[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4DEFA6E3.6090809@xs4all.nl>
Date: Wed, 08 Jun 2011 18:44:19 +0200
From: Jim Bos <jim876@...all.nl>
To: Yinghai Lu <yinghai@...nel.org>
CC: Maarten Lankhorst <m.b.lankhorst@...il.com>,
Matthew Garrett <mjg59@...f.ucam.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Greg KH <gregkh@...e.de>,
"H. Peter Anvin" <hpa@...ux.intel.com>
Subject: Re: 2.6.39.1 immediately reboots/resets on EFI system
On 06/08/2011 12:25 AM, Yinghai Lu wrote:
> On 06/07/2011 05:22 AM, Maarten Lankhorst wrote:
>> It seems the error still occurs when calling free_bootmem_late
>> even if I only reserve blocks that haven't been reserved yet.
>
> please check this one.
>
> ---
> arch/x86/platform/efi/efi.c | 2 +-
> include/linux/mm.h | 2 ++
> mm/page_alloc.c | 32 ++++++++++++++++++++++++++++++++
> 3 files changed, 35 insertions(+), 1 deletion(-)
>
> Index: linux-2.6/arch/x86/platform/efi/efi.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/platform/efi/efi.c
> +++ linux-2.6/arch/x86/platform/efi/efi.c
> @@ -368,7 +368,7 @@ static void __init efi_free_boot_service
> md->type != EFI_BOOT_SERVICES_DATA)
> continue;
>
> - free_bootmem_late(start, size);
> + free_bootmem_late_with_active_regions(start, size);
> }
> }
>
> Index: linux-2.6/include/linux/mm.h
> ===================================================================
> --- linux-2.6.orig/include/linux/mm.h
> +++ linux-2.6/include/linux/mm.h
> @@ -1322,6 +1322,8 @@ extern void get_pfn_range_for_nid(unsign
> extern unsigned long find_min_pfn_with_active_regions(void);
> extern void free_bootmem_with_active_regions(int nid,
> unsigned long max_low_pfn);
> +void free_bootmem_late_with_active_regions(unsigned long addr,
> + unsigned long size);
> int add_from_early_node_map(struct range *range, int az,
> int nr_range, int nid);
> u64 __init find_memory_core_early(int nid, u64 size, u64 align,
> Index: linux-2.6/mm/page_alloc.c
> ===================================================================
> --- linux-2.6.orig/mm/page_alloc.c
> +++ linux-2.6/mm/page_alloc.c
> @@ -3828,6 +3828,38 @@ void __init free_bootmem_with_active_reg
> }
> }
>
> +/**
> + * free_bootmem_late_with_active_regions - Call free_bootmem_late for each active range
> + * @addr: starting address of the range
> + * @size: size of the range in bytes
> + *
> + * this function make sure on active regions only
> + */
> +void __init free_bootmem_late_with_active_regions(unsigned long addr,
> + unsigned long size)
> +{
> + int i;
> + int nid = MAX_NUMNODES;
> + unsigned long start_pfn = PFN_UP(addr);
> + unsigned long end_pfn = PFN_DOWN(addr + size);
> +
> + if (start_pfn >= end_pfn)
> + return;
> +
> + for_each_active_range_index_in_nid(i, nid) {
> + unsigned long common_start, common_end;
> +
> + common_start = max(start_pfn, early_node_map[i].start_pfn);
> + common_end = min(end_pfn, early_node_map[i].end_pfn);
> +
> + if (common_start >= common_end)
> + continue;
> +
> + free_bootmem_late(common_start << PAGE_SHIFT,
> + (common_end - common_start) << PAGE_SHIFT);
> + }
> +}
> +
> #ifdef CONFIG_HAVE_MEMBLOCK
> /*
> * Basic iterator support. Return the last range of PFNs for a node
>
I applied this on top of 3.0-rc1 (with offsets), kernels boots but see
now in dmesg output a new 'BUG: Bad page state in process swapper
pfn:01800' see attached dmesg output.
Jim
View attachment "bug.txt" of type "text/plain" (1075 bytes)
Download attachment "dmesg-3.0.0-rc1.gz" of type "application/gzip" (14862 bytes)
Powered by blists - more mailing lists