2020-03-28 16:46:18

by George Spelvin

Subject: [RFC PATCH v1 47/50] kernel/bpf/core.c: Use get_random_max32()

Smaller, faster, and a smidge more uniform than %.

Signed-off-by: George Spelvin <[email protected]>
Cc: Daniel Borkmann <[email protected]>
Cc: Alexei Starovoitov <[email protected]>
Cc: Martin KaFai Lau <[email protected]>
Cc: Song Liu <[email protected]>
Cc: Yonghong Song <[email protected]>
Cc: Andrii Nakryiko <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
kernel/bpf/core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index af6b738cf435c..61713a7f73d85 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -874,7 +874,7 @@ bpf_jit_binary_alloc(unsigned int proglen, u8 **image_ptr,
hdr->pages = pages;
hole = min_t(unsigned int, size - (proglen + sizeof(*hdr)),
PAGE_SIZE - sizeof(*hdr));
- start = (get_random_int() % hole) & ~(alignment - 1);
+ start = get_random_max32(hole) & ~(alignment - 1);

/* Leave a random number of instructions before BPF code. */
*image_ptr = &hdr->image[start];
@@ -947,7 +947,7 @@ static int bpf_jit_blind_insn(const struct bpf_insn *from,
bool emit_zext)
{
struct bpf_insn *to = to_buff;
- u32 imm_rnd = get_random_int();
+ u32 imm_rnd = get_random_u32();
s16 off;

BUILD_BUG_ON(BPF_REG_AX + 1 != MAX_BPF_JIT_REG);
--
2.26.0

2020-03-28 17:16:15

by Alexei Starovoitov

Subject: Re: [RFC PATCH v1 47/50] kernel/bpf/core.c: Use get_random_max32()

On Sat, Mar 28, 2020 at 9:46 AM George Spelvin <[email protected]> wrote:
>
> Smaller, faster, and a smidge more uniform than %.
>
> Signed-off-by: George Spelvin <[email protected]>
> Cc: Daniel Borkmann <[email protected]>
> Cc: Alexei Starovoitov <[email protected]>
> Cc: Martin KaFai Lau <[email protected]>
> Cc: Song Liu <[email protected]>
> Cc: Yonghong Song <[email protected]>
> Cc: Andrii Nakryiko <[email protected]>
> Cc: [email protected]
> Cc: [email protected]

If Linus is going to pick up the whole set at once:
Acked-by: Alexei Starovoitov <[email protected]>