From: "Alexei Starovoitov" <alexei.starovoitov@gmail.com>
To: "Eduard Zingerman" <eddyz87@gmail.com>, <bpf@vger.kernel.org>
Cc: <daniel@iogearbox.net>, <andrii@kernel.org>, <memxor@gmail.com>,
<a.s.protopopov@gmail.com>
Subject: Re: [PATCH bpf-next 04/17] bpf: Prepare static analysis passes for callx instruction
Date: Wed, 23 Sep 2026 23:51:01 +0000 [thread overview]
Message-ID: <DLN3KYEZPIQR.30NJCCM978VV1@gmail.com> (raw)
In-Reply-To: <7917b7dbfbc2d206b0b4c22d2352f5b4700125f0.camel@gmail.com>
On Wed, Sep 23, 2026 at 04:10 PM Eduard Zingerman <eddyz87@gmail.com> wrote:
>> - alive = bpf_calls_callback(env, callsite)
>> + alive = callee_stack_access_at_callsite(env, callsite)
>> ? is_live_before(instance, callsite, rel, half_spi)
>> : is_live_before(instance, callsite + 1, rel, half_spi);
>> if (alive)
>
> I don't understand this change.
> The original reasoning is that callback calling function can call the
> callback many times and the callback might access stack slots from
> outer frames. Hence is_live_before(... callsite ...).
> callx only calls target once.
It's not about the number of calls.
analyze_subprog() doesn't know the callee of callx and doesn't create
an instance for it at this callsite. While the verifier is in the
callee lookup_instance() finds its standalone depth 0 instance that
knows nothing about outer frames. What the callee may read in the
caller's stack is recorded as may_read of callx insn itself by
record_call_access(). That's in live_before(callsite), but not in
live_before(callsite + 1).
callx_callee_reads_caller_stack_ok in patch 15 is a test case for this.
Without this hunk fp-8 is dead.
next prev parent reply other threads:[~2026-09-23 23:51 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-22 1:13 [PATCH bpf-next 00/17] bpf: Indirect calls of bpf subprogs (callx) Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 01/17] bpf: Fix infinite loop in check_max_stack_depth() Alexei Starovoitov
2026-09-22 2:01 ` bot+bpf-ci
2026-09-22 2:56 ` Alexei Starovoitov
2026-09-23 22:35 ` Eduard Zingerman
2026-09-22 1:13 ` [PATCH bpf-next 02/17] selftests/bpf: Test recursion through a global function and a callback Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 03/17] bpf: Don't fold loads from insn_array maps into constants Alexei Starovoitov
2026-09-23 22:39 ` Eduard Zingerman
2026-09-23 23:12 ` Alexei Starovoitov
2026-09-24 0:19 ` bot+bpf-ci
2026-09-22 1:13 ` [PATCH bpf-next 04/17] bpf: Prepare static analysis passes for callx instruction Alexei Starovoitov
2026-09-23 23:10 ` Eduard Zingerman
2026-09-23 23:51 ` Alexei Starovoitov [this message]
2026-09-22 1:13 ` [PATCH bpf-next 05/17] bpf: Add callx instruction to call bpf subprogs indirectly Alexei Starovoitov
2026-09-24 0:09 ` Eduard Zingerman
2026-09-22 1:13 ` [PATCH bpf-next 06/17] bpf: Add callx calls to the call graph Alexei Starovoitov
2026-09-22 1:27 ` sashiko-bot
2026-09-22 2:54 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 07/17] bpf, x86: Add JIT support for callx Alexei Starovoitov
2026-09-22 1:27 ` sashiko-bot
2026-09-22 2:53 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 08/17] bpf, arm64: " Alexei Starovoitov
2026-09-22 15:05 ` Puranjay Mohan
2026-09-22 1:13 ` [PATCH bpf-next 09/17] bpf: Discover subprogs described by func_info Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 10/17] bpf: Recognize pointers to functions in read-only maps Alexei Starovoitov
2026-09-22 1:31 ` sashiko-bot
2026-09-22 3:01 ` Alexei Starovoitov
2026-09-24 0:46 ` bot+bpf-ci
2026-09-24 2:12 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 11/17] libbpf: Support pointers to static functions in data when linking Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 12/17] libbpf: Resolve pointers to functions in read-only data Alexei Starovoitov
2026-09-24 0:33 ` bot+bpf-ci
2026-09-24 2:13 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 13/17] libbpf: Treat .data.rel.ro as " Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 14/17] libbpf: Support pointers to functions in read-only data in light skeleton Alexei Starovoitov
2026-09-22 2:01 ` bot+bpf-ci
2026-09-22 2:55 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 15/17] selftests/bpf: Add tests for callx Alexei Starovoitov
2026-09-24 0:33 ` bot+bpf-ci
2026-09-24 2:13 ` Alexei Starovoitov
2026-09-22 1:13 ` [PATCH bpf-next 16/17] selftests/bpf: Add tests for callx through pointers in read-only data Alexei Starovoitov
2026-09-22 2:01 ` bot+bpf-ci
2026-09-24 0:33 ` bot+bpf-ci
2026-09-22 1:13 ` [PATCH bpf-next 17/17] bpf, docs: Document callx instruction Alexei Starovoitov
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=DLN3KYEZPIQR.30NJCCM978VV1@gmail.com \
--to=alexei.starovoitov@gmail.com \
--cc=a.s.protopopov@gmail.com \
--cc=andrii@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=eddyz87@gmail.com \
--cc=memxor@gmail.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