Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
reservations") introduced early_reserve_memory() to do all needed
initial memblock_reserve() calls in one function. Unfortunately the
call of early_reserve_memory() is done too late for Xen dom0, as in
some cases a Xen hook called by e820__memory_setup() will need those
memory reservations to have happened already.

Move the call of early_reserve_memory() to the beginning of
setup_arch() in order to avoid such problems.

Cc: stable@vger.kernel.org
Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
---
 arch/x86/kernel/setup.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 79f164141116..f369c51ec580 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
 
 void __init setup_arch(char **cmdline_p)
 {
+	/*
+	 * Do some memory reservations *before* memory is added to
+	 * memblock, so memblock allocations won't overwrite it.
+	 * Do it after early param, so we could get (unlikely) panic from
+	 * serial.
+	 *
+	 * After this point everything still needed from the boot loader or
+	 * firmware or kernel text should be early reserved or marked not
+	 * RAM in e820. All other memory is free game.
+	 */
+	early_reserve_memory();
+
 #ifdef CONFIG_X86_32
 	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
 
@@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
 
 	parse_early_param();
 
-	/*
-	 * Do some memory reservations *before* memory is added to
-	 * memblock, so memblock allocations won't overwrite it.
-	 * Do it after early param, so we could get (unlikely) panic from
-	 * serial.
-	 *
-	 * After this point everything still needed from the boot loader or
-	 * firmware or kernel text should be early reserved or marked not
-	 * RAM in e820. All other memory is free game.
-	 */
-	early_reserve_memory();
-
 #ifdef CONFIG_MEMORY_HOTPLUG
 	/*
 	 * Memory used by the kernel cannot be hot-removed because Linux
-- 
2.26.2

On 14.09.2021 11:41, Juergen Gross wrote:
> Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
> reservations") introduced early_reserve_memory() to do all needed
> initial memblock_reserve() calls in one function. Unfortunately the
> call of early_reserve_memory() is done too late for Xen dom0, as in
> some cases a Xen hook called by e820__memory_setup() will need those
> memory reservations to have happened already.
> 
> Move the call of early_reserve_memory() to the beginning of
> setup_arch() in order to avoid such problems.
> 
> Cc: stable@vger.kernel.org
> Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
> Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
> Signed-off-by: Juergen Gross <jgross@suse.com>
> ---
>  arch/x86/kernel/setup.c | 24 ++++++++++++------------
>  1 file changed, 12 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 79f164141116..f369c51ec580 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
>  
>  void __init setup_arch(char **cmdline_p)
>  {
> +	/*
> +	 * Do some memory reservations *before* memory is added to
> +	 * memblock, so memblock allocations won't overwrite it.
> +	 * Do it after early param, so we could get (unlikely) panic from
> +	 * serial.

Hmm, this part of the comment is not only stale now, but gets actively
undermined. No idea how likely such a panic() would be, and hence how
relevant it is to retain this particular property.

Jan

> +	 * After this point everything still needed from the boot loader or
> +	 * firmware or kernel text should be early reserved or marked not
> +	 * RAM in e820. All other memory is free game.
> +	 */
> +	early_reserve_memory();
> +
>  #ifdef CONFIG_X86_32
>  	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
>  
> @@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
>  
>  	parse_early_param();
>  
> -	/*
> -	 * Do some memory reservations *before* memory is added to
> -	 * memblock, so memblock allocations won't overwrite it.
> -	 * Do it after early param, so we could get (unlikely) panic from
> -	 * serial.
> -	 *
> -	 * After this point everything still needed from the boot loader or
> -	 * firmware or kernel text should be early reserved or marked not
> -	 * RAM in e820. All other memory is free game.
> -	 */
> -	early_reserve_memory();
> -
>  #ifdef CONFIG_MEMORY_HOTPLUG
>  	/*
>  	 * Memory used by the kernel cannot be hot-removed because Linux
> 

