From: "Paul E. McKenney" <[email protected]>
To: [email protected]
Cc: [email protected], [email protected], [email protected],
"Ammar Faizi" <[email protected]>,
"Zhangjin Wu" <[email protected]>,
"Nicholas Rosenberg" <[email protected]>,
"Thomas Weißschuh" <[email protected]>,
"Alviro Iskandar Setiawan" <[email protected]>,
"Willy Tarreau" <[email protected]>
Subject: [PATCH nolibc 01/19] tools/nolibc: i386: Fix a stack misalign bug on _start
Date: Thu, 12 Oct 2023 12:32:15 -0700 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <b34ce3cf-3fcc-4eb0-a658-229c197455ef@paulmck-laptop>
From: Ammar Faizi <[email protected]>
The ABI mandates that the %esp register must be a multiple of 16 when
executing a 'call' instruction.
Commit 2ab446336b17 ("tools/nolibc: i386: shrink _start with _start_c")
simplified the _start function, but it didn't take care of the %esp
alignment, causing SIGSEGV on SSE and AVX programs that use aligned move
instruction (e.g., movdqa, movaps, and vmovdqa).
The 'and $-16, %esp' aligns the %esp at a multiple of 16. Then 'push
%eax' will subtract the %esp by 4; thus, it breaks the 16-byte
alignment. Make sure the %esp is correctly aligned after the push by
subtracting 12 before the push.
Extra:
Add 'add $12, %esp' before the 'and $-16, %esp' to avoid over-estimating
for particular cases as suggested by Willy.
A test program to validate the %esp alignment on _start can be found at:
https://lore.kernel.org/lkml/[email protected]
Cc: Zhangjin Wu <[email protected]>
Fixes: 2ab446336b17aad362c6decee29b4efd83a01979 ("tools/nolibc: i386: shrink _start with _start_c")
Reported-by: Nicholas Rosenberg <[email protected]>
Acked-by: Thomas Weißschuh <[email protected]>
Signed-off-by: Ammar Faizi <[email protected]>
Reviewed-by: Alviro Iskandar Setiawan <[email protected]>
Signed-off-by: Willy Tarreau <[email protected]>
Signed-off-by: Thomas Weißschuh <[email protected]>
---
tools/include/nolibc/arch-i386.h | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/tools/include/nolibc/arch-i386.h b/tools/include/nolibc/arch-i386.h
index 64415b9fac77..28c26a00a762 100644
--- a/tools/include/nolibc/arch-i386.h
+++ b/tools/include/nolibc/arch-i386.h
@@ -167,7 +167,9 @@ void __attribute__((weak, noreturn, optimize("Os", "omit-frame-pointer"))) __no_
__asm__ volatile (
"xor %ebp, %ebp\n" /* zero the stack frame */
"mov %esp, %eax\n" /* save stack pointer to %eax, as arg1 of _start_c */
- "and $-16, %esp\n" /* last pushed argument must be 16-byte aligned */
+ "add $12, %esp\n" /* avoid over-estimating after the 'and' & 'sub' below */
+ "and $-16, %esp\n" /* the %esp must be 16-byte aligned on 'call' */
+ "sub $12, %esp\n" /* sub 12 to keep it aligned after the push %eax */
"push %eax\n" /* push arg1 on stack to support plain stack modes too */
"call _start_c\n" /* transfer to c runtime */
"hlt\n" /* ensure it does not return */
--
2.40.1
next prev parent reply other threads:[~2023-10-12 19:32 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-12 19:32 [PATCH nolibc 0/19] Updates to nolibc for v6.7 (and three for v6.6) Paul E. McKenney
2023-10-12 19:32 ` Paul E. McKenney [this message]
2023-10-12 19:32 ` [PATCH nolibc 02/19] MAINTAINERS: nolibc: update tree location Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 03/19] tools/nolibc: mark start_c as weak Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 04/19] tools/nolibc: add stdarg.h header Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 05/19] selftests/nolibc: use -nostdinc for nolibc-test Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 06/19] tools/nolibc: x86-64: Use `rep movsb` for `memcpy()` and `memmove()` Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 07/19] tools/nolibc: x86-64: Use `rep stosb` for `memset()` Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 08/19] tools/nolibc: string: Remove the `_nolibc_memcpy_down()` function Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 09/19] tools/nolibc: string: Remove the `_nolibc_memcpy_up()` function Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 10/19] selftests/nolibc: libc-test: avoid -Wstringop-overflow warnings Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 11/19] selftests/nolibc: don't embed initramfs into kernel image Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 12/19] selftests/nolibc: allow building i386 with multiarch compiler Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 13/19] tools/nolibc: avoid unused parameter warnings for ENOSYS fallbacks Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 14/19] tools/nolibc: don't define new syscall number Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 15/19] tools/nolibc: automatically detect necessity to use pselect6 Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 16/19] tools/nolibc: drop test for getauxval(AT_PAGESZ) Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 17/19] tools/nolibc: add support for constructors and destructors Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 18/19] selftests/nolibc: use qemu-system-ppc64 for ppc64le Paul E. McKenney
2023-10-12 19:32 ` [PATCH nolibc 19/19] selftests/nolibc: add tests for multi-object linkage Paul E. McKenney
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 \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/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