[PATCH] x86: Move crashkernel reservation before dma32_reserve_bootmem()

[PATCH] x86: Move crashkernel reservation before dma32_reserve_bootmem()

[Bernhard Walle]

On a x86-64 machine (nothing special I could encounter) I had the problem that
crashkernel reservation with the usual "64M@16M" failed. While debugging that,
I encountered that dma32_reserve_bootmem() reserves a memory region which is in
that area.

Because dma32_reserve_bootmem() does not rely on a specific offset but
crashkernel does, it makes sense to move the crashkernel reservation up a bit.
I tested that patch and it works without problems. I don't see any negative
effects of that move, but maybe I oversaw something ...

While we strictly don't need that patch in 2.6.27 because we have the
automatic, dynamic offset detection, it makes sense to also include it here
because:

  - it's easier to get it in -stable then,
  - many people are still used to the 'crashkernel=...@16M' syntax,
  - not everybody may be using a reloatable kernel.

Signed-off-by: Bernhard Walle <bwalle@suse.de>
---
 arch/x86/kernel/setup.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 531b55b..16101c0 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -792,6 +792,12 @@ void __init setup_arch(char **cmdline_p)
 
 	initmem_init(0, max_pfn);
 
+	/*
+	 * dma32_reserve_bootmem() allocates bootmem which may conflict
+	 * with the crashkernel command line, so do that before
+	 */
+	reserve_crashkernel();
+
 #ifdef CONFIG_X86_64
 	dma32_reserve_bootmem();
 #endif
@@ -808,7 +814,6 @@ void __init setup_arch(char **cmdline_p)
 	 */
 	find_smp_config();
 #endif
-	reserve_crashkernel();
 
 	reserve_ibft_region();
 
-- 
1.5.6

Re: [PATCH] x86: Move crashkernel reservation before dma32_reserve_bootmem()

[Yinghai Lu]

On Thu, Jul 17, 2008 at 3:15 PM, Bernhard Walle <bwalle@suse.de> wrote:
> On a x86-64 machine (nothing special I could encounter) I had the problem that
> crashkernel reservation with the usual "64M@16M" failed. While debugging that,
> I encountered that dma32_reserve_bootmem() reserves a memory region which is in
> that area.
>
> Because dma32_reserve_bootmem() does not rely on a specific offset but
> crashkernel does, it makes sense to move the crashkernel reservation up a bit.
> I tested that patch and it works without problems. I don't see any negative
> effects of that move, but maybe I oversaw something ...
>
> While we strictly don't need that patch in 2.6.27 because we have the
> automatic, dynamic offset detection, it makes sense to also include it here
> because:
>
>  - it's easier to get it in -stable then,
>  - many people are still used to the 'crashkernel=...@16M' syntax,
>  - not everybody may be using a reloatable kernel.
>
> Signed-off-by: Bernhard Walle <bwalle@suse.de>
> ---
>  arch/x86/kernel/setup.c |    7 ++++++-
>  1 files changed, 6 insertions(+), 1 deletions(-)
>
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index 531b55b..16101c0 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -792,6 +792,12 @@ void __init setup_arch(char **cmdline_p)
>
>        initmem_init(0, max_pfn);
>
> +       /*
> +        * dma32_reserve_bootmem() allocates bootmem which may conflict
> +        * with the crashkernel command line, so do that before
> +        */
> +       reserve_crashkernel();
> +
>  #ifdef CONFIG_X86_64
>        dma32_reserve_bootmem();
>  #endif
> @@ -808,7 +814,6 @@ void __init setup_arch(char **cmdline_p)
>         */
>        find_smp_config();
>  #endif
> -       reserve_crashkernel();
>
>        reserve_ibft_region();
>

Joe Jin already had another one to move dma32_reserve_bootmem later

YH

Re: [PATCH] x86: Move crashkernel reservation before dma32_reserve_bootmem()

[Bernhard Walle]

* Yinghai Lu [2008-07-17 15:50]:
>
> > +       /*
> > +        * dma32_reserve_bootmem() allocates bootmem which may conflict
> > +        * with the crashkernel command line, so do that before
> > +        */
> > +       reserve_crashkernel();
> > +
> >  #ifdef CONFIG_X86_64
> >        dma32_reserve_bootmem();
> >  #endif
> > @@ -808,7 +814,6 @@ void __init setup_arch(char **cmdline_p)
> >         */
> >        find_smp_config();
> >  #endif
> > -       reserve_crashkernel();
> >
> >        reserve_ibft_region();
> 
> Joe Jin already had another one to move dma32_reserve_bootmem later

Commit id? Current status?


