Re: [PATCH v8 0/5] support reserving crashkernel above 4G on arm64 kdump

[Will Deacon]

On Thu, Jun 04, 2020 at 01:17:06AM +0530, Bhupesh Sharma wrote:
> On Wed, Jun 3, 2020 at 9:03 PM John Donnelly <john.p.donnelly@xxxxxxxxxx> wrote:
> > > On Jun 3, 2020, at 8:20 AM, chenzhou <chenzhou10@xxxxxxxxxx> wrote:
> > > On 2020/6/3 19:47, Prabhakar Kushwaha wrote:
> > >>>> diff --git a/kernel/crash_core.c b/kernel/crash_core.c
> > >>>> index 7f9e5a6dc48c..bd67b90d35bd 100644
> > >>>> --- a/kernel/crash_core.c
> > >>>> +++ b/kernel/crash_core.c
> > >>>> @@ -354,7 +354,7 @@ int __init reserve_crashkernel_low(void)
> > >>>>                       return 0;
> > >>>>       }
> > >>>>
> > >>>> -       low_base = memblock_find_in_range(0, 1ULL << 32, low_size, CRASH_ALIGN);
> > >>>> +       low_base = memblock_find_in_range(0,0xc0000000, low_size, CRASH_ALIGN);
> > >>>>       if (!low_base) {
> > >>>>               pr_err("Cannot reserve %ldMB crashkernel low memory,
> > >>>> please try smaller size.\n",
> > >>>>                      (unsigned long)(low_size >> 20));
> > >>>>
> > >>>>
> > >>>    I suspect  0xc0000000  would need to be a CONFIG item  and not hard-coded.
> > >>>
> > >> if you consider this as valid change,  can you please incorporate as
> > >> part of your patch-set.
> > >
> > > After commit 1a8e1cef7 ("arm64: use both ZONE_DMA and
> > > ZONE_DMA32")ïthe 0-4G memory is splited to DMA [mem
> > > 0x0000000000000000-0x000000003fffffff] and DMA32 [mem
> > > 0x0000000040000000-0x00000000ffffffff] on arm64.
> > >
> > > From the above discussion, on your platform, the low crashkernel fall
> > > in DMA32 region, but your environment needs to access DMA region, so
> > > there is the call trace.
> > >
> > > I have a question, why do you choose 0xc0000000 here?
> > >
> > > Besides, this is common code, we also need to consider about x86.
> > >
> >
> >  + nsaenzjulienne@xxxxxxx
> >
> >   Exactly .  This is why it needs to be a CONFIG option for  Raspberry
> >   ..,  or device tree option.
> >
> >
> >   We could revert 1a8e1cef7 since it broke  Arm kdump too.
> 
> Well, unfortunately the patch for commit 1a8e1cef7603 ("arm64: use
> both ZONE_DMA and ZONE_DMA32") was not Cc'ed to the kexec mailing
> list, thus we couldn't get many eyes on it for a thorough review from
> kexec/kdump p-o-v.
> 
> Also we historically never had distinction in common arch code on the
> basis of the intended end use-case: embedded, server or automotive, so
> I am not sure introducing a Raspberry specific CONFIG option would be
> a good idea.

Right, we need a fix that works for everybody, since we try hard for a
single Image that works for all platforms.

What I don't really understand is why, with Chen's patches applied, we can't
just keep the crashkernel out of the DMA zones altogether when no base is
specified. I guess I'll just look out for your patch!

Will