unofficial mirror of libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Sergey Bugaev <bugaevc@gmail.com>
To: libc-alpha@sourceware.org, bug-hurd@gnu.org
Subject: [RFC PATCH 07/23] hurd: Pass the data pointer to _hurd_stack_setup explicitly
Date: Wed,  3 Jan 2024 20:14:40 +0300	[thread overview]
Message-ID: <20240103171502.1358371-8-bugaevc@gmail.com> (raw)
In-Reply-To: <20240103171502.1358371-1-bugaevc@gmail.com>

Instead of relying on the stack frame layout to figure out where the stack
pointer was prior to the _hurd_stack_setup () call, just pass the pointer
as an argument explicitly. This is less brittle and much more portable.

Signed-off-by: Sergey Bugaev <bugaevc@gmail.com>
---
 sysdeps/mach/hurd/i386/static-start.S   |  3 +++
 sysdeps/mach/hurd/x86/init-first.c      | 16 +++++++---------
 sysdeps/mach/hurd/x86_64/static-start.S |  1 +
 3 files changed, 11 insertions(+), 9 deletions(-)

diff --git a/sysdeps/mach/hurd/i386/static-start.S b/sysdeps/mach/hurd/i386/static-start.S
index d83505b2..3ffcb47d 100644
--- a/sysdeps/mach/hurd/i386/static-start.S
+++ b/sysdeps/mach/hurd/i386/static-start.S
@@ -19,7 +19,10 @@
 	.text
 	.globl _start
 _start:
+	pushl %esp
 	call _hurd_stack_setup
+	/* No need to "addl %4, %esp", since _hurd_stack_setup
+	 * returns with an already adjusted stack pointer.  */
 	xorl %edx, %edx
 	jmp _start1
 
diff --git a/sysdeps/mach/hurd/x86/init-first.c b/sysdeps/mach/hurd/x86/init-first.c
index bb051418..6f71d71b 100644
--- a/sysdeps/mach/hurd/x86/init-first.c
+++ b/sysdeps/mach/hurd/x86/init-first.c
@@ -197,7 +197,7 @@ strong_alias (posixland_init, __libc_init_first);
    which should not exist at all.  */
 void
 inhibit_stack_protector
-_hurd_stack_setup (void)
+_hurd_stack_setup (void **argptr)
 {
   /* This is the very first C code that runs in a statically linked
      executable -- calling this function is the first thing that _start in
@@ -206,14 +206,12 @@ _hurd_stack_setup (void)
 
      _start1 expects the arguments, environment, and a Hurd data block to be
      located at the top of the stack.  The data may already be located there,
-     or we may need to receive it from the exec server.  */
-  void *caller = __builtin_extract_return_addr (__builtin_return_address (0));
-  /* If the arguments and environment are already located on the stack, this is
-     where they are, just above our call frame.  Note that this may not be a
-     valid pointer in case we're supposed to receive the arguments from the exec
-     server, so we can not dereference it yet.  */
-  void **p = (void **) __builtin_frame_address (0) + 2;
+     or we may need to receive it from the exec server.  If the data is located
+     on the stack (just above our call frame), argptr points to it.  Note that
+     this may not be a valid pointer in case we're supposed to receive the
+     arguments from the exec server, so we can not dereference it yet.  */
 
+  void *caller = __builtin_extract_return_addr (__builtin_return_address (0));
   /* Init the essential things.  */
   first_init ();
 
@@ -245,7 +243,7 @@ _hurd_stack_setup (void)
      the stack pointer to the data (which is somewhere on the current stack
      anyway).  This way, _start1 find the data on the top of the stack, just as
      it expects to.  */
-  _hurd_startup (p, &doinit);
+  _hurd_startup (argptr, &doinit);
   __builtin_unreachable ();
 }
 #endif
diff --git a/sysdeps/mach/hurd/x86_64/static-start.S b/sysdeps/mach/hurd/x86_64/static-start.S
index 9b9db937..0ec00905 100644
--- a/sysdeps/mach/hurd/x86_64/static-start.S
+++ b/sysdeps/mach/hurd/x86_64/static-start.S
@@ -25,6 +25,7 @@ _start:
 	leaq __strlen_sse2(%rip), %rax
 	movq %rax, strlen@GOTPCREL(%rip)
 
