[jdk16] RFR: 8255763: C2: OSR miscompilation caused by invalid memory instruction placement

[Roberto Castañeda Lozano]

On Wed, 16 Dec 2020 20:41:09 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

> Is it possible to avoid "hoist inversion" if we take into account BCI and other information instead of only frequency?

We could avoid it if

1) we followed the original heuristic proposed in Click's [Global code motion/global value numbering](https://courses.cs.washington.edu/courses/cse501/04wi/papers/click-pldi95.pdf) paper:

_"We choose the block that is in the shallowest loop nest possible, and then is as control dependent as possible."_

and

2) we had perfect loop information (including information about irreducible loops).

Unfortunately, information about irreducible loops is currently discarded when building the CFG-loop tree: 

https://github.com/openjdk/jdk16/blob/ce0ab2dd848484ad1837536942d42f19a509989b/src/hotspot/share/opto/gcm.cpp#L1657-L1659

> Is it possible compute_freq() take into account irreducible loops?

I think so. compute_freq() computes the frequencies of each block within a loop L in a single forward pass (in reverse DFS postorder) over the members (blocks and loops) of L:

https://github.com/openjdk/jdk16/blob/ce0ab2dd848484ad1837536942d42f19a509989b/src/hotspot/share/opto/gcm.cpp#L1790-L1808

In this computation, it assumes that the members form an acyclic graph (except for L's back-edges). This assumption does not hold for irreducible graphs, where there might be additional cycles corresponding to non-natural loops.  This could be refined by transforming the single forward pass into a proper system of equations to be solved iteratively. See some more details in the description field of the [bug report](https://bugs.openjdk.java.net/browse/JDK-8255763).

-------------

PR: https://git.openjdk.java.net/jdk16/pull/22