[1/2] LoongArch: kdump: Add the same binary implementation

Message ID 1677048791-26951-2-git-send-email-tangyouling@loongson.cn
State New
Headers
Series LoongArch: Add the same binary implementation for kdump |

Commit Message

Youling Tang Feb. 22, 2023, 6:53 a.m. UTC
  This feature depends on the relocation function, so the relocation configuration
CONFIG_RELOCATABLE will be enabled.

Add the same set of binary implementations for kdump, and then no longer need to
compile two kernels (the production kernel and the capture kernel share the same
binary).

CONFIG_CRASH_DUMP is enabled by default (CONFIG_RELOCATABLE is also enabled).

Signed-off-by: Youling Tang <tangyouling@loongson.cn>
---
 arch/loongarch/Kconfig                     | 12 +-----------
 arch/loongarch/Makefile                    |  4 ----
 arch/loongarch/configs/loongson3_defconfig |  1 +
 arch/loongarch/include/asm/addrspace.h     |  2 ++
 arch/loongarch/kernel/head.S               |  2 +-
 5 files changed, 5 insertions(+), 16 deletions(-)
  

Comments

WANG Xuerui Feb. 22, 2023, 9:05 a.m. UTC | #1
On 2023/2/22 14:53, Youling Tang wrote:
> This feature depends on the relocation function, so the relocation configuration
> CONFIG_RELOCATABLE will be enabled.

In general try to describe things briefly: "This depends on the kernel 
being relocatable" would be enough in this case.

> 
> Add the same set of binary implementations for kdump, and then no longer need to
> compile two kernels (the production kernel and the capture kernel share the same
> binary).

Sorry but what do you mean by "same set of binary implementation", 
where's the "first set of binary implementation"?

Judging from the patch content, I guess it's kinda wrong translation, 
and what you actually mean is something like "enable using the same 
image for crashkernel"?

> 
> CONFIG_CRASH_DUMP is enabled by default (CONFIG_RELOCATABLE is also enabled).

No it's not: you didn't add "default y" anywhere. What you actually did 
is to enable it *in the defconfig*. And the latter part about 
CONFIG_RELOCATABLE shouldn't be necessary, it's implementation detail 
after all, and the users likely don't have to be aware of it.

Better reword a little bit, like "Also enable CONFIG_CRASH_DUMP in 
loongson3_defconfig".

> 
> Signed-off-by: Youling Tang <tangyouling@loongson.cn>
> ---
>   arch/loongarch/Kconfig                     | 12 +-----------
>   arch/loongarch/Makefile                    |  4 ----
>   arch/loongarch/configs/loongson3_defconfig |  1 +
>   arch/loongarch/include/asm/addrspace.h     |  2 ++
>   arch/loongarch/kernel/head.S               |  2 +-
>   5 files changed, 5 insertions(+), 16 deletions(-)
> 
> diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
> index ab4c2ab146ab..84f220273296 100644
> --- a/arch/loongarch/Kconfig
> +++ b/arch/loongarch/Kconfig
> @@ -502,6 +502,7 @@ config KEXEC
>   
>   config CRASH_DUMP
>   	bool "Build kdump crash kernel"
> +	select RELOCATABLE
>   	help
>   	  Generate crash dump after being started by kexec. This should
>   	  be normally only set in special crash dump kernels which are
> @@ -511,17 +512,6 @@ config CRASH_DUMP
>   
>   	  For more details see Documentation/admin-guide/kdump/kdump.rst
>   
> -config PHYSICAL_START
> -	hex "Physical address where the kernel is loaded"
> -	default "0x90000000a0000000"
> -	depends on CRASH_DUMP
> -	help
> -	  This gives the XKPRANGE address where the kernel is loaded.
> -	  If you plan to use kernel for capturing the crash dump change
> -	  this value to start of the reserved region (the "X" value as
> -	  specified in the "crashkernel=YM@XM" command line boot parameter
> -	  passed to the panic-ed kernel).
> -
>   config RELOCATABLE
>   	bool "Relocatable kernel"
>   	help
> diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
> index 2aba6ff30436..8304fed7aa42 100644
> --- a/arch/loongarch/Makefile
> +++ b/arch/loongarch/Makefile
> @@ -79,11 +79,7 @@ endif
>   cflags-y += -ffreestanding
>   cflags-y += $(call cc-option, -mno-check-zero-division)
>   
> -ifndef CONFIG_PHYSICAL_START
>   load-y		= 0x9000000000200000
> -else
> -load-y		= $(CONFIG_PHYSICAL_START)
> -endif
>   bootvars-y	= VMLINUX_LOAD_ADDRESS=$(load-y)

