Hi,

Here is a candidate patch list from Intel for GCC 14.

This list ensures alignment of patches for inclusion in the Red Hat GCC 14 branch prior to its freeze.

Thx,

Lin

Patch list based on GCC14.2 (after GCC 14.2 release):

• r14-10550-g30f4fa3f53e4c1 i386: Add non-optimize prefetchi intrins
• r14-10551-ga295076bee293a Refine constraint "Bk" to define_special_memory_constraint.
• r14-10588-g4e7735a8d87559 Move ix86_align_loops into a separate pass and insert the pass after pass_endbr_and_patchable_area.
• r14-10591-g45a771d22e3090 i386: Fix some vex insns that prohibit egpr
• r14-10608-g27dc1533b6dfc4 Align ix86_{move_max,store_max} with vectorizer.
• r14-10625-gba9a3f105ea552 Check avx upper register for parallel.
• r14-10627-g59157c038d683e i386: Fix vfpclassph non-optimizied intrin
• r14-10658-g3951efed1cce97 doc: Enhance Intel CPU documentation
• r14-10672-g3f3f546bf830d0 x86-64: Don't use temp for argument in a TImode register
• r14-10689-ga3efd2ff9db054 doc: Add more alias option and reorder Intel CPU -march documentation
• r14-10716-g25cb153f93bb9f x86: Don't use address override with segment regsiter
• r14-10746-g1032b72548c47a x86: Disable stack protector for naked functions
• r14-10782-gfe0692f689a18c Add new microarchitecture tune for SRF/GRR/CWF.
• r14-10783-g9b7d5ecbecfbd1 Add a new tune avx256_avoid_vec_perm for SRF.
• r14-10807-g79e7e02b7cc578 Refine splitters related to "combine vpcmpuw + zero_extend to vpcmpuw"
• r14-10831-gb718f6ec1674c0 Fix ICE due to isa mismatch for the builtins.
• r14-10852-g71a0cf699b6a2d Fix ICE due to subreg:us_truncate.
• r14-10862-gdf542909224a7f [APX PPX] Avoid generating unmatched pushp/popp in pro/epilogue
• r14-10864-g82bfb6c5ba6d1f i386: Do not allow pointer conversion for CMPccXADD intrin under -O0
• r14-10895-g05fd99e3d5e9f0 i386: Add OPTION_MASK_ISA2_EVEX512 for some AVX512 instructions.
• r14-10896-g6a0e143a6449bc i386: Modify regexp of pr117304-1.c
• r14-10916-g4380d6f8acc878 i386: Add new model number for Arrow Lake 
• r14-10937-g8b4bb54e6c4541 i386: Zero extend 32-bit address to 64-bit with option -mx32 -maddress-mode=long. [PR 117418]
• r14-10979-g4a63cc6de77481 Fix uninitialized operands[2] in vec_unpacks_hi_v4sf.
• r14-11300-gcec0326137ef91 i386: Fix AVX512BW intrin header with _OPTIMIZE_ [PR 118813]
• r14-11301-g5f47dc6e9aa82e x86: Correct ASM_OUTPUT_SYMBOL_REF
• r14-11314-g31cbac836bb4f4 i386: Do not check vector size conflict when AVX512 is not explicitly set [PR 118815]
• r14-11315-gde562367d34475 i386: Deprecate -m[no-]avx10.1 and make -mno-avx10.1-512 to disable the whole AVX10.1
• r14-11345-gba488a332ad171 i386: Treat Granite Rapids/Granite Rapids-D similar as Sapphire Rapids in x86-tune.def
• r14-11443-g5ba6fdc5476d33 i386: Add -mavx10.1 back with 512 bit alias
• r14-11641-gd275b3748a23aa Extend check-function-bodies to allow label and directives
• r14-11642-ga33e2808c8987d APX: Don't use red-zone with 32 GPRs and no caller-saved registers
• r14-11643-g8cc672d3d3a2f0 x86: Update gcc.target/i386/apx-interrupt-1.c
• r14-11698-g96729baa9ef0e9 Remove other processors from X86_TUNE_DEST_FALSE_DEP_FOR_GLC except GLC