* [RFC PATCH v3 0/3] LoongArch: Modify handle_syscall
@ 2022-06-15 10:29 Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 1/3] LoongArch: Add TI_SYSCALL in output_thread_info_defines() Tiezhu Yang
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Tiezhu Yang @ 2022-06-15 10:29 UTC (permalink / raw)
To: Huacai Chen, WANG Xuerui
Cc: Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, linux-kernel
Tiezhu Yang (3):
LoongArch: Add TI_SYSCALL in output_thread_info_defines()
LoongArch: Only clone and clone3 need to call SAVE_STATIC
LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls
arch/loongarch/include/asm/stackframe.h | 17 ++++++++++++++++
arch/loongarch/kernel/asm-offsets.c | 1 +
arch/loongarch/kernel/entry.S | 35 ++++++++++++++++++++++++++++++++-
3 files changed, 52 insertions(+), 1 deletion(-)
--
2.1.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* [RFC PATCH v3 1/3] LoongArch: Add TI_SYSCALL in output_thread_info_defines()
2022-06-15 10:29 [RFC PATCH v3 0/3] LoongArch: Modify handle_syscall Tiezhu Yang
@ 2022-06-15 10:29 ` Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 2/3] LoongArch: Only clone and clone3 need to call SAVE_STATIC Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls Tiezhu Yang
2 siblings, 0 replies; 6+ messages in thread
From: Tiezhu Yang @ 2022-06-15 10:29 UTC (permalink / raw)
To: Huacai Chen, WANG Xuerui
Cc: Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, linux-kernel
The initial idea was to store the syscall number in PT_R11,
and then we can get the syscall number from PT_R11 to check
before SAVE and RESTORE in handle_syscall, but PT_R11 may
be overwritten by the signal handler and the syscall number
is lost.
Add TI_SYSCALL in output_thread_info_defines(), then we can
store the syscall number in TI_SYSCALL. This is preparation
for later patch.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/loongarch/kernel/asm-offsets.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/loongarch/kernel/asm-offsets.c b/arch/loongarch/kernel/asm-offsets.c
index bfb65eb..4757ebe 100644
--- a/arch/loongarch/kernel/asm-offsets.c
+++ b/arch/loongarch/kernel/asm-offsets.c
@@ -81,6 +81,7 @@ void output_thread_info_defines(void)
OFFSET(TI_CPU, thread_info, cpu);
OFFSET(TI_PRE_COUNT, thread_info, preempt_count);
OFFSET(TI_REGS, thread_info, regs);
+ OFFSET(TI_SYSCALL, thread_info, syscall);
DEFINE(_THREAD_SIZE, THREAD_SIZE);
DEFINE(_THREAD_MASK, THREAD_MASK);
DEFINE(_IRQ_STACK_SIZE, IRQ_STACK_SIZE);
--
2.1.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [RFC PATCH v3 2/3] LoongArch: Only clone and clone3 need to call SAVE_STATIC
2022-06-15 10:29 [RFC PATCH v3 0/3] LoongArch: Modify handle_syscall Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 1/3] LoongArch: Add TI_SYSCALL in output_thread_info_defines() Tiezhu Yang
@ 2022-06-15 10:29 ` Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls Tiezhu Yang
2 siblings, 0 replies; 6+ messages in thread
From: Tiezhu Yang @ 2022-06-15 10:29 UTC (permalink / raw)
To: Huacai Chen, WANG Xuerui
Cc: Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, linux-kernel
In handle_syscall, it is unnecessary to call SAVE_STATIC for all syscalls,
only clone and clone3 need to do this operation, so it is better to check
the syscall number before call SAVE_STATIC.
With this patch, it can reduce many store instructions.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/loongarch/kernel/entry.S | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/arch/loongarch/kernel/entry.S b/arch/loongarch/kernel/entry.S
index d5b3dbc..a0f7fcc 100644
--- a/arch/loongarch/kernel/entry.S
+++ b/arch/loongarch/kernel/entry.S
@@ -14,6 +14,7 @@
#include <asm/regdef.h>
#include <asm/stackframe.h>
#include <asm/thread_info.h>
+#include <asm/unistd.h>
.text
.cfi_sections .debug_frame
@@ -56,8 +57,17 @@ SYM_FUNC_START(handle_syscall)
cfi_st u0, PT_R21
cfi_st fp, PT_R22
+ /* Syscall number held in a7, only clone and clone3 need to call SAVE_STATIC. */
+ li.w t3, __NR_clone
+ beq a7, t3, 1f
+ li.w t3, __NR_clone3
+ beq a7, t3, 1f
+ b 2f
+
+1:
SAVE_STATIC
+2:
move u0, t0
li.d tp, ~_THREAD_MASK
and tp, tp, sp
--
2.1.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls
2022-06-15 10:29 [RFC PATCH v3 0/3] LoongArch: Modify handle_syscall Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 1/3] LoongArch: Add TI_SYSCALL in output_thread_info_defines() Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 2/3] LoongArch: Only clone and clone3 need to call SAVE_STATIC Tiezhu Yang
@ 2022-06-15 10:29 ` Tiezhu Yang
2022-06-16 12:52 ` Huacai Chen
2 siblings, 1 reply; 6+ messages in thread
From: Tiezhu Yang @ 2022-06-15 10:29 UTC (permalink / raw)
To: Huacai Chen, WANG Xuerui
Cc: Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, linux-kernel
In handle_syscall, it is unnecessary to call RESTORE_ALL_AND_RET
for all syscalls.
