Re: [PATCH v24 3/6] arm64: kdump: Reimplement crashkernel=X

[Baoquan He]

On 05/07/22 at 07:50pm, Catalin Marinas wrote:
> On Sat, May 07, 2022 at 07:10:32AM +0800, Baoquan He wrote:
> > On 05/06/22 at 07:43pm, Zhen Lei wrote:
> > ......  
> > > @@ -118,8 +162,7 @@ static void __init reserve_crashkernel(void)
> > >  	if (crash_base)
> > >  		crash_max = crash_base + crash_size;
> > >  
> > > -	/* Current arm64 boot protocol requires 2MB alignment */
> > > -	crash_base = memblock_phys_alloc_range(crash_size, SZ_2M,
> > > +	crash_base = memblock_phys_alloc_range(crash_size, CRASH_ALIGN,
> > >  					       crash_base, crash_max);
> > >  	if (!crash_base) {
> > >  		pr_warn("cannot allocate crashkernel (size:0x%llx)\n",
> > > @@ -127,6 +170,11 @@ static void __init reserve_crashkernel(void)
> > >  		return;
> > >  	}
> > >  
> > 
> > There's corner case missed, e.g
> > 1) ,high and ,low are specified, CONFIG_ZONE_DMA|DMA32 is not enabled;
> > 2) ,high and ,low are specified, the whole system memory is under 4G.
> 
> My view of ,low is that it should only used to override the default
> ZONE_DMA allocation if that one is not suitable. If no ZONE_DMA exists
> or everything is ZONE_DMA, ignore it altogether. That's a specialist
> case for people that know more about the memory layout, otherwise
> crashkernel=X works in most case with crashkernel=X,high as an
> alternative to allow high allocation.

Totally agree with the conclusion.