[PATCH v2 1/1] x86: Rename __{start,end}_init_task to __{start,end}_init_stack

[PATCH v2 1/1] x86: Rename __{start,end}_init_task to __{start,end}_init_stack

[Xin Li (Intel)]

The stack of a task has been separated from the memory of a task_struct
struture for a long time on x86, as a result __{start,end}_init_task no
longer mark the start and end of the init_task structure, but its stack
only.

Rename __{start,end}_init_task to __{start,end}_init_stack.

Note other architectures are not affected because __{start,end}_init_task
are used on x86 only.

Signed-off-by: Xin Li (Intel) <xin@zytor.com>
---

Change since v1:
* Revert an accident insane change, init_task to init_stack (Jürgen Groß).
---
 arch/x86/include/asm/processor.h  | 4 ++--
 arch/x86/kernel/head_64.S         | 2 +-
 arch/x86/xen/xen-head.S           | 2 +-
 include/asm-generic/vmlinux.lds.h | 6 +++---
 4 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index 811548f131f4..8b3a3f3bb859 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -636,10 +636,10 @@ static __always_inline void prefetchw(const void *x)
 #define KSTK_ESP(task)		(task_pt_regs(task)->sp)
 
 #else
-extern unsigned long __end_init_task[];
+extern unsigned long __end_init_stack[];
 
 #define INIT_THREAD {							\
-	.sp	= (unsigned long)&__end_init_task -			\
+	.sp	= (unsigned long)&__end_init_stack -			\
 		  TOP_OF_KERNEL_STACK_PADDING -				\
 		  sizeof(struct pt_regs),				\
 }
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index d8198fbd70e5..c7babd7ebb0f 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -66,7 +66,7 @@ SYM_CODE_START_NOALIGN(startup_64)
 	mov	%rsi, %r15
 
 	/* Set up the stack for verify_cpu() */
-	leaq	(__end_init_task - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
+	leaq	(__end_init_stack - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
 
 	/* Setup GSBASE to allow stack canary access for C code */
 	movl	$MSR_GS_BASE, %ecx
diff --git a/arch/x86/xen/xen-head.S b/arch/x86/xen/xen-head.S
index 04101b984f24..43eadf03f46d 100644
--- a/arch/x86/xen/xen-head.S
+++ b/arch/x86/xen/xen-head.S
@@ -49,7 +49,7 @@ SYM_CODE_START(startup_xen)
 	ANNOTATE_NOENDBR
 	cld
 
-	leaq	(__end_init_task - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
+	leaq	(__end_init_stack - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
 
 	/* Set up %gs.
 	 *
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 514d3002ad8a..cdfdcca23045 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -399,13 +399,13 @@
 
 #define INIT_TASK_DATA(align)						\
 	. = ALIGN(align);						\
-	__start_init_task = .;						\
+	__start_init_stack = .;						\
 	init_thread_union = .;						\
 	init_stack = .;							\
 	KEEP(*(.data..init_task))					\
 	KEEP(*(.data..init_thread_info))				\
-	. = __start_init_task + THREAD_SIZE;				\
-	__end_init_task = .;
+	. = __start_init_stack + THREAD_SIZE;				\
+	__end_init_stack = .;
 
 #define JUMP_TABLE_DATA							\
 	. = ALIGN(8);							\

base-commit: 7e19a79344df2ed5e106091c29338962261b0290
-- 
2.44.0

Re: [PATCH v2 1/1] x86: Rename __{start,end}_init_task to __{start,end}_init_stack

[Jürgen Groß]

On 18.03.24 08:14, Xin Li (Intel) wrote:
> The stack of a task has been separated from the memory of a task_struct
> struture for a long time on x86, as a result __{start,end}_init_task no
> longer mark the start and end of the init_task structure, but its stack
> only.
> 
> Rename __{start,end}_init_task to __{start,end}_init_stack.
> 
> Note other architectures are not affected because __{start,end}_init_task
> are used on x86 only.
> 
> Signed-off-by: Xin Li (Intel) <xin@zytor.com>

Reviewed-by: Juergen Gross <jgross@suse.com>


Juergen

Re: [PATCH v2 1/1] x86: Rename __{start,end}_init_task to __{start,end}_init_stack

[Ingo Molnar]

* Xin Li (Intel) <xin@zytor.com> wrote:

> The stack of a task has been separated from the memory of a task_struct
> struture for a long time on x86, as a result __{start,end}_init_task no
> longer mark the start and end of the init_task structure, but its stack
> only.
> 
> Rename __{start,end}_init_task to __{start,end}_init_stack.
> 
> Note other architectures are not affected because __{start,end}_init_task
> are used on x86 only.
> 
> Signed-off-by: Xin Li (Intel) <xin@zytor.com>
> ---
> 
> Change since v1:
> * Revert an accident insane change, init_task to init_stack (Jürgen Groß).
> ---
>  arch/x86/include/asm/processor.h  | 4 ++--
>  arch/x86/kernel/head_64.S         | 2 +-
>  arch/x86/xen/xen-head.S           | 2 +-
>  include/asm-generic/vmlinux.lds.h | 6 +++---
>  4 files changed, 7 insertions(+), 7 deletions(-)

Note that this is now in conflict with this cleanup by Brian Gerst:

   2cb16181a1d1 x86/boot: Simplify boot stack setup

... which removed __end_init_task[] entirely.

Thanks,

	Ingo

Re: [PATCH v2 1/1] x86: Rename __{start,end}_init_task to __{start,end}_init_stack

[Xin Li]

On 3/21/2024 12:54 PM, Ingo Molnar wrote:
> 
> * Xin Li (Intel) <xin@zytor.com> wrote:
> 
>> The stack of a task has been separated from the memory of a task_struct
>> struture for a long time on x86, as a result __{start,end}_init_task no
>> longer mark the start and end of the init_task structure, but its stack
>> only.
>>
>> Rename __{start,end}_init_task to __{start,end}_init_stack.
>>
>> Note other architectures are not affected because __{start,end}_init_task
>> are used on x86 only.
>>
>> Signed-off-by: Xin Li (Intel) <xin@zytor.com>
>> ---
>>
>> Change since v1:
>> * Revert an accident insane change, init_task to init_stack (Jürgen Groß).
>> ---
>>   arch/x86/include/asm/processor.h  | 4 ++--
>>   arch/x86/kernel/head_64.S         | 2 +-
>>   arch/x86/xen/xen-head.S           | 2 +-
>>   include/asm-generic/vmlinux.lds.h | 6 +++---
>>   4 files changed, 7 insertions(+), 7 deletions(-)
> 
> Note that this is now in conflict with this cleanup by Brian Gerst:
> 
>     2cb16181a1d1 x86/boot: Simplify boot stack setup

I like this patch, which does something that I'm also doing, further 
x86_64 doesn't have the concept of TOP_OF_KERNEL_STACK_PADDING, and
I guess it was extended from x86_32 but defined to 0 just for code 
consistency.

Yes, FRED code uses TOP_OF_KERNEL_STACK_PADDING, but a FRED event frame 
could contain different amount of information for different event types,
e.g., #MCE could push extra bytes of information, or perhaps even for 
different instances of the same event type. Thus the size of an event 
frame pushed by a FRED CPU is not fixed and the address of the pt_regs 
structure that is used to save the user level context of current task is 
not at a fixed offset from the top of current task kernel stack.

> 
> .. which removed __end_init_task[] entirely.


Not really, but __end_init_task[] is only used to initialize 
__top_init_kernel_stack.

I will rebase and send v3.

Thanks!
     Xin