(1) clone and clone3 call RESTORE_STATIC_SOME_SP_AND_RET.
(2) rt_sigreturn and rt_sigsuspend call RESTORE_TEMP_SOME_SP_AND_RET.
(3) The other syscalls call RESTORE_SOME_SP_AND_RET.
With this patch, it can reduce many load instructions.
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
---
arch/loongarch/include/asm/stackframe.h | 17 +++++++++++++++++
arch/loongarch/kernel/entry.S | 25 ++++++++++++++++++++++++-
2 files changed, 41 insertions(+), 1 deletion(-)
diff --git a/arch/loongarch/include/asm/stackframe.h b/arch/loongarch/include/asm/stackframe.h
index 4ca9530..52649a5f 100644
--- a/arch/loongarch/include/asm/stackframe.h
+++ b/arch/loongarch/include/asm/stackframe.h
@@ -216,4 +216,21 @@
RESTORE_SP_AND_RET \docfi
.endm
+ .macro RESTORE_SOME_SP_AND_RET docfi=0
+ RESTORE_SOME \docfi
+ RESTORE_SP_AND_RET \docfi
+ .endm
+
+ .macro RESTORE_STATIC_SOME_SP_AND_RET docfi=0
+ RESTORE_STATIC \docfi
+ RESTORE_SOME \docfi
+ RESTORE_SP_AND_RET \docfi
+ .endm
+
+ .macro RESTORE_TEMP_SOME_SP_AND_RET docfi=0
+ RESTORE_TEMP \docfi
+ RESTORE_SOME \docfi
+ RESTORE_SP_AND_RET \docfi
+ .endm
+
#endif /* _ASM_STACKFRAME_H */
diff --git a/arch/loongarch/kernel/entry.S b/arch/loongarch/kernel/entry.S
index a0f7fcc..e27e38b 100644
--- a/arch/loongarch/kernel/entry.S
+++ b/arch/loongarch/kernel/entry.S
@@ -72,10 +72,33 @@ SYM_FUNC_START(handle_syscall)
li.d tp, ~_THREAD_MASK
and tp, tp, sp
+ /* Syscall number held in a7, we can store it in TI_SYSCALL. */
+ LONG_S a7, tp, TI_SYSCALL
+
move a0, sp
bl do_syscall
- RESTORE_ALL_AND_RET
+ /*
+ * Syscall number held in a7 which is stored in TI_SYSCALL.
+ * clone and clone3 call RESTORE_STATIC_SOME_SP_AND_RET.
+ * rt_sigreturn and rt_sigsuspend call RESTORE_TEMP_SOME_SP_AND_RET.
+ * The other syscalls call RESTORE_SOME_SP_AND_RET.
+ */
+ LONG_L t3, tp, TI_SYSCALL
+ li.w t4, __NR_clone
+ beq t3, t4, 3f
+ li.w t4, __NR_clone3
+ beq t3, t4, 3f
+ li.w t4, __NR_rt_sigreturn
+ beq t3, t4, 4f
+ li.w t4, __NR_rt_sigsuspend
+ beq t3, t4, 4f
+
+ RESTORE_SOME_SP_AND_RET
+3:
+ RESTORE_STATIC_SOME_SP_AND_RET
+4:
+ RESTORE_TEMP_SOME_SP_AND_RET
SYM_FUNC_END(handle_syscall)
SYM_CODE_START(ret_from_fork)
--
2.1.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls
2022-06-15 10:29 ` [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls Tiezhu Yang
@ 2022-06-16 12:52 ` Huacai Chen
2022-06-16 13:36 ` Tiezhu Yang
0 siblings, 1 reply; 6+ messages in thread
From: Huacai Chen @ 2022-06-16 12:52 UTC (permalink / raw)
To: Tiezhu Yang; +Cc: WANG Xuerui, Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, LKML
Hi, Tiezhu,
On Wed, Jun 15, 2022 at 6:29 PM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>
> In handle_syscall, it is unnecessary to call RESTORE_ALL_AND_RET
> for all syscalls.
>
> (1) clone and clone3 call RESTORE_STATIC_SOME_SP_AND_RET.
> (2) rt_sigreturn and rt_sigsuspend call RESTORE_TEMP_SOME_SP_AND_RET.
I doubt that you have not tested, it is obvious that rt_sigreturn need
RESTORE_ALL.
Huacai
> (3) The other syscalls call RESTORE_SOME_SP_AND_RET.
>
> With this patch, it can reduce many load instructions.
