diff --git a/Makefile b/Makefile
--- a/Makefile
+++ b/Makefile
@@ -467,7 +467,7 @@ endif
 export KERNELDOC
 
 KBUILD_USERHOSTCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \
-			 -O2 -fomit-frame-pointer -std=gnu11
+			 -O3 -fomit-frame-pointer -std=gnu11
 KBUILD_USERCFLAGS  := $(KBUILD_USERHOSTCFLAGS) $(USERCFLAGS)
 KBUILD_USERLDFLAGS := $(USERLDFLAGS)

@@ -506,7 +506,7 @@ endif
 
 KBUILD_HOSTCFLAGS   := $(KBUILD_USERHOSTCFLAGS) $(HOST_LFS_CFLAGS) \
 		       $(HOSTCFLAGS) -I $(srctree)/scripts/include
-KBUILD_HOSTCXXFLAGS := -Wall -O2 $(HOST_LFS_CFLAGS) $(HOSTCXXFLAGS) \
+KBUILD_HOSTCXXFLAGS := -Wall -O3 $(HOST_LFS_CFLAGS) $(HOSTCXXFLAGS) \
 		       -I $(srctree)/scripts/include
 KBUILD_HOSTRUSTFLAGS := $(rust_common_flags) -O -Cstrip=debuginfo \
 			-Zallow-features=
@@ -922,19 +922,19 @@ KBUILD_CFLAGS       += $(call cc-disable-warning, format-overflow)
 KBUILD_CFLAGS	+= -fno-delete-null-pointer-checks
 
 ifdef CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE
-KBUILD_CFLAGS += -O2
+KBUILD_CFLAGS += -O3
 # LLVM's Polly for high-level loop and data-locality optimizer based on integer
 # polyhedra to analyze and optimize the memory access pattern.
 KBUILD_CFLAGS += $(call cc-option,-mllvm=-polly -mllvm=-polly-ast-use-context \
 		 -mllvm=-polly-invariant-load-hoisting -mllvm=-polly-loopfusion-greedy \
 		 -mllvm=-polly-run-inliner -mllvm=-polly-vectorizer=stripmine)
 ifdef CONFIG_LD_DEAD_CODE_DATA_ELIMINATION
 KBUILD_CFLAGS += $(call cc-option,-mllvm=-polly-run-dce)
 endif
 # Enable software pipelining (Swing Modulo Scheduling) for innermost loops.
 KBUILD_CFLAGS += $(call cc-option,-fmodulo-sched -fmodulo-sched-allow-regmoves) \
 		 $(call cc-option,-mllvm=-enable-pipeliner)
-KBUILD_RUSTFLAGS += -Copt-level=2
+KBUILD_RUSTFLAGS += -Copt-level=3
 else ifdef CONFIG_CC_OPTIMIZE_FOR_SIZE
 KBUILD_CFLAGS += -Os
 KBUILD_CFLAGS += $(call cc-option,-mllvm=-enable-pipeliner-opt-size)
diff --git a/init/Kconfig b/init/Kconfig
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -1613,10 +1613,10 @@ choice
 	default CC_OPTIMIZE_FOR_PERFORMANCE

 config CC_OPTIMIZE_FOR_PERFORMANCE
-	bool "Optimize for performance (-O2)"
+	bool "Optimize for performance (-O3)"
 	help
 	  This is the default optimization level for the kernel, building
-	  with the "-O2" compiler flag for best performance and most
+	  with the "-O3" compiler flag for best performance and most
 	  helpful compile-time warnings.

 config CC_OPTIMIZE_FOR_SIZE