startup_64_setup_env is first introduced by 'commit 866b556efa12
("x86/head/64: Install startup GDT")'. Then 'commit 469693d8f622
("x86/head/64: Re-enable stack protection")' insert some code which
leverage %rdi, which is removed by 'commit 8f6be6d870e8 ("x86/smpboot:
Remove initial_gs")'.
This left the parameter of startup_64_setup_env a little far away from
where it is used. Let's move them together for better understanding.
Signed-off-by: Wei Yang <[email protected]>
CC: Joerg Roedel <[email protected]>
CC: Michael Roth <[email protected]>
CC: Brian Gerst <[email protected]>
CC: Ard Biesheuvel <[email protected]>
---
arch/x86/kernel/head_64.S | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index 3dbd05f93859..5fb8bb7cdabe 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -69,8 +69,6 @@ SYM_CODE_START_NOALIGN(startup_64)
/* Set up the stack for verify_cpu() */
leaq (__end_init_task - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
- leaq _text(%rip), %rdi
-
/* Setup GSBASE to allow stack canary access for C code */
movl $MSR_GS_BASE, %ecx
leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx
@@ -78,6 +76,7 @@ SYM_CODE_START_NOALIGN(startup_64)
shrq $32, %rdx
wrmsr
+ leaq _text(%rip), %rdi
call startup_64_setup_env
/* Now switch to __KERNEL_CS so IRET works reliably */
--
2.34.1
Hello Richard,
On Fri, 15 Mar 2024 at 05:14, Wei Yang <[email protected]> wrote:
>
> startup_64_setup_env is first introduced by 'commit 866b556efa12
> ("x86/head/64: Install startup GDT")'. Then 'commit 469693d8f622
> ("x86/head/64: Re-enable stack protection")' insert some code which
> leverage %rdi, which is removed by 'commit 8f6be6d870e8 ("x86/smpboot:
> Remove initial_gs")'.
>
> This left the parameter of startup_64_setup_env a little far away from
> where it is used. Let's move them together for better understanding.
>
> Signed-off-by: Wei Yang <[email protected]>
> CC: Joerg Roedel <[email protected]>
> CC: Michael Roth <[email protected]>
> CC: Brian Gerst <[email protected]>
> CC: Ard Biesheuvel <[email protected]>
> ---
> arch/x86/kernel/head_64.S | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
The code that you are fixing here no longer exists.
If you are going to send patches in the middle of the merge window,
please make sure that you are at least up to date with Linus's tree at
the time you send the patch.
> diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
> index 3dbd05f93859..5fb8bb7cdabe 100644
> --- a/arch/x86/kernel/head_64.S
> +++ b/arch/x86/kernel/head_64.S
> @@ -69,8 +69,6 @@ SYM_CODE_START_NOALIGN(startup_64)
> /* Set up the stack for verify_cpu() */
> leaq (__end_init_task - TOP_OF_KERNEL_STACK_PADDING - PTREGS_SIZE)(%rip), %rsp
>
> - leaq _text(%rip), %rdi
> -
> /* Setup GSBASE to allow stack canary access for C code */
> movl $MSR_GS_BASE, %ecx
> leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx
> @@ -78,6 +76,7 @@ SYM_CODE_START_NOALIGN(startup_64)
> shrq $32, %rdx
> wrmsr
>
> + leaq _text(%rip), %rdi
> call startup_64_setup_env
>
> /* Now switch to __KERNEL_CS so IRET works reliably */
> --
> 2.34.1
>
On Fri, Mar 15, 2024 at 10:10:06AM +0100, Ard Biesheuvel wrote:
>Hello Richard,
>
>On Fri, 15 Mar 2024 at 05:14, Wei Yang <[email protected]> wrote:
>>
>> startup_64_setup_env is first introduced by 'commit 866b556efa12
>> ("x86/head/64: Install startup GDT")'. Then 'commit 469693d8f622
>> ("x86/head/64: Re-enable stack protection")' insert some code which
>> leverage %rdi, which is removed by 'commit 8f6be6d870e8 ("x86/smpboot:
>> Remove initial_gs")'.
>>
>> This left the parameter of startup_64_setup_env a little far away from
>> where it is used. Let's move them together for better understanding.
>>
>> Signed-off-by: Wei Yang <[email protected]>
>> CC: Joerg Roedel <[email protected]>
>> CC: Michael Roth <[email protected]>
>> CC: Brian Gerst <[email protected]>
>> CC: Ard Biesheuvel <[email protected]>
>> ---
>> arch/x86/kernel/head_64.S | 3 +--
>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>
>
>The code that you are fixing here no longer exists.
>
>If you are going to send patches in the middle of the merge window,
>please make sure that you are at least up to date with Linus's tree at
>the time you send the patch.
>
Thanks, will do it next time.
--
Wei Yang
Help you, Help me