readnone

Project Infrastructure LLVM Dev List Archives
1

If a function accesses a global constant array of constants--in my case
a constant array of function pointers, and does NOT reference this
array via a function argument (it instead directly references this global),
is this enough to disallow the application of the readnone attribute to the
function in question?

Thanks in advance

Garrison

2

Garrison Venn wrote:

If a function accesses a global constant array of constants--in my case
a constant array of function pointers, and does NOT reference this
array via a function argument (it instead directly references this global),
is this enough to disallow the application of the readnone attribute to the
function in question?

A function is readnone if it doesn't access any mutable state, so reading a constant global is fine. (Incidentally, it may even contain stores to constants because any such operation is undefined.)

I was sure that I remember "opt -functionattrs" being taught to do this, but I just tried it out and it doesn't. This is a missed optz'n opportunity, testcase:

   @x = constant i32 0
   define void @foo() {
     load i32* @x
     ret void
   }

is only marked readonly when it should be readnone. Could I interest you in fixing that? :slight_smile:

Nick

3

Nick Lewycky wrote:

I was sure that I remember "opt -functionattrs" being taught to do this,
but I just tried it out and it doesn't. This is a missed optz'n
opportunity, testcase:

    @x = constant i32 0
    define void @foo() {
      load i32* @x
      ret void
    }

is only marked readonly when it should be readnone. Could I interest you
in fixing that? :slight_smile:

Never mind, it's because the default AA is no-aa, which answers no to everything. "opt -basicaa -functionattrs" gets this testcase right (and the one that stores, too!)

Nick

4

Thanks for the answer Nick. Now you got me on a chase through FunctionAttrs.cpp
trying to find out what analysis passes "... which answers no to everything", means. :slight_smile:

Garrison

5

Reading the comments for:

AliasAnalysis::onlyAccessesArgPointees(…) and
AliasAnalysis::doesAccessArgPointees(…)

And uses of these signatures in FunctionAttrs.cpp such as:

if (AliasAnalysis::onlyAccessesArgPointees(MRB)) {
// If the call does access argument pointees, check each argument.
if (AliasAnalysis::doesAccessArgPointees(MRB))

Wouldn’t AliasAnalysis::onlyAccessesArgPointees(…) being renamed something like:

AliasAnalysis::atMostAccessesArgPointees(…), be clearer?

The “only” prefix throws me off at least.

Sorry if this is noise.

Garrison