- User Since
- Apr 27 2021, 11:57 AM (13 w, 1 d)
Fri, Jul 16
Is there still an issue preventing committing of this change?
Wed, Jul 7
Can someone commit this accepted change?
Jun 24 2021
The build failures seem unrelated to this patch. Can someone review, please?
Jun 23 2021
@peter.smith, thank you for the analysis! I'm consequently backing out of this patch.
Jun 22 2021
And what about this: https://bugs.llvm.org/show_bug.cgi?id=50721
Back ends are expected to crash when given invalid IR.
And how should the crash be fixed? Or shouldn't it?
Jun 15 2021
Not quite. Address spaces in LLVM are a somewhat opaque concept, and can mean anything. I'm just not sure they are exchangeable like you propose on all cases.
Basically, what that's saying is that *all* access to *any* address space is the same as an access through the default address space.
It may be true for your case, but not necessarily others. Your code would make it true for all.
I'm not an expert on atomics or the various uses of address spaces in the wild,...
I'm not an expert on atomics or the various uses of address spaces in the wild, but I'm curious as to what your use case is.
Jun 14 2021
Jun 11 2021
The build builders/53/builds/3020 fails due a time out. I don't think this is an actual failure. Maybe someone could increase the time out value for certain long-running tests.
The commit 789708617d20 caused two failures:
Jun 9 2021
@rengolin: Can you please commit this updated patch on my behalf?
I'll do that from now on. I'm sorry for the confusion.
@rengolin I tested the ARM tests that I thought would be affected. This appears to be affecting other tests.
@rengolin: I don't have commit access. Can you please commit this patch on my behalf?
Jun 8 2021
Doesn't this depend on which runtime library you're using? No Arm32 library implement that?
I remember a patch to add 128-bit arithmetic into Armv7 many years ago (from Android) but I'm not sure it ever got into compiler-rt.
This patch is intended to solve the bug 20871 for the ARM back-end.