On 14.09.21 12:03, Jan Beulich wrote:
> On 14.09.2021 11:41, Juergen Gross wrote:
>> Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
>> reservations") introduced early_reserve_memory() to do all needed
>> initial memblock_reserve() calls in one function. Unfortunately the
>> call of early_reserve_memory() is done too late for Xen dom0, as in
>> some cases a Xen hook called by e820__memory_setup() will need those
>> memory reservations to have happened already.
>>
>> Move the call of early_reserve_memory() to the beginning of
>> setup_arch() in order to avoid such problems.
>>
>> Cc: stable@vger.kernel.org
>> Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
>> Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
>> ---
>>   arch/x86/kernel/setup.c | 24 ++++++++++++------------
>>   1 file changed, 12 insertions(+), 12 deletions(-)
>>
>> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
>> index 79f164141116..f369c51ec580 100644
>> --- a/arch/x86/kernel/setup.c
>> +++ b/arch/x86/kernel/setup.c
>> @@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
>>   
>>   void __init setup_arch(char **cmdline_p)
>>   {
>> +	/*
>> +	 * Do some memory reservations *before* memory is added to
>> +	 * memblock, so memblock allocations won't overwrite it.
>> +	 * Do it after early param, so we could get (unlikely) panic from
>> +	 * serial.
> 
> Hmm, this part of the comment is not only stale now, but gets actively
> undermined. No idea how likely such a panic() would be, and hence how
> relevant it is to retain this particular property.

Ah, right.

The alternative would be to split it up again. Let's let the x86
maintainers decide which way is the better one.


Juergen

> 
> Jan
> 
>> +	 * After this point everything still needed from the boot loader or
>> +	 * firmware or kernel text should be early reserved or marked not
>> +	 * RAM in e820. All other memory is free game.
>> +	 */
>> +	early_reserve_memory();
>> +
>>   #ifdef CONFIG_X86_32
>>   	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
>>   
>> @@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
>>   
>>   	parse_early_param();
>>   
>> -	/*
>> -	 * Do some memory reservations *before* memory is added to
>> -	 * memblock, so memblock allocations won't overwrite it.
>> -	 * Do it after early param, so we could get (unlikely) panic from
>> -	 * serial.
>> -	 *
>> -	 * After this point everything still needed from the boot loader or
>> -	 * firmware or kernel text should be early reserved or marked not
>> -	 * RAM in e820. All other memory is free game.
>> -	 */
>> -	early_reserve_memory();
>> -
>>   #ifdef CONFIG_MEMORY_HOTPLUG
>>   	/*
>>   	 * Memory used by the kernel cannot be hot-removed because Linux
>>
> 

You forgot to Cc Mike, lemme add him.

And drop stable@ too.

