From: Peter Zijlstra <peterz@infradead.org>
To: Josh Poimboeuf <jpoimboe@redhat.com>,
Thomas Gleixner <tglx@linutronix.de>
Cc: Miroslav Benes <mbenes@suse.cz>,
Nick Desaulniers <ndesaulniers@google.com>,
Julien Thierry <jthierry@redhat.com>,
Kees Cook <keescook@chromium.org>,
x86@kernel.org, linux-kernel@vger.kernel.org,
peterz@infradead.org
Subject: [PATCH 3/5] objtool: Prepare for scratch regs
Date: Wed, 03 Feb 2021 13:02:25 +0100 [thread overview]
Message-ID: <20210203120401.200627050@infradead.org> (raw)
In-Reply-To: 20210203120222.451068583@infradead.org
Introduce __save_reg() which allows using the
!arch_callee_saved_reg()s and make sure they're wiped after every
stack op so they don't linger, allow for a single op exception so they
can be used on the very next stack-op.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
tools/objtool/check.c | 27 ++++++++++++++++++++++++---
1 file changed, 24 insertions(+), 3 deletions(-)
--- a/tools/objtool/check.c
+++ b/tools/objtool/check.c
@@ -1775,15 +1775,20 @@ static int update_cfi_state_regs(struct
return 0;
}
-static void save_reg(struct cfi_state *cfi, unsigned char reg, int base, int offset)
+static void __save_reg(struct cfi_state *cfi, unsigned char reg, int base, int offset)
{
- if (arch_callee_saved_reg(reg) &&
- cfi->regs[reg].base == CFI_UNDEFINED) {
+ if (cfi->regs[reg].base == CFI_UNDEFINED) {
cfi->regs[reg].base = base;
cfi->regs[reg].offset = offset;
}
}
+static void save_reg(struct cfi_state *cfi, unsigned char reg, int base, int offset)
+{
+ if (arch_callee_saved_reg(reg))
+ __save_reg(cfi, reg, base, offset);
+}
+
static void restore_reg(struct cfi_state *cfi, unsigned char reg)
{
cfi->regs[reg].base = initial_func_cfi.regs[reg].base;
@@ -1848,6 +1853,7 @@ static int update_cfi_state(struct instr
{
struct cfi_reg *cfa = &cfi->cfa;
struct cfi_reg *regs = cfi->regs;
+ bool skip_wipe = false;
/* stack operations don't make sense with an undefined CFA */
if (cfa->base == CFI_UNDEFINED) {
@@ -2192,6 +2198,21 @@ static int update_cfi_state(struct instr
return -1;
}
+ /*
+ * Only callee saved registers are preserved; the rest is scratch space
+ * preserved at most one instruction.
+ */
+ if (!skip_wipe) {
+ int i;
+
+ for (i = 0; i < CFI_NUM_REGS; i++) {
+ if (arch_callee_saved_reg(i))
+ continue;
+
+ restore_reg(cfi, i);
+ }
+ }
+
return 0;
}
next prev parent reply other threads:[~2021-02-03 12:05 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-03 12:02 [PATCH 0/5] objtool: The stack swizzle again Peter Zijlstra
2021-02-03 12:02 ` [PATCH 1/5] objtool: Change REG_SP_INDIRECT Peter Zijlstra
2021-02-03 14:42 ` Josh Poimboeuf
2021-02-03 14:49 ` Peter Zijlstra
2021-02-03 14:55 ` Josh Poimboeuf
2021-02-03 12:02 ` [PATCH 2/5] objtool: More consistent use of arch_callee_saved_reg() Peter Zijlstra
2021-02-03 12:02 ` Peter Zijlstra [this message]
2021-02-03 12:02 ` [PATCH 4/5] objtool,x86: Additionally decode: mov %rsp, (%reg) Peter Zijlstra
2021-02-03 12:02 ` [PATCH 5/5] objtool: Support stack-swizzle Peter Zijlstra
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=20210203120401.200627050@infradead.org \
--to=peterz@infradead.org \
--cc=jpoimboe@redhat.com \
--cc=jthierry@redhat.com \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mbenes@suse.cz \
--cc=ndesaulniers@google.com \
--cc=tglx@linutronix.de \
--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;
as well as URLs for NNTP newsgroup(s).