On Tue, Jan 28, 2014 at 6:45 PM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 28 January 2014 01:46, Peter Crosthwaite > <peter.crosthwa...@xilinx.com> wrote: >> There seem to be multiple instances in this series where you fallback >> to open coded R/W accessor functions for the sake of access checks. Is >> it better to define a bool check_access() fn hook in ARMCPRegInfo and >> leave the actual write/read behaviour to the data driven mechanisms? >> This may also minimise the need for raw_write hook usages as it serves >> to isolate the actual state change into its own self contained >> definition (whether open coded or not). > > Yes, I think it's probably going to be better to do that. We may need > to make it more than just bool, though since for AArch64 the > kind of exception can be different I think -- the specific syndrome > information can vary. >
I guess then it's simplest to just return in same format as read/write accessors. Regards, Peter > thanks > -- PMM >