Both load-y and VMLINUX_LOAD_ADDRESS are kinda LoongArch-ism (stemming 
from similar usage in arch/mips apparently), so why not just drop load-y 
and fold the constant into the bootvars-y definition? So we have one 
piece of "special" definition instead of two.

>   
>   drivers-$(CONFIG_PCI)		+= arch/loongarch/pci/
> diff --git a/arch/loongarch/configs/loongson3_defconfig b/arch/loongarch/configs/loongson3_defconfig
> index cb52774c80e8..7885f6e5de93 100644
> --- a/arch/loongarch/configs/loongson3_defconfig
> +++ b/arch/loongarch/configs/loongson3_defconfig
> @@ -48,6 +48,7 @@ CONFIG_HOTPLUG_CPU=y
>   CONFIG_NR_CPUS=64
>   CONFIG_NUMA=y
>   CONFIG_KEXEC=y
> +CONFIG_CRASH_DUMP=y
>   CONFIG_SUSPEND=y
>   CONFIG_HIBERNATION=y
>   CONFIG_ACPI=y
> diff --git a/arch/loongarch/include/asm/addrspace.h b/arch/loongarch/include/asm/addrspace.h
> index d342935e5a72..4edcb3c21cf5 100644
> --- a/arch/loongarch/include/asm/addrspace.h
> +++ b/arch/loongarch/include/asm/addrspace.h
> @@ -125,4 +125,6 @@ extern unsigned long vm_map_base;
>   #define ISA_IOSIZE	SZ_16K
>   #define IO_SPACE_LIMIT	(PCI_IOSIZE - 1)
>   
> +#define PHYS_LINK_ADDR	PHYSADDR(VMLINUX_LOAD_ADDRESS)
> +
>   #endif /* _ASM_ADDRSPACE_H */
> diff --git a/arch/loongarch/kernel/head.S b/arch/loongarch/kernel/head.S
> index b12f459ad73a..57962ff08f6d 100644
> --- a/arch/loongarch/kernel/head.S
> +++ b/arch/loongarch/kernel/head.S
> @@ -24,7 +24,7 @@ _head:
>   	.org	0x8
>   	.dword	kernel_entry		/* Kernel entry point */
>   	.dword	_end - _text		/* Kernel image effective size */
> -	.quad	0			/* Kernel image load offset from start of RAM */
> +	.quad	PHYS_LINK_ADDR		/* Kernel image load offset from start of RAM */
>   	.org	0x38			/* 0x20 ~ 0x37 reserved */
>   	.long	LINUX_PE_MAGIC
>   	.long	pe_header - _head	/* Offset to the PE header */
  
Youling Tang Feb. 22, 2023, 9:54 a.m. UTC | #2
Hi, Xuerui

On 02/22/2023 05:05 PM, WANG Xuerui wrote:
> On 2023/2/22 14:53, Youling Tang wrote:
>> This feature depends on the relocation function, so the relocation
>> configuration
>> CONFIG_RELOCATABLE will be enabled.
>
> In general try to describe things briefly: "This depends on the kernel
> being relocatable" would be enough in this case.
OK.

