From: Yonghong Song <yonghong.song@linux.dev>
To: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc: Andrii Nakryiko <andrii.nakryiko@gmail.com>,
Anton Protopopov <aspsk@isovalent.com>,
Alexei Starovoitov <ast@kernel.org>,
Andrii Nakryiko <andrii@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Jiri Olsa <jolsa@kernel.org>,
Martin KaFai Lau <martin.lau@linux.dev>,
Stanislav Fomichev <sdf@google.com>, bpf <bpf@vger.kernel.org>
Subject: Re: [PATCH bpf-next 6/7] libbpf: BPF Static Keys support
Date: Fri, 8 Dec 2023 21:04:50 -0800 [thread overview]
Message-ID: <85aa91f9-d5c0-4e7b-950d-475da7787f64@linux.dev> (raw)
In-Reply-To: <CAADnVQJd1aUFzznLhwNvkN+zot-u3=4A16utY93HoLJrP_vo3w@mail.gmail.com>
On 12/8/23 8:25 PM, Alexei Starovoitov wrote:
> On Fri, Dec 8, 2023 at 8:15 PM Yonghong Song <yonghong.song@linux.dev> wrote:
>>
>> On 12/8/23 8:05 PM, Alexei Starovoitov wrote:
>>> On Fri, Dec 8, 2023 at 2:04 PM Andrii Nakryiko
>>> <andrii.nakryiko@gmail.com> wrote:
>>>> I feel like embedding some sort of ID inside the instruction is very..
>>>> unusual, shall we say?
>>> yeah. no magic numbers inside insns pls.
>>>
>>> I don't like JA_CFG name, since I read CFG as control flow graph,
>>> while you probably meant CFG as configurable.
>>> How about BPF_JA_OR_NOP ?
>>> Then in combination with BPF_JMP or BPF_JMP32 modifier
>>> the insn->off|imm will be used.
>>> 1st bit in src_reg can indicate the default action: nop or jmp.
>>> In asm it may look like asm("goto_or_nop +5")
>> How does the C source code looks like in order to generate
>> BPF_JA_OR_NOP insn? Any source examples?
> It will be in inline asm only. The address of that insn will
> be taken either via && or via asm (".long %l[label]").
> From llvm pov both should go through the same relo creation logic. I hope :)
A hack in llvm below with an example, could you check whether the C
syntax and object dump result
is what you want to see?
diff --git a/llvm/lib/Target/BPF/AsmParser/BPFAsmParser.cpp
b/llvm/lib/Target/BPF/AsmParser/BPFAsmParser.cpp
index 90697c6645be..38b1cbc31f9a 100644
--- a/llvm/lib/Target/BPF/AsmParser/BPFAsmParser.cpp
+++ b/llvm/lib/Target/BPF/AsmParser/BPFAsmParser.cpp
@@ -231,6 +231,7 @@ public:
.Case("call", true)
.Case("goto", true)
.Case("gotol", true)
+ .Case("goto_or_nop", true)
.Case("*", true)
.Case("exit", true)
.Case("lock", true)
@@ -259,6 +260,7 @@ public:
.Case("bswap64", true)
.Case("goto", true)
.Case("gotol", true)
+ .Case("goto_or_nop", true)
.Case("ll", true)
.Case("skb", true)
.Case("s", true)
diff --git a/llvm/lib/Target/BPF/BPFInstrInfo.td
b/llvm/lib/Target/BPF/BPFInstrInfo.td
index 5972c9d49c51..a953d10429bf 100644
--- a/llvm/lib/Target/BPF/BPFInstrInfo.td
+++ b/llvm/lib/Target/BPF/BPFInstrInfo.td
@@ -592,6 +592,19 @@ class BRANCH<BPFJumpOp Opc, string OpcodeStr,
list<dag> Pattern>
let BPFClass = BPF_JMP;
}
+class BRANCH_OR_NOP<BPFJumpOp Opc, string OpcodeStr, list<dag> Pattern>
+ : TYPE_ALU_JMP<Opc.Value, BPF_K.Value,
+ (outs),
+ (ins brtarget:$BrDst),
+ !strconcat(OpcodeStr, " $BrDst"),
+ Pattern> {
+ bits<16> BrDst;
+
+ let Inst{47-32} = BrDst;
+ let Inst{31-0} = 1;
+ let BPFClass = BPF_JMP;
+}
+
class BRANCH_LONG<BPFJumpOp Opc, string OpcodeStr, list<dag> Pattern>
: TYPE_ALU_JMP<Opc.Value, BPF_K.Value,
(outs),
@@ -632,6 +645,7 @@ class CALLX<string OpcodeStr>
let isBranch = 1, isTerminator = 1, hasDelaySlot=0, isBarrier = 1 in {
def JMP : BRANCH<BPF_JA, "goto", [(br bb:$BrDst)]>;
