From: Peter Zijlstra <peterz@infradead.org>
To: jpoimboe@redhat.com, x86@kernel.org, pjt@google.com,
mbenes@suze.cz, jgross@suse.com
Cc: linux-kernel@vger.kernel.org, peterz@infradead.org
Subject: [RFC][PATCH 5/6] objtool: Allow archs to rewrite retpolines
Date: Fri, 19 Feb 2021 21:43:05 +0100 [thread overview]
Message-ID: <20210219210535.431965352@infradead.org> (raw)
In-Reply-To: 20210219204300.749932493@infradead.org
When retpolines are employed, compilers typically emit calls to
retpoline thunks. Objtool recognises these calls and marks them as
dynamic calls.
Provide infrastructure for architectures to rewrite/augment what the
compiler wrote for us.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
tools/objtool/check.c | 23 +++++++++++++++++++++--
tools/objtool/include/objtool/arch.h | 3 +++
2 files changed, 24 insertions(+), 2 deletions(-)
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -775,6 +775,14 @@ static int add_ignore_alternatives(struc
return 0;
}
+
+__weak int arch_rewrite_retpoline(struct objtool_file *file,
+ struct instruction *insn,
+ struct reloc *reloc)
+{
+ return 0;
+}
+
/*
* Find the destination instructions for all jumps.
*/
@@ -808,6 +816,9 @@ static int add_jump_destinations(struct
insn->type = INSN_JUMP_DYNAMIC_CONDITIONAL;
insn->retpoline_safe = true;
+
+ arch_rewrite_retpoline(file, insn, reloc);
+
continue;
} else if (insn->func) {
/* internal or external sibling call (with reloc) */
@@ -959,6 +970,8 @@ static int add_call_destinations(struct
insn->type = INSN_CALL_DYNAMIC;
insn->retpoline_safe = true;
+ arch_rewrite_retpoline(file, insn, reloc);
+
remove_insn_ops(insn);
continue;
@@ -1119,6 +1132,8 @@ static int handle_group_alt(struct objto
dest_off = arch_jump_destination(insn);
if (dest_off == special_alt->new_off + special_alt->new_len)
insn->jump_dest = next_insn_same_sec(file, last_orig_insn);
+ else
+ insn->jump_dest = find_insn(file, insn->sec, dest_off);
if (!insn->jump_dest) {
WARN_FUNC("can't find alternative jump destination",
@@ -1704,11 +1719,15 @@ static int decode_sections(struct objtoo
if (ret)
return ret;
- ret = add_jump_destinations(file);
+ /*
+ * Must be before add_{jump,call}_destination; for they can add
+ * magic alternatives we can't actually parse.
+ */
+ ret = add_special_section_alts(file);
if (ret)
return ret;
- ret = add_special_section_alts(file);
+ ret = add_jump_destinations(file);
if (ret)
return ret;
--- a/tools/objtool/include/objtool/arch.h
+++ b/tools/objtool/include/objtool/arch.h
@@ -82,6 +82,9 @@ unsigned long arch_jump_destination(stru
unsigned long arch_dest_reloc_offset(int addend);
const char *arch_nop_insn(int len);
+int arch_rewrite_retpoline(struct objtool_file *file,
+ struct instruction *insn,
+ struct reloc *reloc);
int arch_decode_hint_reg(struct instruction *insn, u8 sp_reg);
next prev parent reply other threads:[~2021-02-19 21:11 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-19 20:43 [RFC][PATCH 0/6] objtool: Optimize !retpoline Peter Zijlstra
2021-02-19 20:43 ` [RFC][PATCH 1/6] objtool: Correctly handle retpoline thunk calls Peter Zijlstra
2021-02-19 20:43 ` [RFC][PATCH 2/6] objtool: Fix static_call list generation Peter Zijlstra
2021-02-19 20:43 ` [RFC][PATCH 3/6] objtool: Rework rebuild_reloc logic Peter Zijlstra
2021-02-19 20:43 ` [RFC][PATCH 4/6] objtool: Add elf_create_undef_symbol() Peter Zijlstra
2021-02-19 20:43 ` Peter Zijlstra [this message]
2021-02-19 20:43 ` [RFC][PATCH 6/6] objtool,x86: Rewrite retpoline thunk calls Peter Zijlstra
2021-02-19 21:55 ` Josh Poimboeuf
2021-02-19 22:01 ` Peter Zijlstra
2021-02-20 0:39 ` Borislav Petkov
2021-02-20 16:48 ` Peter Zijlstra
2021-02-20 17:41 ` Borislav Petkov
2021-02-20 22:28 ` Peter Zijlstra
2021-02-20 22:51 ` Peter Zijlstra
2021-02-21 9:54 ` Borislav Petkov
2021-02-20 22:32 ` Peter Zijlstra
2021-02-21 5:45 ` Jürgen Groß
2021-02-21 9:44 ` Borislav Petkov
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=20210219210535.431965352@infradead.org \
--to=peterz@infradead.org \
--cc=jgross@suse.com \
--cc=jpoimboe@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mbenes@suze.cz \
--cc=pjt@google.com \
--cc=x86@kernel.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