>
>>
>> Add the same set of binary implementations for kdump, and then no
>> longer need to
>> compile two kernels (the production kernel and the capture kernel
>> share the same
>> binary).
>
> Sorry but what do you mean by "same set of binary implementation",
> where's the "first set of binary implementation"?

kdump requires two kernels, the production kernel and the capture
kernel, which made the final link address different through different
configuration options in the previous implementation. Now it is
possible to share a kernel, and after entering the capture kernel, it
can be corrected by relocation.

>
> Judging from the patch content, I guess it's kinda wrong translation,
> and what you actually mean is something like "enable using the same
> image for crashkernel"?

Or "LoongArch: kdump: Add support for using a single image" ?

>
>>
>> CONFIG_CRASH_DUMP is enabled by default (CONFIG_RELOCATABLE is also
>> enabled).
>
> No it's not: you didn't add "default y" anywhere. What you actually did
> is to enable it *in the defconfig*. And the latter part about
> CONFIG_RELOCATABLE shouldn't be necessary, it's implementation detail
> after all, and the users likely don't have to be aware of it.
>
> Better reword a little bit, like "Also enable CONFIG_CRASH_DUMP in
> loongson3_defconfig".

Thanks, I'll rewrite the commit message.

>
>>
>> Signed-off-by: Youling Tang <tangyouling@loongson.cn>
>> ---
>>   arch/loongarch/Kconfig                     | 12 +-----------
>>   arch/loongarch/Makefile                    |  4 ----
>>   arch/loongarch/configs/loongson3_defconfig |  1 +
>>   arch/loongarch/include/asm/addrspace.h     |  2 ++
>>   arch/loongarch/kernel/head.S               |  2 +-
>>   5 files changed, 5 insertions(+), 16 deletions(-)
>>
>> diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
>> index ab4c2ab146ab..84f220273296 100644
>> --- a/arch/loongarch/Kconfig
>> +++ b/arch/loongarch/Kconfig
>> @@ -502,6 +502,7 @@ config KEXEC
>>     config CRASH_DUMP
>>       bool "Build kdump crash kernel"
>> +    select RELOCATABLE
>>       help
>>         Generate crash dump after being started by kexec. This should
>>         be normally only set in special crash dump kernels which are
>> @@ -511,17 +512,6 @@ config CRASH_DUMP
>>           For more details see Documentation/admin-guide/kdump/kdump.rst
>>   -config PHYSICAL_START
>> -    hex "Physical address where the kernel is loaded"
>> -    default "0x90000000a0000000"
>> -    depends on CRASH_DUMP
>> -    help
>> -      This gives the XKPRANGE address where the kernel is loaded.
>> -      If you plan to use kernel for capturing the crash dump change
>> -      this value to start of the reserved region (the "X" value as
>> -      specified in the "crashkernel=YM@XM" command line boot parameter
>> -      passed to the panic-ed kernel).
>> -
>>   config RELOCATABLE
>>       bool "Relocatable kernel"
>>       help
>> diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
>> index 2aba6ff30436..8304fed7aa42 100644
>> --- a/arch/loongarch/Makefile
>> +++ b/arch/loongarch/Makefile
>> @@ -79,11 +79,7 @@ endif
>>   cflags-y += -ffreestanding
>>   cflags-y += $(call cc-option, -mno-check-zero-division)
>>   -ifndef CONFIG_PHYSICAL_START
>>   load-y        = 0x9000000000200000
>> -else
>> -load-y        = $(CONFIG_PHYSICAL_START)
>> -endif
>>   bootvars-y    = VMLINUX_LOAD_ADDRESS=$(load-y)
>
> Both load-y and VMLINUX_LOAD_ADDRESS are kinda LoongArch-ism (stemming
> from similar usage in arch/mips apparently), so why not just drop load-y
> and fold the constant into the bootvars-y definition? So we have one
> piece of "special" definition instead of two.