def JMPL : BRANCH_LONG<BPF_JA, "gotol", []>;
+ def JMP_OR_NOP : BRANCH_OR_NOP<BPF_JA, "goto_or_nop", []>;
}
// Jump and link
And an example,
[ ~/tmp1/gotol]$ cat t.c
int bar(void);
int foo()
{
int a, b;
asm volatile goto ("r0 = 0; \
goto_or_nop %l[label]; \
r2 = 2; \
r3 = 3; \
"::::label);
a = bar();
label:
b = 20 * a;
return b;
}
[ ~/tmp1/gotol]$ clang --target=bpf -O2 -S t.c
[ ~/tmp1/gotol]$ cat t.s
.text
.file "t.c"
.globl foo # -- Begin function foo
.p2align 3
.type foo,@function
foo: # @foo
# %bb.0: # %entry
r0 = 0
#APP
r0 = 0
goto_or_nop LBB0_2
r2 = 2
r3 = 3
#NO_APP
# %bb.1: # %asm.fallthrough
call bar
r0 *= 20
LBB0_2: # Block address taken
# %label
# Label of block must be emitted
exit
.Lfunc_end0:
.size foo, .Lfunc_end0-foo
# -- End function
.addrsig
[ ~/tmp1/gotol]$ clang --target=bpf -O2 -c t.c
[ ~/tmp1/gotol]$ llvm-objdump -dr t.o
t.o: file format elf64-bpf
Disassembly of section .text:
0000000000000000 <foo>:
0: b7 00 00 00 00 00 00 00 r0 = 0x0
1: b7 00 00 00 00 00 00 00 r0 = 0x0
2: 05 00 04 00 01 00 00 00 goto_or_nop +0x4 <LBB0_2>
3: b7 02 00 00 02 00 00 00 r2 = 0x2
4: b7 03 00 00 03 00 00 00 r3 = 0x3
5: 85 10 00 00 ff ff ff ff call -0x1
0000000000000028: R_BPF_64_32 bar
6: 27 00 00 00 14 00 00 00 r0 *= 0x14
0000000000000038 <LBB0_2>:
7: 95 00 00 00 00 00 00 00 exit
[ ~/tmp1/gotol]$
next prev parent reply other threads:[~2023-12-09 5:05 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-06 14:10 [PATCH bpf-next 0/7] BPF Static Keys Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 1/7] bpf: extract bpf_prog_bind_map logic into an inline helper Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 2/7] bpf: rename and export a struct definition Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 3/7] bpf: adjust functions offsets when patching progs Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 4/7] bpf: implement BPF Static Keys support Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 5/7] bpf: x86: implement static keys support Anton Protopopov
2023-12-06 14:10 ` [PATCH bpf-next 6/7] libbpf: BPF Static Keys support Anton Protopopov
2023-12-08 3:45 ` Alexei Starovoitov
2023-12-08 16:19 ` Anton Protopopov
2023-12-08 22:04 ` Andrii Nakryiko
2023-12-08 23:07 ` Eduard Zingerman
2023-12-09 4:07 ` Alexei Starovoitov
2023-12-09 4:05 ` Alexei Starovoitov
2023-12-09 4:15 ` Yonghong Song
2023-12-09 4:25 ` Alexei Starovoitov
2023-12-09 5:04 ` Yonghong Song [this message]
2023-12-09 17:18 ` Alexei Starovoitov
2023-12-10 6:32 ` Yonghong Song
2023-12-10 10:30 ` Eduard Zingerman
2023-12-11 3:33 ` Alexei Starovoitov
2023-12-11 18:49 ` Andrii Nakryiko
2023-12-12 10:25 ` Anton Protopopov
2023-12-14 2:15 ` Alexei Starovoitov
2023-12-14 3:04 ` Yonghong Song
2023-12-14 16:56 ` Eduard Zingerman
2023-12-14 16:33 ` Anton Protopopov
2023-12-11 17:31 ` Anton Protopopov
2023-12-11 19:08 ` Alexei Starovoitov
2023-12-12 9:06 ` Anton Protopopov
2023-12-11 21:51 ` Yonghong Song
2023-12-11 22:52 ` Yonghong Song
2023-12-06 14:10 ` [PATCH bpf-next 7/7] selftests/bpf: Add tests for BPF Static Keys Anton Protopopov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=85aa91f9-d5c0-4e7b-950d-475da7787f64@linux.dev \
--to=yonghong.song@linux.dev \
--cc=alexei.starovoitov@gmail.com \
--cc=andrii.nakryiko@gmail.com \
--cc=andrii@kernel.org \
--cc=aspsk@isovalent.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=jolsa@kernel.org \
--cc=martin.lau@linux.dev \
--cc=sdf@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox