gcc: xtensa: Backport patches from upstream/master (v3)

[jjsuwa-sys3175]

See #29, #36 and #37.

List of patches added after #37:

• 461d3c84a0e5ad045ee54631901cc953d6befa20 "xtensa: fix PR target/108919"
• ce83c3e492c2fa5a08c15b5f4619d58f42a5dcd0 "xtensa: Make use of CLAMPS instruction if configured"
• c981f61c071757f1f724fe7c3959622c13f079fa "xtensa: Remove REG_OK_STRICT and its derivatives"
• e33d2dcb463161a110ac345a451132ce8b2b23d9 "xtensa: Optimize '(x & CST1_POW2) != 0 ? CST2_POW2 : 0'"
• 8a20b4bc50bdc8d61610974d60d5851f3fd8b70f "xtensa: Merge 'addx' and 'subx' insn patterns into one"
• bf78e24a90d4d064b0372a03c0327b6f90475949 "xtensa: tidy extzvsi-1bit patterns"
• 9b867c8281ee313cf6ec737d8f4a9ba7ef78408e "xtensa: Add 'subtraction from constant' insn pattern"
• 78648c09a0494e6bf1edbd97acb708bc282eb5f5 "xtensa: Rework 'setmemsi' insn pattern"
• feae4e83f76cf8e6698429d5acb08ff1968d13f2 "xtensa: Improve "*shlrd_reg" insn pattern and its variant"
• fe3ce0861081dd17e581c32b299b9c743d000470 "xtensa: Add 'adddi3' and 'subdi3' insn patterns"
• 830d36b3c307c70af57b832821d8590b29a5bda5 "xtensa: Optimize boolean evaluation or branching when EQ/NE to INT_MIN"
• 1ce54ad8cd694a1defb9374f18607194ef702ea7 "xtensa: Remove TARGET_MEMORY_MOVE_COST hook"
• 7360cba833cd921631818428a851e03ea88f1e8a "xtensa: constantsynth: Add new 2-insns synthesis pattern"
• 0778f4dcdafa7dabb015a3375089addc1a4df49b "xtensa: Fix missing mode warning in "*eqne_INT_MIN""
• 2f615b33dd61590a74e7758e19016250ade2b828 "xtensa: The use of CLAMPS instruction also requires TARGET_MINMAX, as well as TARGET_CLAMPS"
• cd22b97726472138d3fe22fb1ff0c27176408fc9 "xtensa: Use HARD_REG_SET instead of bare integer"

[earlephilhower]

Thanks! Let me know when it's worth trying a new build.

Also, what is the impact of these changes? Are we expecting to see a noticeable code reduction or speed improvement. Or are these actual bug fixes?

[jjsuwa-sys3175]

For now, it's a miscellaneous collection of low-impact patches and no critical bug fixes, so I don't think there's any particular reason to rush the build.