Richard Henderson <richard.henderson@linaro.org> writes:

> For semi-hosting, it seems even better if the semi-hosting syscall instruction
> is not "real", because you're explicitly requesting services from "unreal"
> hardware.  It should be specified to generate a SIGILL type of exception
> anywhere semi-hosting is not enabled.

In the QEMU case, yes, it's virtual hardware. However, the other common case
for semihosting is when doing hardware bringup using JTAG or other
remote debugging link -- having an instruction which safely traps to the
debugger is required to make semihosting usable there. Reading through
the history of the current RISC-V semihosting mechanism, there were many
designs considered and rejected because they wouldn't work in the JTAG
environment.

> With that in mind, it may be simpler to handle all of this not in the
> translator, but in the function that delivers the ebreak exception.  At that
> point one can arrange to read memory without raising additional
> exceptions.

I'll go explore and see if I can figure any of this out.

I'd still like to get the non-RISC-V SYS_READC patch landed someday :-)

-- 
-keith