GDB Protocol is not handled correctly

[jounathaen]

When I run cargo test gdb on the main branch and add a print to the read_addrs function,

uhyve/src/linux/gdb/mod.rs

Line 121 in dc6b01e

fn read_addrs(&mut self, start_addr: u64, data: &mut [u8]) -> TargetResult<usize, Self> {

I get the following reads:

[...]
>>>  reading start_addr 0x59d478
>>>  reading start_addr 0x64412074656e726d
>>>  reading start_addr 0x64412074656e7265
>>>  reading start_addr 0x64412074656e7265
>>>  reading start_addr 0x64412074656e726d
>>>  reading start_addr 0x3f8000
>>>  reading start_addr 0x0
>>>  reading start_addr 0x99d478
>>>  reading start_addr 0x99d478
>>>  reading start_addr 0x99d478
>>>  reading start_addr 0x649a84
[...]

0x64412074656e726d is definitely not a valid virtual address on x86.
Currently, this fails silently, as the virt_to_phys() maps this address to the address 26d.

-> We should check, whether this is valid memory and return an Error if not (read_addr API as reference: https://docs.rs/gdbstub/latest/gdbstub/target/ext/base/singlethread/trait.SingleThreadBase.html#tymethod.read_addrs)

[jounathaen]

This is partly fixed in #527, where this access returns an error to gdbstub.
However, I'm still not sure why this address is accessed at all?