search for: attributesetnod

...hink it's practical to remove these inherently expensive APIs, but if we make AttributeBuilder easier to use, it will be a lot easier to fix these problems as they come up. Lastly, the Attribute APIs are hard to use because they do too much information hiding. In particular, the choice to make AttributeSetNode an internal implementation detail of lib/IR is problematic. This type describes all of the attributes on an individual argument, return value, or function, which IPO transforms often want. Today the getFnAttributes, getRetAttributes, and getParamAttributes APIs find the relevant AttributeSetNode*...

There are two 64-bit bitmasks maintained in AttributeImpl.h<https://sdocc.itg.ti.com/ui#file:review=11893/version=393846>: - AvailableFunctionAttrs is part of the AttributeListImpl class, and - AvailableAttrs is part of the AttributeSetNode class Both of these assume that the number of available enum attributes is limited to 64. In fact, a static_assert in Attributes.cpp<https://sdocc.itg.ti.com/ui#file:review=11893/version=393848> enforces that the number of enum attributes stays at 64 or below. However, the bitcode writer an...

I fell over this issue yesterday myself with lots of asserts being thrown. I think the issue is in lib/IR/AsmWriter.cpp:1618 in the function AssemblyWriter::printFunction(const Function *F) Looking at the code I think the 2nd for loop should be preceded by the test ... if (Idx < AS.getNumSlots()) Not sure why it doesn't fail on other platforms as it looks like it should be a genuine