RFR: 8272563: Possible assertion failure in CardTableBarrierSetC1

[Fairoz Matte]

On Wed, 18 Aug 2021 12:37:00 GMT, Fairoz Matte <fmatte at openjdk.org> wrote:

> This patch is proposed by the submitter of the bug - ugawa at ci.i.u-tokyo.ac.jp
> 
> The method CardTableBarrierSetC1::post_barrier generates a move LIR when TwoOperandLIRForm flag is true to move the address to be marked in the card table to a temporary register.
>> __ move(addr, tmp);
> However, this code only guarantees that `addr` is a valid register for LIR, which can be a virtual register. If the virtual register for `addr` is spilled to the stack by chance, the `move(addr, tmp)` is compiled to a memory-to-register which causes an assertion failure because a memory-to-register move requires their arguments to have the same size. 
> The fix is to check if it is is_oop() and call the mov appropriately.
> 
> No issues found in local testing and Mach5 tier1-3

Hi Igor, 

Thanks for looking into this and thanks for the pointers to handle the spills in logic involved in T_OBJECT to T_LONG mov's and Also to check the src and dst must have the registers.
These conditions can be handled using asserts/guarantee's (as the function doesn't return). 
But as original patch suggests we can have solution by copying over extra temp register. 

In https://github.com/openjdk/jdk/blob/d732c3091fea0a7c6d6767227de89002564504e5/src/hotspot/share/c1/c1_LinearScan.cpp#L1135 It was not handled and enfornced we must have both registers. So I am not sure if we also can handle this scenario or it can be prevented as done in c1_LinearScan.cpp file.

Thanks,

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

PR: https://git.openjdk.java.net/jdk/pull/5164