From: Alexei Starovoitov <alexei.starovoitov@gmail.com>
To: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Cc: bpf@vger.kernel.org, "Alexei Starovoitov" <ast@kernel.org>,
"Daniel Borkmann" <daniel@iogearbox.net>,
"Andrii Nakryiko" <andrii@kernel.org>,
"Martin KaFai Lau" <kafai@fb.com>,
"Song Liu" <songliubraving@fb.com>, "Yonghong Song" <yhs@fb.com>,
"Jesper Dangaard Brouer" <brouer@redhat.com>,
"Toke Høiland-Jørgensen" <toke@redhat.com>,
netdev@vger.kernel.org
Subject: Re: [PATCH bpf-next v5 3/8] libbpf: Add weak ksym support to gen_loader
Date: Thu, 28 Oct 2021 17:22:44 -0700 [thread overview]
Message-ID: <20211029002244.3m63qmwrmykarvt6@ast-mbp.dhcp.thefacebook.com> (raw)
In-Reply-To: <20211028063501.2239335-4-memxor@gmail.com>
On Thu, Oct 28, 2021 at 12:04:56PM +0530, Kumar Kartikeya Dwivedi wrote:
> This extends existing ksym relocation code to also support relocating
> weak ksyms. Care needs to be taken to zero out the src_reg (currently
> BPF_PSEUOD_BTF_ID, always set for gen_loader by bpf_object__relocate_data)
> when the BTF ID lookup fails at runtime. This is not a problem for
> libbpf as it only sets ext->is_set when BTF ID lookup succeeds (and only
> proceeds in case of failure if ext->is_weak, leading to src_reg
> remaining as 0 for weak unresolved ksym).
>
> A pattern similar to emit_relo_kfunc_btf is followed of first storing
> the default values and then jumping over actual stores in case of an
> error. For src_reg adjustment, we also need to perform it when copying
> the populated instruction, so depending on if copied insn[0].imm is 0 or
> not, we decide to jump over the adjustment.
>
> We cannot reach that point unless the ksym was weak and resolved and
> zeroed out, as the emit_check_err will cause us to jump to cleanup
> label, so we do not need to recheck whether the ksym is weak before
> doing the adjustment after copying BTF ID and BTF FD.
>
> This is consistent with how libbpf relocates weak ksym. Logging
> statements are added to show the relocation result and aid debugging.
>
> Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
> ---
> tools/lib/bpf/gen_loader.c | 35 ++++++++++++++++++++++++++++++++---
> 1 file changed, 32 insertions(+), 3 deletions(-)
>
> diff --git a/tools/lib/bpf/gen_loader.c b/tools/lib/bpf/gen_loader.c
> index 11172a868180..1c404752e565 100644
> --- a/tools/lib/bpf/gen_loader.c
> +++ b/tools/lib/bpf/gen_loader.c
> @@ -13,6 +13,7 @@
> #include "hashmap.h"
> #include "bpf_gen_internal.h"
> #include "skel_internal.h"
> +#include <asm/byteorder.h>
>
> #define MAX_USED_MAPS 64
> #define MAX_USED_PROGS 32
> @@ -776,12 +777,24 @@ static void emit_relo_ksym_typeless(struct bpf_gen *gen,
> emit_ksym_relo_log(gen, relo, kdesc->ref);
> }
>
> +static __u32 src_reg_mask(void)
> +{
> +#if defined(__LITTLE_ENDIAN_BITFIELD)
> + return 0x0f; /* src_reg,dst_reg,... */
> +#elif defined(__BIG_ENDIAN_BITFIELD)
> + return 0xf0; /* dst_reg,src_reg,... */
> +#else
> +#error "Unsupported bit endianness, cannot proceed"
> +#endif
> +}
> +
> /* Expects:
> * BPF_REG_8 - pointer to instruction
> */
> static void emit_relo_ksym_btf(struct bpf_gen *gen, struct ksym_relo_desc *relo, int insn)
> {
> struct ksym_desc *kdesc;
> + __u32 reg_mask;
>
> kdesc = get_ksym_desc(gen, relo);
> if (!kdesc)
> @@ -792,19 +805,35 @@ static void emit_relo_ksym_btf(struct bpf_gen *gen, struct ksym_relo_desc *relo,
> kdesc->insn + offsetof(struct bpf_insn, imm));
> move_blob2blob(gen, insn + sizeof(struct bpf_insn) + offsetof(struct bpf_insn, imm), 4,
> kdesc->insn + sizeof(struct bpf_insn) + offsetof(struct bpf_insn, imm));
> - goto log;
> + emit(gen, BPF_LDX_MEM(BPF_W, BPF_REG_9, BPF_REG_8, offsetof(struct bpf_insn, imm)));
Thanks a lot for working on this. I've applied the set.