This series of patches will not modify it, perhaps it can be submitted
and discussed separately later.

Thanks,
Youling.

>
>>     drivers-$(CONFIG_PCI)        += arch/loongarch/pci/
>> diff --git a/arch/loongarch/configs/loongson3_defconfig
>> b/arch/loongarch/configs/loongson3_defconfig
>> index cb52774c80e8..7885f6e5de93 100644
>> --- a/arch/loongarch/configs/loongson3_defconfig
>> +++ b/arch/loongarch/configs/loongson3_defconfig
>> @@ -48,6 +48,7 @@ CONFIG_HOTPLUG_CPU=y
>>   CONFIG_NR_CPUS=64
>>   CONFIG_NUMA=y
>>   CONFIG_KEXEC=y
>> +CONFIG_CRASH_DUMP=y
>>   CONFIG_SUSPEND=y
>>   CONFIG_HIBERNATION=y
>>   CONFIG_ACPI=y
>> diff --git a/arch/loongarch/include/asm/addrspace.h
>> b/arch/loongarch/include/asm/addrspace.h
>> index d342935e5a72..4edcb3c21cf5 100644
>> --- a/arch/loongarch/include/asm/addrspace.h
>> +++ b/arch/loongarch/include/asm/addrspace.h
>> @@ -125,4 +125,6 @@ extern unsigned long vm_map_base;
>>   #define ISA_IOSIZE    SZ_16K
>>   #define IO_SPACE_LIMIT    (PCI_IOSIZE - 1)
>>   +#define PHYS_LINK_ADDR    PHYSADDR(VMLINUX_LOAD_ADDRESS)
>> +
>>   #endif /* _ASM_ADDRSPACE_H */
>> diff --git a/arch/loongarch/kernel/head.S b/arch/loongarch/kernel/head.S
>> index b12f459ad73a..57962ff08f6d 100644
>> --- a/arch/loongarch/kernel/head.S
>> +++ b/arch/loongarch/kernel/head.S
>> @@ -24,7 +24,7 @@ _head:
>>       .org    0x8
>>       .dword    kernel_entry        /* Kernel entry point */
>>       .dword    _end - _text        /* Kernel image effective size */
>> -    .quad    0            /* Kernel image load offset from start of
>> RAM */
>> +    .quad    PHYS_LINK_ADDR        /* Kernel image load offset from
>> start of RAM */
>>       .org    0x38            /* 0x20 ~ 0x37 reserved */
>>       .long    LINUX_PE_MAGIC
>>       .long    pe_header - _head    /* Offset to the PE header */
>
  
WANG Xuerui Feb. 22, 2023, 10:12 a.m. UTC | #3
On 2023/2/22 17:54, Youling Tang wrote:
> Hi, Xuerui
> 
> On 02/22/2023 05:05 PM, WANG Xuerui wrote:
>> On 2023/2/22 14:53, Youling Tang wrote:
>>> This feature depends on the relocation function, so the relocation
>>> configuration
>>> CONFIG_RELOCATABLE will be enabled.
>>
>> In general try to describe things briefly: "This depends on the kernel
>> being relocatable" would be enough in this case.
> OK.
> 
>>
>>>
>>> Add the same set of binary implementations for kdump, and then no
>>> longer need to
>>> compile two kernels (the production kernel and the capture kernel
>>> share the same
>>> binary).
>>
>> Sorry but what do you mean by "same set of binary implementation",
>> where's the "first set of binary implementation"?
> 
> kdump requires two kernels, the production kernel and the capture
> kernel, which made the final link address different through different
> configuration options in the previous implementation. Now it is
> possible to share a kernel, and after entering the capture kernel, it
> can be corrected by relocation.
> 
>>
>> Judging from the patch content, I guess it's kinda wrong translation,
>> and what you actually mean is something like "enable using the same
>> image for crashkernel"?
> 
> Or "LoongArch: kdump: Add support for using a single image" ?
> 

Whatever you prefer, literally anything else would be better than the 
original. I don't have strong opinion over this.