On Tue, Sep 14, 2021 at 01:06:22PM +0200, Juergen Gross wrote:
> On 14.09.21 12:03, Jan Beulich wrote:
> > On 14.09.2021 11:41, Juergen Gross wrote:
> > > Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
> > > reservations") introduced early_reserve_memory() to do all needed
> > > initial memblock_reserve() calls in one function. Unfortunately the
> > > call of early_reserve_memory() is done too late for Xen dom0, as in
> > > some cases a Xen hook called by e820__memory_setup() will need those
> > > memory reservations to have happened already.
> > > 
> > > Move the call of early_reserve_memory() to the beginning of
> > > setup_arch() in order to avoid such problems.
> > > 
> > > Cc: stable@vger.kernel.org
> > > Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
> > > Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
> > > Signed-off-by: Juergen Gross <jgross@suse.com>
> > > ---
> > >   arch/x86/kernel/setup.c | 24 ++++++++++++------------
> > >   1 file changed, 12 insertions(+), 12 deletions(-)
> > > 
> > > diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> > > index 79f164141116..f369c51ec580 100644
> > > --- a/arch/x86/kernel/setup.c
> > > +++ b/arch/x86/kernel/setup.c
> > > @@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
> > >   void __init setup_arch(char **cmdline_p)
> > >   {
> > > +	/*
> > > +	 * Do some memory reservations *before* memory is added to
> > > +	 * memblock, so memblock allocations won't overwrite it.
> > > +	 * Do it after early param, so we could get (unlikely) panic from
> > > +	 * serial.
> > 
> > Hmm, this part of the comment is not only stale now, but gets actively
> > undermined. No idea how likely such a panic() would be, and hence how
> > relevant it is to retain this particular property.
> 
> Ah, right.
> 
> The alternative would be to split it up again. Let's let the x86
> maintainers decide which way is the better one.
> 
> 
> Juergen
> 
> > 
> > Jan
> > 
> > > +	 * After this point everything still needed from the boot loader or
> > > +	 * firmware or kernel text should be early reserved or marked not
> > > +	 * RAM in e820. All other memory is free game.
> > > +	 */
> > > +	early_reserve_memory();
> > > +
> > >   #ifdef CONFIG_X86_32
> > >   	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
> > > @@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
> > >   	parse_early_param();
> > > -	/*
> > > -	 * Do some memory reservations *before* memory is added to
> > > -	 * memblock, so memblock allocations won't overwrite it.
> > > -	 * Do it after early param, so we could get (unlikely) panic from
> > > -	 * serial.
> > > -	 *
> > > -	 * After this point everything still needed from the boot loader or
> > > -	 * firmware or kernel text should be early reserved or marked not
> > > -	 * RAM in e820. All other memory is free game.
> > > -	 */
> > > -	early_reserve_memory();
> > > -
> > >   #ifdef CONFIG_MEMORY_HOTPLUG
> > >   	/*
> > >   	 * Memory used by the kernel cannot be hot-removed because Linux
> > > 
> > 
> 






-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

On Wed, Sep 15, 2021 at 01:00:20PM +0200, Borislav Petkov wrote:
> You forgot to Cc Mike, lemme add him.
> 
> And drop stable@ too.
> 
> On Tue, Sep 14, 2021 at 01:06:22PM +0200, Juergen Gross wrote:
> > On 14.09.21 12:03, Jan Beulich wrote:
> > > On 14.09.2021 11:41, Juergen Gross wrote:
> > > > Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
> > > > reservations") introduced early_reserve_memory() to do all needed
> > > > initial memblock_reserve() calls in one function. Unfortunately the
> > > > call of early_reserve_memory() is done too late for Xen dom0, as in
> > > > some cases a Xen hook called by e820__memory_setup() will need those
> > > > memory reservations to have happened already.
> > > > 
> > > > Move the call of early_reserve_memory() to the beginning of
> > > > setup_arch() in order to avoid such problems.
> > > > 
> > > > Cc: stable@vger.kernel.org
> > > > Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
> > > > Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
> > > > Signed-off-by: Juergen Gross <jgross@suse.com>
> > > > ---
> > > >   arch/x86/kernel/setup.c | 24 ++++++++++++------------
> > > >   1 file changed, 12 insertions(+), 12 deletions(-)
> > > > 
> > > > diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> > > > index 79f164141116..f369c51ec580 100644
> > > > --- a/arch/x86/kernel/setup.c
> > > > +++ b/arch/x86/kernel/setup.c
> > > > @@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
> > > >   void __init setup_arch(char **cmdline_p)
> > > >   {
> > > > +	/*
> > > > +	 * Do some memory reservations *before* memory is added to
> > > > +	 * memblock, so memblock allocations won't overwrite it.
> > > > +	 * Do it after early param, so we could get (unlikely) panic from
> > > > +	 * serial.
> > > 
> > > Hmm, this part of the comment is not only stale now, but gets actively
> > > undermined. No idea how likely such a panic() would be, and hence how
> > > relevant it is to retain this particular property.
> > 
> > Ah, right.
> > 
> > The alternative would be to split it up again. Let's let the x86
> > maintainers decide which way is the better one.

I think the first sentence about reserving memory before memblock
allocations are possible is important and I think we should keep it.

