[v2,2/3] vmcore: allow fadump to export vmcore even if is_kdump_kernel() is false

Currently, is_kdump_kernel() returns true when elfcorehdr_addr is set.
While elfcorehdr_addr is set for kexec based kernel dump mechanism,
alternate dump capturing methods like fadump [1] also set it to export
the vmcore. Since, is_kdump_kernel() is used to restrict resources in
crash dump capture kernel and such restrictions are not desirable for
fadump, allow is_kdump_kernel() to be defined differently for fadump
case. With that change, include is_fadump_active() check in functions
is_vmcore_usable() & vmcore_unusable() to be able to export vmcore for
fadump case too.

[1] https://docs.kernel.org/powerpc/firmware-assisted-dump.html

Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
---

Changes in v2:
* is_fadump_active() check added to is_vmcore_usable() as suggested
  by Baoquan.


 include/linux/crash_dump.h | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

Hari Bathini <hbathini@linux.ibm.com> writes:
> Currently, is_kdump_kernel() returns true when elfcorehdr_addr is set.
> While elfcorehdr_addr is set for kexec based kernel dump mechanism,
> alternate dump capturing methods like fadump [1] also set it to export
> the vmcore. Since, is_kdump_kernel() is used to restrict resources in
> crash dump capture kernel and such restrictions are not desirable for
> fadump, allow is_kdump_kernel() to be defined differently for fadump
> case. With that change, include is_fadump_active() check in functions
> is_vmcore_usable() & vmcore_unusable() to be able to export vmcore for
> fadump case too.
...
> diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
> index 0f3a656293b0..de8a9fabfb6f 100644
> --- a/include/linux/crash_dump.h
> +++ b/include/linux/crash_dump.h
> @@ -50,6 +50,7 @@ void vmcore_cleanup(void);
>  #define vmcore_elf64_check_arch(x) (elf_check_arch(x) || vmcore_elf_check_arch_cross(x))
>  #endif
>  
> +#ifndef is_kdump_kernel
>  /*
>   * is_kdump_kernel() checks whether this kernel is booting after a panic of
>   * previous kernel or not. This is determined by checking if previous kernel
> @@ -64,6 +65,19 @@ static inline bool is_kdump_kernel(void)
>  {
>  	return elfcorehdr_addr != ELFCORE_ADDR_MAX;
>  }
> +#endif
> +
> +#ifndef is_fadump_active
> +/*
> + * If f/w assisted dump capturing mechanism (fadump), instead of kexec based
> + * dump capturing mechanism (kdump) is exporting the vmcore, then this function
> + * will be defined in arch specific code to return true, when appropriate.
> + */
> +static inline bool is_fadump_active(void)
> +{
> +	return false;
> +}
> +#endif
>  
>  /* is_vmcore_usable() checks if the kernel is booting after a panic and
>   * the vmcore region is usable.
> @@ -75,7 +89,8 @@ static inline bool is_kdump_kernel(void)
>  
>  static inline int is_vmcore_usable(void)
>  {
> -	return is_kdump_kernel() && elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
> +	return (is_kdump_kernel() || is_fadump_active())
> +		&& elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
>  }
>  
>  /* vmcore_unusable() marks the vmcore as unusable,
> @@ -84,7 +99,7 @@ static inline int is_vmcore_usable(void)
>  
>  static inline void vmcore_unusable(void)
>  {
> -	if (is_kdump_kernel())
> +	if (is_kdump_kernel() || is_fadump_active())
>  		elfcorehdr_addr = ELFCORE_ADDR_ERR;
>  }

I think it would be cleaner to decouple is_vmcore_usable() and
vmcore_usable() from is_kdump_kernel().

ie, make them operate solely based on the value of elforehdr_addr:

static inline int is_vmcore_usable(void)
{
	elfcorehdr_addr != ELFCORE_ADDR_ERR && \
		elfcorehdr_addr != ELFCORE_ADDR_MAX;
}

static inline void vmcore_unusable(void)
{
	elfcorehdr_addr = ELFCORE_ADDR_ERR;
}


Then all we need on powerpc is a way to override is_kdump_kernel().