>>
>>>
>>> CONFIG_CRASH_DUMP is enabled by default (CONFIG_RELOCATABLE is also
>>> enabled).
>>
>> No it's not: you didn't add "default y" anywhere. What you actually did
>> is to enable it *in the defconfig*. And the latter part about
>> CONFIG_RELOCATABLE shouldn't be necessary, it's implementation detail
>> after all, and the users likely don't have to be aware of it.
>>
>> Better reword a little bit, like "Also enable CONFIG_CRASH_DUMP in
>> loongson3_defconfig".
> 
> Thanks, I'll rewrite the commit message.
> 
>>
>>>
>>> Signed-off-by: Youling Tang <tangyouling@loongson.cn>
>>> ---
>>>   arch/loongarch/Kconfig                     | 12 +-----------
>>>   arch/loongarch/Makefile                    |  4 ----
>>>   arch/loongarch/configs/loongson3_defconfig |  1 +
>>>   arch/loongarch/include/asm/addrspace.h     |  2 ++
>>>   arch/loongarch/kernel/head.S               |  2 +-
>>>   5 files changed, 5 insertions(+), 16 deletions(-)
>>>
>>> diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
>>> index ab4c2ab146ab..84f220273296 100644
>>> --- a/arch/loongarch/Kconfig
>>> +++ b/arch/loongarch/Kconfig
>>> @@ -502,6 +502,7 @@ config KEXEC
>>>     config CRASH_DUMP
>>>       bool "Build kdump crash kernel"
>>> +    select RELOCATABLE
>>>       help
>>>         Generate crash dump after being started by kexec. This should
>>>         be normally only set in special crash dump kernels which are
>>> @@ -511,17 +512,6 @@ config CRASH_DUMP
>>>           For more details see Documentation/admin-guide/kdump/kdump.rst
>>>   -config PHYSICAL_START
>>> -    hex "Physical address where the kernel is loaded"
>>> -    default "0x90000000a0000000"
>>> -    depends on CRASH_DUMP
>>> -    help
>>> -      This gives the XKPRANGE address where the kernel is loaded.
>>> -      If you plan to use kernel for capturing the crash dump change
>>> -      this value to start of the reserved region (the "X" value as
>>> -      specified in the "crashkernel=YM@XM" command line boot parameter
>>> -      passed to the panic-ed kernel).
>>> -
>>>   config RELOCATABLE
>>>       bool "Relocatable kernel"
>>>       help
>>> diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
>>> index 2aba6ff30436..8304fed7aa42 100644
>>> --- a/arch/loongarch/Makefile
>>> +++ b/arch/loongarch/Makefile
>>> @@ -79,11 +79,7 @@ endif
>>>   cflags-y += -ffreestanding
>>>   cflags-y += $(call cc-option, -mno-check-zero-division)
>>>   -ifndef CONFIG_PHYSICAL_START
>>>   load-y        = 0x9000000000200000
>>> -else
>>> -load-y        = $(CONFIG_PHYSICAL_START)
>>> -endif
>>>   bootvars-y    = VMLINUX_LOAD_ADDRESS=$(load-y)
>>
>> Both load-y and VMLINUX_LOAD_ADDRESS are kinda LoongArch-ism (stemming
>> from similar usage in arch/mips apparently), so why not just drop load-y
>> and fold the constant into the bootvars-y definition? So we have one
>> piece of "special" definition instead of two.
> 
> This series of patches will not modify it, perhaps it can be submitted
> and discussed separately later.

I mean since you're already touching this part, you might be better off 
just simplify along the way because it's used by nowhere else. But 
splitting commits would be indeed cleaner and of course you can add such 
a followup commit in the next revision.
  

