From: "Jose E. Marchesi" <jose.marchesi@oracle.com>
To: David Faust <david.faust@oracle.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] bpf: set PREFERRED_DEBUGGING_TYPE to BTF_DEBUG
Date: Thu, 25 Apr 2024 20:45:34 +0200 [thread overview]
Message-ID: <87h6fp7035.fsf@oracle.com> (raw)
In-Reply-To: <20240425172320.7967-1-david.faust@oracle.com> (David Faust's message of "Thu, 25 Apr 2024 10:23:20 -0700")
Hi David.
OK. Thanks for the patch.
> BTF is the standard debug info used with BPF programs, so it makes sense
> to default to BTF rather than DWARF.
>
> Tested on x86_64-linux-gnu host for bpf-unknown-none target.
>
> gcc/
> * config/bpf/bpf.h (PREFERRED_DEBUGGING_TYPE): Set to BTF_DEBUG.
>
> gcc/testsuite/
> * gcc.target/bpf/bpf-debug-options-1.c: New test.
> * gcc.target/bpf/bpf-debug-options-2.c: Likewise.
> * gcc.target/bpf/bpf-debug-options-3.c: Likewise.
> * gcc.target/bpf/core-options-4.c: Likewise.
> ---
> gcc/config/bpf/bpf.h | 5 +++++
> .../gcc.target/bpf/bpf-debug-options-1.c | 17 +++++++++++++++++
> .../gcc.target/bpf/bpf-debug-options-2.c | 18 ++++++++++++++++++
> .../gcc.target/bpf/bpf-debug-options-3.c | 14 ++++++++++++++
> gcc/testsuite/gcc.target/bpf/core-options-4.c | 4 ++++
> 5 files changed, 58 insertions(+)
> create mode 100644 gcc/testsuite/gcc.target/bpf/bpf-debug-options-1.c
> create mode 100644 gcc/testsuite/gcc.target/bpf/bpf-debug-options-2.c
> create mode 100644 gcc/testsuite/gcc.target/bpf/bpf-debug-options-3.c
> create mode 100644 gcc/testsuite/gcc.target/bpf/core-options-4.c
>
> diff --git a/gcc/config/bpf/bpf.h b/gcc/config/bpf/bpf.h
> index c67e17526bf..e163fbf688d 100644
> --- a/gcc/config/bpf/bpf.h
> +++ b/gcc/config/bpf/bpf.h
> @@ -245,6 +245,11 @@ enum reg_class
>
> /**** Debugging Info ****/
>
> +/* Use BTF debug info by default. */
> +
> +#undef PREFERRED_DEBUGGING_TYPE
> +#define PREFERRED_DEBUGGING_TYPE BTF_DEBUG
> +
> /* In eBPF it is not possible to unwind frames. Disable CFA. */
>
> #define DWARF2_FRAME_INFO 0
> diff --git a/gcc/testsuite/gcc.target/bpf/bpf-debug-options-1.c b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-1.c
> new file mode 100644
> index 00000000000..409466c4ead
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-1.c
> @@ -0,0 +1,17 @@
> +/* Default to BTF debug info. */
> +/* { dg-do compile } */
> +/* { dg-options "-g -dA" }*/
> +
> +struct A {
> + int x;
> + int y;
> +};
> +
> +int
> +foo (struct A *a)
> +{
> + return a->x;
> +}
> +
> +/* { dg-final { scan-assembler-not "DWARF version" } } */
> +/* { dg-final { scan-assembler "btf_version" } } */
> diff --git a/gcc/testsuite/gcc.target/bpf/bpf-debug-options-2.c b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-2.c
> new file mode 100644
> index 00000000000..03bde12315b
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-2.c
> @@ -0,0 +1,18 @@
> +/* -g defaults to BTF, which in turn implies -mco-re. */
> +/* { dg-do compile } */
> +/* { dg-options "-g -dA" }*/
> +
> +struct A {
> + int x;
> + int y;
> +};
> +
> +int
> +foo (struct A *a)
> +{
> + return __builtin_preserve_access_index (a->x);
> +}
> +
> +/* { dg-final { scan-assembler-not "DWARF version" } } */
> +/* { dg-final { scan-assembler "btf_version" } } */
> +/* { dg-final { scan-assembler "btfext_version" } } */
> diff --git a/gcc/testsuite/gcc.target/bpf/bpf-debug-options-3.c b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-3.c
> new file mode 100644
> index 00000000000..d41790e0928
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/bpf/bpf-debug-options-3.c
> @@ -0,0 +1,14 @@
> +/* Using -g does not incorrectly force CO-RE enabled. */
> +/* { dg-do compile } */
> +/* { dg-options "-g -dA -mno-co-re" }*/
> +
> +struct A {
> + int x;
> + int y;
> +};
> +
> +int
> +foo (struct A *a)
> +{
> + return __builtin_preserve_access_index (a->x); /* { dg-error "BPF CO-RE is required" } */
> +}
> diff --git a/gcc/testsuite/gcc.target/bpf/core-options-4.c b/gcc/testsuite/gcc.target/bpf/core-options-4.c
> new file mode 100644
> index 00000000000..fde4195da42
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/bpf/core-options-4.c
> @@ -0,0 +1,4 @@
> +/* -g implies BTF, -gtoggle turns it off. CO-RE should not work. */
> +/* { dg-do compile } */
> +/* { dg-options "-g -mco-re -gtoggle" } */
> +/* { dg-excess-errors "BPF CO-RE requires BTF debugging information" } */
prev parent reply other threads:[~2024-04-25 18:45 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-25 17:23 David Faust
2024-04-25 18:45 ` Jose E. Marchesi [this message]
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=87h6fp7035.fsf@oracle.com \
--to=jose.marchesi@oracle.com \
--cc=david.faust@oracle.com \
--cc=gcc-patches@gcc.gnu.org \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).