diff options
| author | Alexei Starovoitov <ast@kernel.org> | 2024-04-01 16:38:00 -0700 |
|---|---|---|
| committer | Daniel Borkmann <daniel@iogearbox.net> | 2024-04-04 16:13:26 +0200 |
| commit | af682b767a41772499f8e54ca7d7e1deb3395f44 (patch) | |
| tree | 6d1b833f7a32bae58032ff46b09ffea035ef5147 /tools/testing/selftests/bpf/prog_tests/skb_load_bytes.c | |
| parent | 1e9e0b85255e6eca6036b59d8a5fbca6501905ac (diff) | |
bpf: Optimize emit_mov_imm64().
Turned out that bpf prog callback addresses, bpf prog addresses
used in bpf_trampoline, and in other cases the 64-bit address
can be represented as sign extended 32-bit value.
According to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82339
"Skylake has 0.64c throughput for mov r64, imm64, vs. 0.25 for mov r32, imm32."
So use shorter encoding and faster instruction when possible.
Special care is needed in jit_subprogs(), since bpf_pseudo_func()
instruction cannot change its size during the last step of JIT.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/CAADnVQKFfpY-QZBrOU2CG8v2du8Lgyb7MNVmOZVK_yTyOdNbBA@mail.gmail.com
Link: https://lore.kernel.org/bpf/20240401233800.42737-1-alexei.starovoitov@gmail.com
Diffstat (limited to 'tools/testing/selftests/bpf/prog_tests/skb_load_bytes.c')
0 files changed, 0 insertions, 0 deletions