Patch

diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
index ab4c2ab146ab..84f220273296 100644
--- a/arch/loongarch/Kconfig
+++ b/arch/loongarch/Kconfig
@@ -502,6 +502,7 @@  config KEXEC
 
 config CRASH_DUMP
 	bool "Build kdump crash kernel"
+	select RELOCATABLE
 	help
 	  Generate crash dump after being started by kexec. This should
 	  be normally only set in special crash dump kernels which are
@@ -511,17 +512,6 @@  config CRASH_DUMP
 
 	  For more details see Documentation/admin-guide/kdump/kdump.rst
 
-config PHYSICAL_START
-	hex "Physical address where the kernel is loaded"
-	default "0x90000000a0000000"
-	depends on CRASH_DUMP
-	help
-	  This gives the XKPRANGE address where the kernel is loaded.
-	  If you plan to use kernel for capturing the crash dump change
-	  this value to start of the reserved region (the "X" value as
-	  specified in the "crashkernel=YM@XM" command line boot parameter
-	  passed to the panic-ed kernel).
-
 config RELOCATABLE
 	bool "Relocatable kernel"
 	help
diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
index 2aba6ff30436..8304fed7aa42 100644
--- a/arch/loongarch/Makefile
+++ b/arch/loongarch/Makefile
@@ -79,11 +79,7 @@  endif
 cflags-y += -ffreestanding
 cflags-y += $(call cc-option, -mno-check-zero-division)
 
-ifndef CONFIG_PHYSICAL_START
 load-y		= 0x9000000000200000
-else
-load-y		= $(CONFIG_PHYSICAL_START)
-endif
 bootvars-y	= VMLINUX_LOAD_ADDRESS=$(load-y)
 
 drivers-$(CONFIG_PCI)		+= arch/loongarch/pci/
diff --git a/arch/loongarch/configs/loongson3_defconfig b/arch/loongarch/configs/loongson3_defconfig
index cb52774c80e8..7885f6e5de93 100644
--- a/arch/loongarch/configs/loongson3_defconfig
+++ b/arch/loongarch/configs/loongson3_defconfig
@@ -48,6 +48,7 @@  CONFIG_HOTPLUG_CPU=y
 CONFIG_NR_CPUS=64
 CONFIG_NUMA=y
 CONFIG_KEXEC=y
+CONFIG_CRASH_DUMP=y
 CONFIG_SUSPEND=y
 CONFIG_HIBERNATION=y
 CONFIG_ACPI=y
diff --git a/arch/loongarch/include/asm/addrspace.h b/arch/loongarch/include/asm/addrspace.h
index d342935e5a72..4edcb3c21cf5 100644
--- a/arch/loongarch/include/asm/addrspace.h
+++ b/arch/loongarch/include/asm/addrspace.h
@@ -125,4 +125,6 @@  extern unsigned long vm_map_base;
 #define ISA_IOSIZE	SZ_16K
 #define IO_SPACE_LIMIT	(PCI_IOSIZE - 1)
 
+#define PHYS_LINK_ADDR	PHYSADDR(VMLINUX_LOAD_ADDRESS)
+
 #endif /* _ASM_ADDRSPACE_H */
diff --git a/arch/loongarch/kernel/head.S b/arch/loongarch/kernel/head.S
index b12f459ad73a..57962ff08f6d 100644
--- a/arch/loongarch/kernel/head.S
+++ b/arch/loongarch/kernel/head.S
@@ -24,7 +24,7 @@  _head:
 	.org	0x8
 	.dword	kernel_entry		/* Kernel entry point */
 	.dword	_end - _text		/* Kernel image effective size */
-	.quad	0			/* Kernel image load offset from start of RAM */
+	.quad	PHYS_LINK_ADDR		/* Kernel image load offset from start of RAM */
 	.org	0x38			/* 0x20 ~ 0x37 reserved */
 	.long	LINUX_PE_MAGIC
 	.long	pe_header - _head	/* Offset to the PE header */