Received-SPF: pass (google.com: domain of linux-kernel+bounces-34824-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Fri, 19 Jan 2024 10:26:41 +0800
From: Huang Pei <huangpei@loongson.cn>
To: Mike Rapoport <rppt@kernel.org>
Cc: Yajun Deng <yajun.deng@linux.dev>, akpm@linux-foundation.org,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] memblock: fix crash when reserved memory is not added to
 memory
Message-ID: <20240119022641.xyw4n4dj2iw2deas@Board-3A3000>
References: <20240118061853.2652295-1-yajun.deng@linux.dev>
 <ZajIQGtqUNWgBWkk@kernel.org>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <ZajIQGtqUNWgBWkk@kernel.org>

On Thu, Jan 18, 2024 at 08:42:08AM +0200, Mike Rapoport wrote:
> (adding Huang Pei to verify this fixes the crash on loongarch)
> 
It fixed 61167ad5fecd on MIPS/loongson64, tested on 3b1500 with only
this one;

My loongarch64 machine is not available now, I can not test it, sorry.
It is already applied a arch-specific workaround in v6.5+

> On Thu, Jan 18, 2024 at 02:18:53PM +0800, Yajun Deng wrote:
> > After commit 61167ad5fecd ("mm: pass nid to reserve_bootmem_region()"),
> > we set nid on all reserved pages based on memory region, but some reserved
> > memory may not be added to memory on certain architectures. The nid is
> > invalid.
> > 
> > Add the nid check, make sure the nid is valid by early_pfn_to_nid().
> > 
> > Fixes: 61167ad5fecd ("mm: pass nid to reserve_bootmem_region()")
> > Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
> > ---
> >  mm/memblock.c | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/mm/memblock.c b/mm/memblock.c
> > index abd92869874d..4dcb2ee35eca 100644
> > --- a/mm/memblock.c
> > +++ b/mm/memblock.c
> > @@ -2176,6 +2176,9 @@ static void __init memmap_init_reserved_pages(void)
> >  			start = region->base;
> >  			end = start + region->size;
> >  
> > +			if (nid == NUMA_NO_NODE || nid >= MAX_NUMNODES)
> > +				nid = early_pfn_to_nid(PFN_DOWN(start));
> > +
> >  			reserve_bootmem_region(start, end, nid);
> >  		}
> >  	}
> > -- 
> > 2.25.1
> > 
> 
> -- 
> Sincerely yours,
> Mike.