The above load is redundant, right? BPF_REG_0 already has that value
and could have been used in the JNE below, right?
> + /* jump over src_reg adjustment if imm is not 0 */
> + emit(gen, BPF_JMP_IMM(BPF_JNE, BPF_REG_9, 0, 3));
> + goto clear_src_reg;
Is there a test for this part of the code?
It's only for weak && unresolved && multi referenced ksym, right?
Or bpf_link_fops2 test_ksyms_weak.c fits this category?
> }
> /* remember insn offset, so we can copy BTF ID and FD later */
> kdesc->insn = insn;
> emit_bpf_find_by_name_kind(gen, relo);
> - emit_check_err(gen);
> + if (!relo->is_weak)
> + emit_check_err(gen);
> + /* set default values as 0 */
> + emit(gen, BPF_ST_MEM(BPF_W, BPF_REG_8, offsetof(struct bpf_insn, imm), 0));
> + emit(gen, BPF_ST_MEM(BPF_W, BPF_REG_8, sizeof(struct bpf_insn) + offsetof(struct bpf_insn, imm), 0));
> + /* skip success case stores if ret < 0 */
> + emit(gen, BPF_JMP_IMM(BPF_JSLT, BPF_REG_7, 0, 4));
> /* store btf_id into insn[insn_idx].imm */
> emit(gen, BPF_STX_MEM(BPF_W, BPF_REG_8, BPF_REG_7, offsetof(struct bpf_insn, imm)));
> /* store btf_obj_fd into insn[insn_idx + 1].imm */
> emit(gen, BPF_ALU64_IMM(BPF_RSH, BPF_REG_7, 32));
> emit(gen, BPF_STX_MEM(BPF_W, BPF_REG_8, BPF_REG_7,
> sizeof(struct bpf_insn) + offsetof(struct bpf_insn, imm)));
The double store (first with zeros and then with real values) doesn't look pretty.
I think an extra jump over two stores would have been cleaner.
next prev parent reply other threads:[~2021-10-29 0:22 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-28 6:34 [PATCH bpf-next v5 0/8] Typeless/weak ksym for gen_loader + misc fixups Kumar Kartikeya Dwivedi
2021-10-28 6:34 ` [PATCH bpf-next v5 1/8] bpf: Add bpf_kallsyms_lookup_name helper Kumar Kartikeya Dwivedi
2021-10-28 18:10 ` Andrii Nakryiko
2021-10-28 6:34 ` [PATCH bpf-next v5 2/8] libbpf: Add typeless ksym support to gen_loader Kumar Kartikeya Dwivedi
2021-10-28 6:34 ` [PATCH bpf-next v5 3/8] libbpf: Add weak " Kumar Kartikeya Dwivedi
2021-10-29 0:22 ` Alexei Starovoitov [this message]
2021-10-30 15:12 ` Kumar Kartikeya Dwivedi
2021-10-28 6:34 ` [PATCH bpf-next v5 4/8] libbpf: Ensure that BPF syscall fds are never 0, 1, or 2 Kumar Kartikeya Dwivedi
2021-10-28 18:00 ` Andrii Nakryiko
2021-10-28 6:34 ` [PATCH bpf-next v5 5/8] libbpf: Use O_CLOEXEC uniformly when opening fds Kumar Kartikeya Dwivedi
2021-10-28 18:04 ` Andrii Nakryiko
2021-10-28 6:34 ` [PATCH bpf-next v5 6/8] selftests/bpf: Add weak/typeless ksym test for light skeleton Kumar Kartikeya Dwivedi
2021-10-28 18:06 ` Andrii Nakryiko
2021-10-28 6:35 ` [PATCH bpf-next v5 7/8] selftests/bpf: Fix fd cleanup in sk_lookup test Kumar Kartikeya Dwivedi
2021-10-28 6:35 ` [PATCH bpf-next v5 8/8] selftests/bpf: Fix memory leak in test_ima Kumar Kartikeya Dwivedi
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=20211029002244.3m63qmwrmykarvt6@ast-mbp.dhcp.thefacebook.com \
--to=alexei.starovoitov@gmail.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=brouer@redhat.com \
--cc=daniel@iogearbox.net \
--cc=kafai@fb.com \
--cc=memxor@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=songliubraving@fb.com \
--cc=toke@redhat.com \
--cc=yhs@fb.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;
as well as URLs for NNTP newsgroup(s).