Fix #74020: Optimize consecutive shifts in JIT Lowering

[csa7mdm]

Description

This PR addresses issue #74020 by implementing a peephole optimization in JIT Lowering to combine consecutive right shift operations (GT_RSH/GT_RSZ) with constant shift amounts.

Fix

The optimization logic is added to Lowering::LowerShift and handles:

• Nested shifts of the same operation type: (x >> c1) >> c2 -> x >> (c1 + c2)
• Nested shifts separated by a cast (common in division lowering): (cast (x >> c1)) >> c2 -> cast (x >> (c1 + c2))

Verification

• Verified against the scenario reported in the issue where repeated division by constants generates redundant shift instructions.
• Ensure combined shift count does not exceed type width.

[tannergooding]

LGTM. I do still think this would be better to handle in morph and that it would be good to track RSH(x, n) for 0 < n < bitWidth as being "never negative", but will defer to other input.

CC. @dotnet/jit-contrib for secondary review

[JulieLeeMSFT]

LGTM. I do still think this would be better to handle in morph and that it would be good to track RSH(x, n) for 0 < n < bitWidth as being "never negative", but will defer to other input.

@csa7mdm, I also think it would be better to handle it in morph. Do you want to try that?

[csa7mdm]

n workflow(s) awaiting approva

[csa7mdm]

n workflow(s) awaiting approval

[JulieLeeMSFT]

Approved workflow to run.

[adamperlin]

@csa7mdm it looks like the latest CI run failed if you want to take a look?