signal handler, the compiler may blissfully and silently optimize the code to use memset() anyways. Odd crashes or worse security leaks may result. Reminds me a bit of old floating point implemented in software. Some implementations had global scratchpad registers. Very weird things happened if you did any floating point operations in multiple threads.
- megous 10y agoIt does a lot more replacements, depending on the architecture. Even a simple a/b or a%b will get replaced with function call on many archs.