DavidSpickett added a comment.
The original change was https://reviews.llvm.org/D29078. Are you able to run
the tests with this applied?
================
Comment at: lldb/tools/intel-features/intel-mpx/cli-wrapper-mpxtable.cpp:66
- if ((lbound == one_cmpl64 || one_cmpl32) && ubound == 0) {
result.Printf("Null bounds on map: pointer value = 0x%" PRIu64 "\n",
value);
----------------
hawkinsw wrote:
> According to https://en.cppreference.com/w/cpp/language/operator_precedence,
> I would read the left operand of the `&&` as
>
> 1. The `==` has higher precedence than `||` so, `b = (lbound == one_compl64)`
> 2. `b || one_cmpl32`
>
> which does not seem like what the original author intended. I absolutely
> think that the fix is correct, but I just wanted to get everyone's feedback
> on whether this seems like more than just a "suspicious bitwise expression"
> (and more like a "mistaken bitwise expression").
>
> All that said, I could be completely, 100% wrong. And, if I am, feel free to
> ignore me!
The corrected code also makes sense given that MPX is some kind of memory
protection across ranges.
If `((lbound == one_cmpl64 || lbound == one_cmpl32) && ubound == 0)` is true
then upper bound < lower bound making an invalid range. Which is what I'd
expect for some default/uninitialised state (especially if zero size ranges are
allowed, so upper == 0 and lower == 0 couldn't be used).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D131312/new/
https://reviews.llvm.org/D131312
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