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
> ---
> arch/loongarch/include/asm/stackframe.h | 17 +++++++++++++++++
> arch/loongarch/kernel/entry.S | 25 ++++++++++++++++++++++++-
> 2 files changed, 41 insertions(+), 1 deletion(-)
>
> diff --git a/arch/loongarch/include/asm/stackframe.h b/arch/loongarch/include/asm/stackframe.h
> index 4ca9530..52649a5f 100644
> --- a/arch/loongarch/include/asm/stackframe.h
> +++ b/arch/loongarch/include/asm/stackframe.h
> @@ -216,4 +216,21 @@
> RESTORE_SP_AND_RET \docfi
> .endm
>
> + .macro RESTORE_SOME_SP_AND_RET docfi=0
> + RESTORE_SOME \docfi
> + RESTORE_SP_AND_RET \docfi
> + .endm
> +
> + .macro RESTORE_STATIC_SOME_SP_AND_RET docfi=0
> + RESTORE_STATIC \docfi
> + RESTORE_SOME \docfi
> + RESTORE_SP_AND_RET \docfi
> + .endm
> +
> + .macro RESTORE_TEMP_SOME_SP_AND_RET docfi=0
> + RESTORE_TEMP \docfi
> + RESTORE_SOME \docfi
> + RESTORE_SP_AND_RET \docfi
> + .endm
> +
> #endif /* _ASM_STACKFRAME_H */
> diff --git a/arch/loongarch/kernel/entry.S b/arch/loongarch/kernel/entry.S
> index a0f7fcc..e27e38b 100644
> --- a/arch/loongarch/kernel/entry.S
> +++ b/arch/loongarch/kernel/entry.S
> @@ -72,10 +72,33 @@ SYM_FUNC_START(handle_syscall)
> li.d tp, ~_THREAD_MASK
> and tp, tp, sp
>
> + /* Syscall number held in a7, we can store it in TI_SYSCALL. */
> + LONG_S a7, tp, TI_SYSCALL
> +
> move a0, sp
> bl do_syscall
>
> - RESTORE_ALL_AND_RET
> + /*
> + * Syscall number held in a7 which is stored in TI_SYSCALL.
> + * clone and clone3 call RESTORE_STATIC_SOME_SP_AND_RET.
> + * rt_sigreturn and rt_sigsuspend call RESTORE_TEMP_SOME_SP_AND_RET.
> + * The other syscalls call RESTORE_SOME_SP_AND_RET.
> + */
> + LONG_L t3, tp, TI_SYSCALL
> + li.w t4, __NR_clone
> + beq t3, t4, 3f
> + li.w t4, __NR_clone3
> + beq t3, t4, 3f
> + li.w t4, __NR_rt_sigreturn
> + beq t3, t4, 4f
> + li.w t4, __NR_rt_sigsuspend
> + beq t3, t4, 4f
> +
> + RESTORE_SOME_SP_AND_RET
> +3:
> + RESTORE_STATIC_SOME_SP_AND_RET
> +4:
> + RESTORE_TEMP_SOME_SP_AND_RET
> SYM_FUNC_END(handle_syscall)
>
> SYM_CODE_START(ret_from_fork)
> --
> 2.1.0
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls
2022-06-16 12:52 ` Huacai Chen
@ 2022-06-16 13:36 ` Tiezhu Yang
0 siblings, 0 replies; 6+ messages in thread
From: Tiezhu Yang @ 2022-06-16 13:36 UTC (permalink / raw)
To: Huacai Chen; +Cc: WANG Xuerui, Xuefeng Li, Jianmin Lv, Jun Yi, Rui Wang, LKML
On 06/16/2022 08:52 PM, Huacai Chen wrote:
> Hi, Tiezhu,
>
> On Wed, Jun 15, 2022 at 6:29 PM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>>
>> In handle_syscall, it is unnecessary to call RESTORE_ALL_AND_RET
>> for all syscalls.
>>
>> (1) clone and clone3 call RESTORE_STATIC_SOME_SP_AND_RET.
>> (2) rt_sigreturn and rt_sigsuspend call RESTORE_TEMP_SOME_SP_AND_RET.
> I doubt that you have not tested, it is obvious that rt_sigreturn need
> RESTORE_ALL.
Yes, your guess is right. :)
I have no workable machine to update the upstream kernel now, only
compile test.
I will try my best to find a workable machine and then test it,
there is a good test case written by Rui Wang [1].
This patch is RFC, any comments will be much appreciated, thank you.
[1] https://github.com/hevz/sigaction-test
Thanks,
Tiezhu
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-06-16 13:36 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-15 10:29 [RFC PATCH v3 0/3] LoongArch: Modify handle_syscall Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 1/3] LoongArch: Add TI_SYSCALL in output_thread_info_defines() Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 2/3] LoongArch: Only clone and clone3 need to call SAVE_STATIC Tiezhu Yang
2022-06-15 10:29 ` [RFC PATCH v3 3/3] LoongArch: No need to call RESTORE_ALL_AND_RET for all syscalls Tiezhu Yang
2022-06-16 12:52 ` Huacai Chen
2022-06-16 13:36 ` Tiezhu Yang
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 NNTP newsgroup(s).