JIT: Remove fgRelocateEHRegions

[amanasifkhalid]

Part of #107749. Block layout will now ensure cold EH regions are moved out of the hot part of the method, so fgRelocateEHRegions is no longer necessary. This isn't a no-diff change, though the changes look inconsequential; for the examples I looked at, the diffs stem from the relative position of cold try regions changing. For example, the baseline JIT will ensure cold try regions are at the very end of the method, like so:

---------------------------------------------------------------------------------------------------------------------------------------------------------------------
BBnum BBid ref try hnd preds           weight   IBC [IL range]   [jump]                            [EH region]        [flags]
---------------------------------------------------------------------------------------------------------------------------------------------------------------------
BB01 [0000]  1                             1    760 [000..017)-> BB03(1),BB02(0)         ( cond )                     i LIR IBC
BB03 [0009]  2       BB01,BB02             1    760 [000..020)-> BB08(1),BB04(0)         ( cond )                     i LIR IBC
BB08 [0005]  3       BB06,BB07,BB03        1    760 [052..053)                           (return)                     i LIR IBC
BB02 [0008]  1       BB01                  0      0 [000..001)-> BB03(1)                 (always)                     i LIR IBC rare hascall gcsafe
BB07 [0004]  1     0                       0.00   0 [04F..052)-> BB08(1)                 (always)    H0   catch { }   i LIR IBC keep hascall xentry
BB04 [0001]  1  0    BB03                  0      0 [020..02B)-> BB06(0.48),BB05(0.52)   ( cond ) T0      try {       i LIR IBC rare keep hascall
BB05 [0002]  1  0    BB04                  0      0 [02B..042)-> BB06(1)                 (always) T0                  i LIR IBC rare hascall newobj
BB06 [0003]  2  0    BB04,BB05             0      0 [042..04F)-> BB08(1)                 (always) T0      }           i LIR IBC rare hascall gcsafe nullcheck
---------------------------------------------------------------------------------------------------------------------------------------------------------------------

Now, the relative ordering of cold code doesn't change:

---------------------------------------------------------------------------------------------------------------------------------------------------------------------
BBnum BBid ref try hnd preds           weight   IBC [IL range]   [jump]                            [EH region]        [flags]
---------------------------------------------------------------------------------------------------------------------------------------------------------------------
BB01 [0000]  1                             1    760 [000..017)-> BB03(1),BB02(0)         ( cond )                     i LIR IBC
BB03 [0009]  2       BB01,BB02             1    760 [000..020)-> BB08(1),BB04(0)         ( cond )                     i LIR IBC
BB08 [0005]  3       BB06,BB07,BB03        1    760 [052..053)                           (return)                     i LIR IBC
BB02 [0008]  1       BB01                  0      0 [000..001)-> BB03(1)                 (always)                     i LIR IBC rare hascall gcsafe
BB04 [0001]  1  0    BB03                  0      0 [020..02B)-> BB06(0.48),BB05(0.52)   ( cond ) T0      try {       i LIR IBC rare keep hascall
BB05 [0002]  1  0    BB04                  0      0 [02B..042)-> BB06(1)                 (always) T0                  i LIR IBC rare hascall newobj
BB06 [0003]  2  0    BB04,BB05             0      0 [042..04F)-> BB08(1)                 (always) T0      }           i LIR IBC rare hascall gcsafe nullcheck
BB07 [0004]  1     0                       0.00   0 [04F..052)-> BB08(1)                 (always)    H0   catch { }   i LIR IBC keep hascall xentry
---------------------------------------------------------------------------------------------------------------------------------------------------------------------

[Copilot]

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

[amanasifkhalid]

@dotnet/jit-contrib PTAL. Diffs show almost no PerfScore diffs, suggesting the churn is mostly due to cold code moving around. Thanks!