From: Yonghong Song <yhs@fb.com>
To: Daniel Xu <dxu@dxuuu.xyz>,
"bpf@vger.kernel.org" <bpf@vger.kernel.org>,
"ast@kernel.org" <ast@kernel.org>,
"daniel@iogearbox.net" <daniel@iogearbox.net>,
Song Liu <songliubraving@fb.com>,
Andrii Nakryiko <andriin@fb.com>
Cc: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Kernel Team <Kernel-team@fb.com>,
"peterz@infradead.org" <peterz@infradead.org>,
"mingo@redhat.com" <mingo@redhat.com>,
"acme@kernel.org" <acme@kernel.org>
Subject: Re: [PATCH v3 bpf-next 1/3] bpf: Add bpf_perf_prog_read_branches() helper
Date: Thu, 23 Jan 2020 23:48:55 +0000 [thread overview]
Message-ID: <5f8e2ffd-9368-cb0c-9591-c6b7ecb8edb0@fb.com> (raw)
In-Reply-To: <20200123212312.3963-2-dxu@dxuuu.xyz>
On 1/23/20 1:23 PM, Daniel Xu wrote:
> Branch records are a CPU feature that can be configured to record
> certain branches that are taken during code execution. This data is
> particularly interesting for profile guided optimizations. perf has had
> branch record support for a while but the data collection can be a bit
> coarse grained.
>
> We (Facebook) have seen in experiments that associating metadata with
> branch records can improve results (after postprocessing). We generally
> use bpf_probe_read_*() to get metadata out of userspace. That's why bpf
> support for branch records is useful.
>
> Aside from this particular use case, having branch data available to bpf
> progs can be useful to get stack traces out of userspace applications
> that omit frame pointers.
>
> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
> ---
> include/uapi/linux/bpf.h | 15 ++++++++++++++-
> kernel/trace/bpf_trace.c | 31 +++++++++++++++++++++++++++++++
> 2 files changed, 45 insertions(+), 1 deletion(-)
>
> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> index f1d74a2bd234..50c580c8a201 100644
> --- a/include/uapi/linux/bpf.h
> +++ b/include/uapi/linux/bpf.h
> @@ -2892,6 +2892,18 @@ union bpf_attr {
> * Obtain the 64bit jiffies
> * Return
> * The 64 bit jiffies
> + *
> + * int bpf_perf_prog_read_branches(struct bpf_perf_event_data *ctx, void *buf, u32 buf_size)
> + * Description
> + * For en eBPF program attached to a perf event, retrieve the
en => an
> + * branch records (struct perf_branch_entry) associated to *ctx*
> + * and store it in the buffer pointed by *buf* up to size
> + * *buf_size* bytes.
> + *
> + * Any unused parts of *buf* will be filled with zeros.
> + * Return
> + * On success, number of bytes written to *buf*. On error, a
> + * negative value.
> */
> #define __BPF_FUNC_MAPPER(FN) \
> FN(unspec), \
> @@ -3012,7 +3024,8 @@ union bpf_attr {
> FN(probe_read_kernel_str), \
> FN(tcp_send_ack), \
> FN(send_signal_thread), \
> - FN(jiffies64),
> + FN(jiffies64), \
> + FN(perf_prog_read_branches),
>
> /* integer value in 'imm' field of BPF_CALL instruction selects which helper
> * function eBPF program intends to call
> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
> index 19e793aa441a..24c51272a1f7 100644
> --- a/kernel/trace/bpf_trace.c
> +++ b/kernel/trace/bpf_trace.c
> @@ -1028,6 +1028,35 @@ static const struct bpf_func_proto bpf_perf_prog_read_value_proto = {
> .arg3_type = ARG_CONST_SIZE,
> };
>
> +BPF_CALL_3(bpf_perf_prog_read_branches, struct bpf_perf_event_data_kern *, ctx,
> + void *, buf, u32, size)
> +{
> + struct perf_branch_stack *br_stack = ctx->data->br_stack;
> + u32 to_copy = 0, to_clear = size;
> + int err = -EINVAL;
> +
> + if (unlikely(!br_stack))
> + goto clear;
> +
> + to_copy = min_t(u32, br_stack->nr * sizeof(struct perf_branch_entry), size);
> + to_clear -= to_copy;
> +
> + memcpy(buf, br_stack->entries, to_copy);
> + err = to_copy;
> +clear:
> + memset(buf + to_copy, 0, to_clear);
> + return err;
If size < u32, br_stack->nr * sizeof(struct perf_branch_entry),
user has no way to know whether some entries are not copied except
repeated trying larger buffers until the return value is smaller
than input buffer size.
I think returning the expected buffer size to users should be a good
thing? We may not have malloc today in bpf, but future malloc thing
should help in this case.
In user space, user may have a fixed buffer, repeated `read` should
read all values.
Using bpf_probe_read(), repeated read with adjusted source pointer
can also read all buffers.
One possible design is to add a flag to the function, e.g., if
flag == GET_BR_STACK_NR, return br_stack->nr in buf/size.
if flag == GET_BR_STACK, return br_stack->entries in buf/size.
What do you think?
> +}
> +
> +static const struct bpf_func_proto bpf_perf_prog_read_branches_proto = {
> + .func = bpf_perf_prog_read_branches,
> + .gpl_only = true,
> + .ret_type = RET_INTEGER,
> + .arg1_type = ARG_PTR_TO_CTX,
> + .arg2_type = ARG_PTR_TO_UNINIT_MEM,
> + .arg3_type = ARG_CONST_SIZE,
> +};
> +
> static const struct bpf_func_proto *
> pe_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
> {
> @@ -1040,6 +1069,8 @@ pe_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
> return &bpf_get_stack_proto_tp;
> case BPF_FUNC_perf_prog_read_value:
> return &bpf_perf_prog_read_value_proto;
> + case BPF_FUNC_perf_prog_read_branches:
> + return &bpf_perf_prog_read_branches_proto;
> default:
> return tracing_func_proto(func_id, prog);
> }
>
next prev parent reply other threads:[~2020-01-23 23:49 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-23 21:23 [PATCH v3 bpf-next 0/3] Add bpf_perf_prog_read_branches() helper Daniel Xu
2020-01-23 21:23 ` [PATCH v3 bpf-next 1/3] bpf: " Daniel Xu
2020-01-23 23:48 ` Yonghong Song [this message]
2020-01-24 0:49 ` John Fastabend
2020-01-24 2:02 ` Daniel Xu
2020-01-24 8:25 ` Martin Lau
2020-01-24 20:28 ` Daniel Xu
2020-01-23 21:23 ` [PATCH v3 bpf-next 2/3] tools/bpf: Sync uapi header bpf.h Daniel Xu
2020-01-23 21:23 ` [PATCH v3 bpf-next 3/3] selftests/bpf: add bpf_perf_prog_read_branches() selftest Daniel Xu
2020-01-23 23:20 ` Martin Lau
2020-01-24 2:55 ` Daniel Xu
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=5f8e2ffd-9368-cb0c-9591-c6b7ecb8edb0@fb.com \
--to=yhs@fb.com \
--cc=Kernel-team@fb.com \
--cc=acme@kernel.org \
--cc=andriin@fb.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=dxu@dxuuu.xyz \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=songliubraving@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).