search for: constantaggregates

On Wed, Mar 25, 2020 at 8:52 AM Doerfert, Johannes <jdoerfert at anl.gov> wrote: > I think the solution space for the value/use-list issue might be larger > than what was mentioned so far. > > > Some random thoughts: > > If no pass ever walks the use list of a constant, except globals which > we could handle differently, we could get rid of their use-list or >

>The type you pass to GlobalVariable's constructor for that variable should be "[10 x %struct.dlist]" because that's what you want storage for. Then the GlobalVariable itself will be a Constant of type "[10 x %struct.dlist]*". Yes, I verified that this is the case. I enabled assertions and the error seems to occur while creating GlobalVariable for both struct dhash

Thank you very much. The code to initialize h1 to non-zero values was what I was looking for. It's almost working except for a type mismatch wrt dlist* llist field of h1. dlist static_lst[10] = { {1, 5, NULL}, ... }; dhash h1[10] = {{"myfile.txt", static_lst}, ... }; Along the lines of the code you had sent, I created a GlobalVariable* llist of type [10 x %struct.dlist]* for the

> On Mar 20, 2020, at 12:34 PM, Nicholas Krause <xerofoify at gmail.com> wrote: > >> >> The problem isn’t constants or functions themselves, it is that they are instances of llvm::Value. Everything that walks a use/def list would have to run code that checks for this, and every call to inst->setOperand() would have to do locking or conditional locking. This would be

Hi: For my own Transform passes I need to lower all ConstantExpressions into Instructions in order to conveniently analyze the use-def def-use chains and operand comparisons. Could anyone possibly kindly share some insights on if there is any standard way to do this without crafting my own implementation and if not, should we add this feature into LLVM upstream? Zhang