Bernhard
-- 
Bernhard Walle, SUSE LINUX Products GmbH, Architecture Development

Re: [PATCH] x86: Move crashkernel reservation before dma32_reserve_bootmem()

[Yinghai Lu]

On Fri, Jul 18, 2008 at 2:52 AM, Bernhard Walle <bwalle@suse.de> wrote:
> * Yinghai Lu [2008-07-17 15:50]:
>>
>> > +       /*
>> > +        * dma32_reserve_bootmem() allocates bootmem which may conflict
>> > +        * with the crashkernel command line, so do that before
>> > +        */
>> > +       reserve_crashkernel();
>> > +
>> >  #ifdef CONFIG_X86_64
>> >        dma32_reserve_bootmem();
>> >  #endif
>> > @@ -808,7 +814,6 @@ void __init setup_arch(char **cmdline_p)
>> >         */
>> >        find_smp_config();
>> >  #endif
>> > -       reserve_crashkernel();
>> >
>> >        reserve_ibft_region();
>>
>> Joe Jin already had another one to move dma32_reserve_bootmem later
>
> Commit id? Current status?

it seems commit description from you is more good.
please create one version move dma32_reserve_boot_mem after
reserve_crashkernel and use your commit description.

YH

[PATCH] x86: Move dma32_reserve_bootmem() after reserve_crashkernel()

[Bernhard Walle]

On a x86-64 machine (nothing special I could encounter) I had the problem that
crashkernel reservation with the usual "64M@16M" failed. While debugging that,
I encountered that dma32_reserve_bootmem() reserves a memory region which is in
that area.

Because dma32_reserve_bootmem() does not rely on a specific offset but
crashkernel does, it makes sense to move the dma32_reserve_bootmem()
reservation down a bit. I tested that patch and it works without problems. I
don't see any negative effects of that move, but maybe I oversaw something ...

While we strictly don't need that patch in 2.6.27 because we have the
automatic, dynamic offset detection, it makes sense to also include it here
because:

  - it's easier to get it in -stable then,
  - many people are still used to the 'crashkernel=...@16M' syntax,
  - not everybody may be using a reloatable kernel.

Signed-off-by: Bernhard Walle <bwalle@suse.de>
---
 arch/x86/kernel/setup.c |   13 +++++++++----
 1 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 531b55b..74d110e 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -792,10 +792,6 @@ void __init setup_arch(char **cmdline_p)
 
 	initmem_init(0, max_pfn);
 
-#ifdef CONFIG_X86_64
-	dma32_reserve_bootmem();
-#endif
-
 #ifdef CONFIG_ACPI_SLEEP
 	/*
 	 * Reserve low memory region for sleep support.
@@ -810,6 +806,15 @@ void __init setup_arch(char **cmdline_p)
 #endif
 	reserve_crashkernel();
 
+#ifdef CONFIG_X86_64
+	/*
+	 * dma32_reserve_bootmem() allocates bootmem which may conflict
+	 * with the crashkernel command line, so do that after
+	 * reserve_crashkernel()
+	 */
+	dma32_reserve_bootmem();
+#endif
+
 	reserve_ibft_region();
 
 #ifdef CONFIG_KVM_CLOCK
-- 
1.5.6

Re: [PATCH] x86: Move dma32_reserve_bootmem() after reserve_crashkernel()

[Ingo Molnar]

* Bernhard Walle <bwalle@suse.de> wrote:

> On a x86-64 machine (nothing special I could encounter) I had the 
> problem that crashkernel reservation with the usual "64M@16M" failed. 
> While debugging that, I encountered that dma32_reserve_bootmem() 
> reserves a memory region which is in that area.
> 
> Because dma32_reserve_bootmem() does not rely on a specific offset but 
> crashkernel does, it makes sense to move the dma32_reserve_bootmem() 
> reservation down a bit. I tested that patch and it works without 
> problems. I don't see any negative effects of that move, but maybe I 
> oversaw something ...
> 
> While we strictly don't need that patch in 2.6.27 because we have the 
> automatic, dynamic offset detection, it makes sense to also include it 
> here because:
> 
>   - it's easier to get it in -stable then,
>   - many people are still used to the 'crashkernel=...@16M' syntax,
>   - not everybody may be using a reloatable kernel.

applied to tip/x86/crashdump, thanks Bernhard.

	Ingo