Debugger/Memory: Fix mem2 search

[malleoz]

Memory widget search functionality would not work for MEM2 region. Match() restricts our physical address from exceeding 0x01FFFFFF. This means that we can access cached/uncache MEM1 as our physical addresses are 0x00000000 - 0x017FFFFF. However, this logic does not work for MEM2, as our physical address is 0x10000000 to 0x117FFFFF (I believe). Searching starting from MEM2 does not show a match until 0xC0000000.

At first I assumed we would just modify memory_area, however, the call to MMU::HostIsRAMAddress() in EffectiveAddressSpaceAccessors::Search() appears to sufficiently check for valid memory bounds. Notably, we want to allow RAM searches for instances where you increase the size of MEM1 or MEM2 via Config->Advanced tab. In these instances, the current code does not allow for the upper bound to increase, whereas HostIsRAMAddress() eventually retrieves the m_memory.GetRamSizeReal() which factors in the possibly larger mem allocation.

The comment in Match() implies concern regarding the dcache. In testing, we either pagefault in MMU::IsRAMAddress() after the call to TranslateAddress() OR we reach EffectiveAddressSpaceAccessors::Matches(). In cases where we reach Matches(), we can just change the physical address check to require that we are not at or beyond dcache.

[malleoz]

Applied a wider change. Rather than just allow MEM2, which was the original intention, it turns out that we can remove the check here altogether, except for restricting dcache attempted accesses.

[sepalani]

Tested and the code LGTM.

This PR fixes MEM2 search using effective address space. It doesn't fix MEM2 search using physical address space, AFAICT.