With that

Acked-by: Mike Rapoport <rppt@linux.ibm.com>

> > 
> > 
> > Juergen
> > 
> > > 
> > > Jan
> > > 
> > > > +	 * After this point everything still needed from the boot loader or
> > > > +	 * firmware or kernel text should be early reserved or marked not
> > > > +	 * RAM in e820. All other memory is free game.
> > > > +	 */
> > > > +	early_reserve_memory();
> > > > +
> > > >   #ifdef CONFIG_X86_32
> > > >   	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
> > > > @@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
> > > >   	parse_early_param();
> > > > -	/*
> > > > -	 * Do some memory reservations *before* memory is added to
> > > > -	 * memblock, so memblock allocations won't overwrite it.
> > > > -	 * Do it after early param, so we could get (unlikely) panic from
> > > > -	 * serial.
> > > > -	 *
> > > > -	 * After this point everything still needed from the boot loader or
> > > > -	 * firmware or kernel text should be early reserved or marked not
> > > > -	 * RAM in e820. All other memory is free game.
> > > > -	 */
> > > > -	early_reserve_memory();
> > > > -
> > > >   #ifdef CONFIG_MEMORY_HOTPLUG
> > > >   	/*
> > > >   	 * Memory used by the kernel cannot be hot-removed because Linux
> > > > 
> > > 
> > 
> 
> 
> 
> 
> 
> 
> -- 
> Regards/Gruss,
>     Boris.
> 
> https://people.kernel.org/tglx/notes-about-netiquette

-- 
Sincerely yours,
Mike.

On Tue, Sep 14, 2021 at 11:41:08AM +0200, Juergen Gross wrote:
> Commit a799c2bd29d19c565 ("x86/setup: Consolidate early memory
> reservations") introduced early_reserve_memory() to do all needed
> initial memblock_reserve() calls in one function. Unfortunately the
> call of early_reserve_memory() is done too late for Xen dom0, as in
> some cases a Xen hook called by e820__memory_setup() will need those
> memory reservations to have happened already.
> 
> Move the call of early_reserve_memory() to the beginning of
> setup_arch() in order to avoid such problems.
> 
> Cc: stable@vger.kernel.org
> Fixes: a799c2bd29d19c565 ("x86/setup: Consolidate early memory reservations")
> Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
> Signed-off-by: Juergen Gross <jgross@suse.com>

Thanks, it works!

Tested-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>

> ---
>  arch/x86/kernel/setup.c | 24 ++++++++++++------------
>  1 file changed, 12 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 79f164141116..f369c51ec580 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -757,6 +757,18 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
>  
>  void __init setup_arch(char **cmdline_p)
>  {
> +	/*
> +	 * Do some memory reservations *before* memory is added to
> +	 * memblock, so memblock allocations won't overwrite it.
> +	 * Do it after early param, so we could get (unlikely) panic from
> +	 * serial.
> +	 *
> +	 * After this point everything still needed from the boot loader or
> +	 * firmware or kernel text should be early reserved or marked not
> +	 * RAM in e820. All other memory is free game.
> +	 */
> +	early_reserve_memory();
> +
>  #ifdef CONFIG_X86_32
>  	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
>  
> @@ -876,18 +888,6 @@ void __init setup_arch(char **cmdline_p)
>  
>  	parse_early_param();
>  
> -	/*
> -	 * Do some memory reservations *before* memory is added to
> -	 * memblock, so memblock allocations won't overwrite it.
> -	 * Do it after early param, so we could get (unlikely) panic from
> -	 * serial.
> -	 *
> -	 * After this point everything still needed from the boot loader or
> -	 * firmware or kernel text should be early reserved or marked not
> -	 * RAM in e820. All other memory is free game.
> -	 */
> -	early_reserve_memory();
> -
>  #ifdef CONFIG_MEMORY_HOTPLUG
>  	/*
>  	 * Memory used by the kernel cannot be hot-removed because Linux
> -- 
> 2.26.2
> 

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab