[SYCL][FPGA] Update __builtin_intel_fpga_mem

[premanandrao]

This is a continuation of @Fznamznon's PR #5002. Most of the changes here are from there.

According to #4980 this built-in
should take four more constant integer parameters with following default
values:

const int32_t AnchorID = -1
const int32_t TargetAnchor = 0
const int32_t Type = 0
const int32_t Cycle = 0
The old variant is still allowed for backward compatibility. In case
some parameters are not provided - default value will be emitted in
annotation string.

In addition to the state of PR #5002, this PR makes minor changes. It incorporates @mlychkov's
code review comments there and modifies the CodeGenSYCL test to accept both opaque and
non-opaque pointers.

[premanandrao]

The test failure isn't related to this change.

[elizabethandrews]

Looks ok to me. Just one question. In your commit message you say - "the CodeGenSYCL test to accept both opaque and non-opaque pointers." What do you mean by this? I just see opaque pointer tests.

Also, from what I can tell, @Fznamznon's PR was not merged because spec wasn't ready and because we weren't sure if FPGA backend can handle this IR change. Are those concerns resolved? I assume backend will just ignore intrinsic it does not recognize if support has not been added yet.

[premanandrao]

Looks ok to me. Just one question. In your commit message you say - "the CodeGenSYCL test to accept both opaque and non-opaque pointers." What do you mean by this? I just see opaque pointer tests.

Thank you for this. I had missed adding the non-opaque version of the test.

Also, from what I can tell, @Fznamznon's PR was not merged because spec wasn't ready and because we weren't sure if FPGA backend can handle this IR change. Are those concerns resolved? I assume backend will just ignore intrinsic it does not recognize if support has not been added yet.

Yes, someone confirmed that it will be ignored.

[smanna12]

LGTM