cheers

On 09/11/23 at 05:13pm, Michael Ellerman wrote:
> Hari Bathini <hbathini@linux.ibm.com> writes:
> > Currently, is_kdump_kernel() returns true when elfcorehdr_addr is set.
> > While elfcorehdr_addr is set for kexec based kernel dump mechanism,
> > alternate dump capturing methods like fadump [1] also set it to export
> > the vmcore. Since, is_kdump_kernel() is used to restrict resources in
> > crash dump capture kernel and such restrictions are not desirable for
> > fadump, allow is_kdump_kernel() to be defined differently for fadump
> > case. With that change, include is_fadump_active() check in functions
> > is_vmcore_usable() & vmcore_unusable() to be able to export vmcore for
> > fadump case too.
> ...
> > diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
> > index 0f3a656293b0..de8a9fabfb6f 100644
> > --- a/include/linux/crash_dump.h
> > +++ b/include/linux/crash_dump.h
> > @@ -50,6 +50,7 @@ void vmcore_cleanup(void);
> >  #define vmcore_elf64_check_arch(x) (elf_check_arch(x) || vmcore_elf_check_arch_cross(x))
> >  #endif
> >  
> > +#ifndef is_kdump_kernel
> >  /*
> >   * is_kdump_kernel() checks whether this kernel is booting after a panic of
> >   * previous kernel or not. This is determined by checking if previous kernel
> > @@ -64,6 +65,19 @@ static inline bool is_kdump_kernel(void)
> >  {
> >  	return elfcorehdr_addr != ELFCORE_ADDR_MAX;
> >  }
> > +#endif
> > +
> > +#ifndef is_fadump_active
> > +/*
> > + * If f/w assisted dump capturing mechanism (fadump), instead of kexec based
> > + * dump capturing mechanism (kdump) is exporting the vmcore, then this function
> > + * will be defined in arch specific code to return true, when appropriate.
> > + */
> > +static inline bool is_fadump_active(void)
> > +{
> > +	return false;
> > +}
> > +#endif
> >  
> >  /* is_vmcore_usable() checks if the kernel is booting after a panic and
> >   * the vmcore region is usable.
> > @@ -75,7 +89,8 @@ static inline bool is_kdump_kernel(void)
> >  
> >  static inline int is_vmcore_usable(void)
> >  {
> > -	return is_kdump_kernel() && elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
> > +	return (is_kdump_kernel() || is_fadump_active())
> > +		&& elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
> >  }
> >  
> >  /* vmcore_unusable() marks the vmcore as unusable,
> > @@ -84,7 +99,7 @@ static inline int is_vmcore_usable(void)
> >  
> >  static inline void vmcore_unusable(void)
> >  {
> > -	if (is_kdump_kernel())
> > +	if (is_kdump_kernel() || is_fadump_active())
> >  		elfcorehdr_addr = ELFCORE_ADDR_ERR;
> >  }
> 
> I think it would be cleaner to decouple is_vmcore_usable() and
> vmcore_usable() from is_kdump_kernel().
> 
> ie, make them operate solely based on the value of elforehdr_addr:
> 
> static inline int is_vmcore_usable(void)
> {
> 	elfcorehdr_addr != ELFCORE_ADDR_ERR && \
> 		elfcorehdr_addr != ELFCORE_ADDR_MAX;

Agree. I fell into the blind corner of thinking earlier. Above change
is better.

> }
> 
> static inline void vmcore_unusable(void)
> {
> 	elfcorehdr_addr = ELFCORE_ADDR_ERR;
> }
> 
> 
> Then all we need on powerpc is a way to override is_kdump_kernel().
> 
> cheers
>

On 11/09/23 4:01 pm, Baoquan He wrote:
> On 09/11/23 at 05:13pm, Michael Ellerman wrote:
>> Hari Bathini <hbathini@linux.ibm.com> writes:
>>> Currently, is_kdump_kernel() returns true when elfcorehdr_addr is set.
>>> While elfcorehdr_addr is set for kexec based kernel dump mechanism,
>>> alternate dump capturing methods like fadump [1] also set it to export
>>> the vmcore. Since, is_kdump_kernel() is used to restrict resources in
>>> crash dump capture kernel and such restrictions are not desirable for
>>> fadump, allow is_kdump_kernel() to be defined differently for fadump
>>> case. With that change, include is_fadump_active() check in functions
>>> is_vmcore_usable() & vmcore_unusable() to be able to export vmcore for
>>> fadump case too.
>> ...
>>> diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
>>> index 0f3a656293b0..de8a9fabfb6f 100644
>>> --- a/include/linux/crash_dump.h
>>> +++ b/include/linux/crash_dump.h
>>> @@ -50,6 +50,7 @@ void vmcore_cleanup(void);
>>>   #define vmcore_elf64_check_arch(x) (elf_check_arch(x) || vmcore_elf_check_arch_cross(x))
>>>   #endif
>>>   
>>> +#ifndef is_kdump_kernel
>>>   /*
>>>    * is_kdump_kernel() checks whether this kernel is booting after a panic of
>>>    * previous kernel or not. This is determined by checking if previous kernel
>>> @@ -64,6 +65,19 @@ static inline bool is_kdump_kernel(void)
>>>   {
>>>   	return elfcorehdr_addr != ELFCORE_ADDR_MAX;
>>>   }
>>> +#endif
>>> +
>>> +#ifndef is_fadump_active
>>> +/*
>>> + * If f/w assisted dump capturing mechanism (fadump), instead of kexec based
>>> + * dump capturing mechanism (kdump) is exporting the vmcore, then this function
>>> + * will be defined in arch specific code to return true, when appropriate.
>>> + */
>>> +static inline bool is_fadump_active(void)
>>> +{
>>> +	return false;
>>> +}
>>> +#endif
>>>   
>>>   /* is_vmcore_usable() checks if the kernel is booting after a panic and
>>>    * the vmcore region is usable.
>>> @@ -75,7 +89,8 @@ static inline bool is_kdump_kernel(void)
>>>   
>>>   static inline int is_vmcore_usable(void)
>>>   {
>>> -	return is_kdump_kernel() && elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
>>> +	return (is_kdump_kernel() || is_fadump_active())
>>> +		&& elfcorehdr_addr != ELFCORE_ADDR_ERR ? 1 : 0;
>>>   }
>>>   
>>>   /* vmcore_unusable() marks the vmcore as unusable,
>>> @@ -84,7 +99,7 @@ static inline int is_vmcore_usable(void)
>>>   
>>>   static inline void vmcore_unusable(void)
>>>   {
>>> -	if (is_kdump_kernel())
>>> +	if (is_kdump_kernel() || is_fadump_active())
>>>   		elfcorehdr_addr = ELFCORE_ADDR_ERR;
>>>   }
>>
>> I think it would be cleaner to decouple is_vmcore_usable() and
>> vmcore_usable() from is_kdump_kernel().
>>
>> ie, make them operate solely based on the value of elforehdr_addr:
>>
>> static inline int is_vmcore_usable(void)
>> {
>> 	elfcorehdr_addr != ELFCORE_ADDR_ERR && \
>> 		elfcorehdr_addr != ELFCORE_ADDR_MAX;
> 
> Agree. I fell into the blind corner of thinking earlier. Above change
> is better.

Thanks for the reviews.
Posted v3.

- Hari