+	movq %rsp, %rdi
 	call _hurd_stack_setup
 	xorq %rdx, %rdx
 	jmp _start1
-- 
2.43.0


  parent reply	other threads:[~2024-01-03 17:17 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-03 17:14 [RFC PATCH 00/23] aarch64-gnu port Sergey Bugaev
2024-01-03 17:14 ` [PATCH 01/23] hurd: Add some missing includes Sergey Bugaev
2024-01-03 20:43   ` Samuel Thibault
2024-01-03 21:08     ` Sergey Bugaev
2024-01-03 21:24       ` Sergey Bugaev
2024-01-03 21:35         ` Samuel Thibault
2024-01-03 21:00   ` Samuel Thibault
2024-01-03 17:14 ` [PATCH 02/23] hurd: Declare _hurd_intr_rpc_msg* with protected visibility Sergey Bugaev
2024-01-03 21:00   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 03/23] Allow glibc to be compiled without EXEC_PAGESIZE Sergey Bugaev
2024-03-11 16:13   ` Florian Weimer
2024-03-25 11:58     ` Florian Weimer
2024-03-25 12:24       ` Sergey Bugaev
2024-04-10 11:57         ` Florian Weimer
2024-04-15 12:53           ` Sergey Bugaev
2024-04-22 21:01         ` Samuel Thibault
2024-01-03 17:14 ` [PATCH 04/23] mach: Drop some unnecessary vm_param.h includes Sergey Bugaev
2024-01-03 21:00   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 05/23] hurd: Disable Prefer_MAP_32BIT_EXEC on non-x86_64 for now Sergey Bugaev
2024-01-03 17:14 ` [PATCH 06/23] mach: Drop SNARF_ARGS macro Sergey Bugaev
2024-01-03 21:00   ` Samuel Thibault
2024-01-03 17:14 ` Sergey Bugaev [this message]
2024-01-04 22:47   ` [RFC PATCH 07/23] hurd: Pass the data pointer to _hurd_stack_setup explicitly Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 08/23] hurd: Drop x86-specific assembly from init-first.c Sergey Bugaev
2024-01-04 22:48   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 09/23] hurd: Make init-first.c no longer x86-specific Sergey Bugaev
2024-01-04 22:48   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 10/23] hurd: Only init early static TLS if it's used to store stack or pointer guards Sergey Bugaev
2024-01-04 22:48   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 11/23] hurd: Initializy _dl_pagesize early in static builds Sergey Bugaev
2024-01-04 22:48   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 12/23] aarch64: Make cpu-features definitions not Linux-specific Sergey Bugaev
2024-01-04 22:48   ` Samuel Thibault
2024-01-03 17:14 ` [RFC PATCH 13/23] aarch64: Add dl-procinfo Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 14/23] aarch64: Allow building without kernel support for BTI Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 15/23] mach: Add a basic AArch64 port Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 16/23] hurd: " Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 17/23] hurd: Implement TLS on AArch64 Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 18/23] hurd: Implement longjmp for AArch64 Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 19/23] Add FPE_FLTIDO Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 20/23] hurd: Add an AArch64 signal implementation Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 21/23] htl: Implement some support for TLS_DTV_AT_TP Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 22/23] htl: Add an AArch64 implementation Sergey Bugaev
2024-01-03 17:14 ` [RFC PATCH 23/23] hurd: Add expected aarch64-gnu abistlists Sergey Bugaev
2024-01-03 17:30 ` [RFC PATCH 00/23] aarch64-gnu port Joseph Myers
2024-01-03 17:57   ` Sergey Bugaev
2024-03-11 16:10   ` Florian Weimer
2024-03-11 13:47 ` H.J. Lu
2024-03-11 15:38   ` Sergey Bugaev
2024-03-11 14:14 ` Maxim Kuvyrkov

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

  List information: https://www.gnu.org/software/libc/involved.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240103171502.1358371-8-bugaevc@gmail.com \
    --to=bugaevc@gmail.com \
    --cc=bug-hurd@gnu.org \
    --cc=libc-alpha@sourceware.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).