On Sat, Aug 12, 2017 at 8:41 AM, Richard Henderson wrote:
> SetExclusiveMonitors in the pseudocode is on the address + width,
> and says nothing about the manner of the load. Therefore
>
> ldxpw0, w1, [x2]
> vs
> ldxrx0, [x2]
>
> must record the same
On 08/12/2017 09:29 AM, Alistair Francis wrote:
> Now we aren't ever checking cpu_exclusive_high. Is it even worth having?
We are checking cpu_exclusive_high for 64-bit STXP.
See paired_cmpxchg64_{l,b}e in target/arm/helper-a64.c.
r~
On Sat, Aug 12, 2017 at 8:41 AM, Richard Henderson wrote:
> SetExclusiveMonitors in the pseudocode is on the address + width,
> and says nothing about the manner of the load. Therefore
>
> ldxpw0, w1, [x2]
> vs
> ldxrx0, [x2]
>
> must record the same
SetExclusiveMonitors in the pseudocode is on the address + width,
and says nothing about the manner of the load. Therefore
ldxpw0, w1, [x2]
vs
ldxrx0, [x2]
must record the same metadata so that either may pair with
stxpw3, w0, w1, [x2]
vs
stxrw3,