* Parse floats the way mainline gdb prints them.

For 64-bit floats, it shows both float and double results. Now more
tests pass using mainline gdb.

* Disable ANSI when talking to gdb.

Helps more tests pass with mainline gdb.

* Parse {float=...,double=...} in "info registers"

Makes tests work better with mainline gdb.

Also work with the new command line options that were renamed in
https://github.com/riscv/riscv-isa-sim/pull/299

Passes on spike and Arty. Won't merge until
https://github.com/riscv/riscv-openocd/pull/364 merges.

This test confirms that in SMP configurations OpenOCD halts the harts
near-simulatenously. (It'll also check for resume, but that's not
implemented yet so commented out for now.)

* WIP

* Use hwthread everywhere.

* Test `-rtos hwthread`.

Also tweak timeouts a bit so that we don't have ridiculous timeouts for
simple operations.

* Tweak timeouts so tests pass on a loaded system.

Add testing of run-test/idle cases.

Fix MemTestBlock

This test used to false pass on 64-bit targets because gdb doesn't like
using Intel hex files on 64-bit targets.

Increase some timeouts in case memory access is slow.

Use more than 1KB for download test.

This is a pretty old bug. I limited the size to 256KB because against
spike my machine gets about 8KB/s, and I don't want to wait forever for
`make` to pass.

Tweak debug tests to run out of flash.

Added MemTestBlockReadInvalid verifying the corresponding OpenOCD fix

TriggerDmode, ProgramHwWatchpoint, ProgramSwWatchpoint.

Only TriggerDmode still fails.

Not all tests pass when run out of flash yet, but it's getting a lot
closer. The ones still failing on HiFive1-flash are: DebugSymbols,
Hwbp2, InstantHaltTest, TriggerDmode, TriggerLoadAddressInstant, and
TriggerStoreAddressInstant.

Gdb and OpenOCD were fixed so we don't have to accept broken behavior
anymore.