The unwinder expects a return address at the very top of the kernel
stack just below pt_regs and before any stack frame is created. Instead
of calling a wrapper, set up a return address as if ret_from_fork()
was called from the syscall entry code.
Signed-off-by: Brian Gerst <[email protected]>
---
arch/x86/entry/entry_32.S | 33 ++++++++++-----------------------
1 file changed, 10 insertions(+), 23 deletions(-)
diff --git a/arch/x86/entry/entry_32.S b/arch/x86/entry/entry_32.S
index 91397f58ac30..e56123f03a79 100644
--- a/arch/x86/entry/entry_32.S
+++ b/arch/x86/entry/entry_32.S
@@ -719,26 +719,6 @@ SYM_CODE_START(__switch_to_asm)
SYM_CODE_END(__switch_to_asm)
.popsection
-/*
- * The unwinder expects the last frame on the stack to always be at the same
- * offset from the end of the page, which allows it to validate the stack.
- * Calling schedule_tail() directly would break that convention because its an
- * asmlinkage function so its argument has to be pushed on the stack. This
- * wrapper creates a proper "end of stack" frame header before the call.
- */
-.pushsection .text, "ax"
-SYM_FUNC_START(schedule_tail_wrapper)
- FRAME_BEGIN
-
- pushl %eax
- call schedule_tail
- popl %eax
-
- FRAME_END
- RET
-SYM_FUNC_END(schedule_tail_wrapper)
-.popsection
-
/*
* A newly forked process directly context switches into this address.
*
@@ -748,16 +728,23 @@ SYM_FUNC_END(schedule_tail_wrapper)
*/
.pushsection .text, "ax"
SYM_CODE_START(ret_from_fork)
- call schedule_tail_wrapper
+ /* return address for the stack unwinder */
+ pushl $.Lsyscall_32_done
+
+ FRAME_BEGIN
+ pushl %eax
+ call schedule_tail
+ addl $4, %esp
+ FRAME_END
testl %ebx, %ebx
jnz 1f /* kernel threads are uncommon */
2:
/* When we fork, we trace the syscall return in the child, too. */
- movl %esp, %eax
+ leal 4(%esp), %eax
call syscall_exit_to_user_mode
- jmp .Lsyscall_32_done
+ RET
/* kernel thread */
1: movl %edi, %eax
--
2.41.0
The following commit has been merged into the x86/urgent branch of tip:
Commit-ID: 81f755d561f365f544795fad92f05a085ea4f292
Gitweb: https://git.kernel.org/tip/81f755d561f365f544795fad92f05a085ea4f292
Author: Brian Gerst <[email protected]>
AuthorDate: Fri, 23 Jun 2023 18:55:28 -04:00
Committer: Peter Zijlstra <[email protected]>
CommitterDate: Mon, 10 Jul 2023 09:52:24 +02:00
x86/32: Remove schedule_tail_wrapper()
The unwinder expects a return address at the very top of the kernel
stack just below pt_regs and before any stack frame is created. Instead
of calling a wrapper, set up a return address as if ret_from_fork()
was called from the syscall entry code.
Signed-off-by: Brian Gerst <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Reviewed-by: Sami Tolvanen <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
arch/x86/entry/entry_32.S | 33 ++++++++++-----------------------
1 file changed, 10 insertions(+), 23 deletions(-)
diff --git a/arch/x86/entry/entry_32.S b/arch/x86/entry/entry_32.S
index 91397f5..e56123f 100644
--- a/arch/x86/entry/entry_32.S
+++ b/arch/x86/entry/entry_32.S
@@ -720,26 +720,6 @@ SYM_CODE_END(__switch_to_asm)
.popsection
/*
- * The unwinder expects the last frame on the stack to always be at the same
- * offset from the end of the page, which allows it to validate the stack.
- * Calling schedule_tail() directly would break that convention because its an
- * asmlinkage function so its argument has to be pushed on the stack. This
- * wrapper creates a proper "end of stack" frame header before the call.
- */
-.pushsection .text, "ax"
-SYM_FUNC_START(schedule_tail_wrapper)
- FRAME_BEGIN
-
- pushl %eax
- call schedule_tail
- popl %eax
-
- FRAME_END
- RET
-SYM_FUNC_END(schedule_tail_wrapper)
-.popsection
-
-/*
* A newly forked process directly context switches into this address.
*
* eax: prev task we switched from
@@ -748,16 +728,23 @@ SYM_FUNC_END(schedule_tail_wrapper)
*/
.pushsection .text, "ax"
SYM_CODE_START(ret_from_fork)
- call schedule_tail_wrapper
+ /* return address for the stack unwinder */
+ pushl $.Lsyscall_32_done
+
+ FRAME_BEGIN
+ pushl %eax
+ call schedule_tail
+ addl $4, %esp
+ FRAME_END
testl %ebx, %ebx
jnz 1f /* kernel threads are uncommon */
2:
/* When we fork, we trace the syscall return in the child, too. */
- movl %esp, %eax
+ leal 4(%esp), %eax
call syscall_exit_to_user_mode
- jmp .Lsyscall_32_done
+ RET
/* kernel thread */
1: movl %edi, %eax