Commit 3193c0836 ("bpf: Disable GCC -fgcse optimization for ___bpf_prog_run()") introduced a __no_fgcse macro that expands to a function scope __attribute__((optimize("-fno-gcse"))), to disable a GCC specific optimization that was causing trouble on x86 builds, and was not expected to have any positive effect in the first place. However, as the GCC manual documents, __attribute__((optimize)) is not for production use, and results in all other optimization options to be forgotten for the function in question. This can cause all kinds of trouble, but in one particular reported case, it causes -fno-asynchronous-unwind-tables to be disregarded, resulting in .eh_frame info to be emitted for the function. This reverts commit 3193c0836, and instead, it disables the -fgcse optimization for the entire source file, but only when building for X86 using GCC with CONFIG_BPF_JIT_ALWAYS_ON disabled. Note that the original commit states that CONFIG_RETPOLINE=n triggers the issue, whereas CONFIG_RETPOLINE=y performs better without the optimization, so it is kept disabled in both cases.
Unified · Split
| @@ -175,5 +175,3 @@ · 174 unmodified lines | |||
| 175 | #else | 175 | #else |
| 176 | #define __diag_GCC_8(s) | 176 | #define __diag_GCC_8(s) |
| 177 | #endif | 177 | #endif |
| 178 | |||
| 179 | #define __no_fgcse __attribute__((optimize("-fno-gcse"))) | ||
| @@ -247,10 +247,6 @@ struct ftrace_likely_data { · 246 unmodified lines | |||
| 247 | #define asm_inline asm | 247 | #define asm_inline asm |
| 248 | #endif | 248 | #endif |
| 249 | 249 | ||
| 250 | #ifndef __no_fgcse | ||
| 251 | # define __no_fgcse | ||
| 252 | #endif | ||
| 253 | |||
| 254 | /* Are two types/vars the same type (ignoring qualifiers)? */ | 250 | /* Are two types/vars the same type (ignoring qualifiers)? */ |
| 255 | #define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b)) | 251 | #define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b)) |
| 256 | 252 | ||
| @@ -1,6 +1,10 @@ | |||
| 1 | # SPDX-License-Identifier: GPL-2.0 | 1 | # SPDX-License-Identifier: GPL-2.0 |
| 2 | obj-y := core.o | 2 | obj-y := core.o |
| 3 | CFLAGS_core.o += $(call cc-disable-warning, override-init) | 3 | ifneq ($(CONFIG_BPF_JIT_ALWAYS_ON),y) |
| 4 | # ___bpf_prog_run() needs GCSE disabled on x86; see 3193c0836f203 for details | ||
| 5 | cflags-nogcse-$(CONFIG_X86)$(CONFIG_CC_IS_GCC) := -fno-gcse | ||
| 6 | endif | ||
| 7 | CFLAGS_core.o += $(call cc-disable-warning, override-init) $(cflags-nogcse-yy) | ||
| 4 | 8 | ||
| 5 | obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o bpf_iter.o map_iter.o task_iter.o prog_iter.o | 9 | obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o bpf_iter.o map_iter.o task_iter.o prog_iter.o |
| 6 | obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o | 10 | obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o |
| @@ -1369,7 +1369,7 @@ u64 __weak bpf_probe_read_kernel(void *dst, u32 size, const void *unsafe_ptr) · 1368 unmodified lines | |||
| 1369 | * | 1369 | * |
| 1370 | * Decode and execute eBPF instructions. | 1370 | * Decode and execute eBPF instructions. |
| 1371 | */ | 1371 | */ |
| 1372 | static u64 __no_fgcse ___bpf_prog_run(u64 *regs, const struct bpf_insn *insn, u64 *stack) | 1372 | static u64 ___bpf_prog_run(u64 *regs, const struct bpf_insn *insn, u64 *stack) |
| 1373 | { | 1373 | { |
| 1374 | #define BPF_INSN_2_LBL(x, y) [BPF_##x | BPF_##y] = &&x##_##y | 1374 | #define BPF_INSN_2_LBL(x, y) [BPF_##x | BPF_##y] = &&x##_##y |
| 1375 | #define BPF_INSN_3_LBL(x, y, z) [BPF_##x | BPF_##y | BPF_##z] = &&x##_##y##_##z | 1375 | #define BPF_INSN_3_LBL(x, y, z) [BPF_##x | BPF_##y | BPF_##z] = &&x##_##y##_##z |