public inbox for linux-pm@vger.kernel.org
 help / color / mirror / Atom feed
From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: linux-pm@lists.osdl.org
Cc: Pavel Machek <pavel@ucw.cz>
Subject: Re: swsusp: which page should be saved?
Date: Thu, 16 Mar 2006 17:05:27 +0100	[thread overview]
Message-ID: <200603161705.28080.rjw@sisk.pl> (raw)
In-Reply-To: <1142481207.26706.16.camel@sli10-desk.sh.intel.com>

[-- Attachment #1: Type: text/plain, Size: 4147 bytes --]

On Thursday 16 March 2006 04:53, Shaohua Li wrote:
> I thought there was a discussion before, but I still saw many pages like
> BIOS reserved are saved/restored in swsusp.

I've never seen this causing any problems but generally it's a good idea to
avoid saving these pages IMO.

> Pages reserved by BIOS, not 
> used by OS and kernel text should be skipped in swsusp to me. Below
> patch works in my test. Any thought?

You should do this for x86_64 too.

Greetings,
Rafael


> ---
> 
>  linux-2.6.15-root/arch/i386/kernel/setup.c |   78 +++++++++++++++++++++++++++++
>  linux-2.6.15-root/kernel/power/snapshot.c  |    1 
>  2 files changed, 78 insertions(+), 1 deletion(-)
> 
> diff -puN arch/i386/kernel/setup.c~nosave_pages arch/i386/kernel/setup.c
> --- linux-2.6.15/arch/i386/kernel/setup.c~nosave_pages	2006-03-14 14:42:24.000000000 +0800
> +++ linux-2.6.15-root/arch/i386/kernel/setup.c	2006-03-15 09:38:18.000000000 +0800
> @@ -1483,6 +1483,83 @@ static void set_mca_bus(int x)
>  static void set_mca_bus(int x) { }
>  #endif
>  
> +extern char __end_rodata;
> +static void __init mark_nosave_page_range(unsigned long start, unsigned long end)
> +{
> +	struct page *page;
> +	while (start <= end) {
> +		printk("%ld,", start);
> +		page = pfn_to_page(start);
> +		SetPageNosave(page);
> +		start ++;
> +	}
> +}
> +
> +static int __init efi_mark_nosave_page(unsigned long start, unsigned long end,
> +	void *arg)
> +{
> +	unsigned long start_pfn, end_pfn, last = *(unsigned long*)arg;
> +
> +	start_pfn = PFN_DOWN(start);
> +	end_pfn = PFN_DOWN(end);
> +	/* check max_low_pfn */
> +	if (start_pfn >= max_low_pfn)
> +		return 0;
> +	if (end_pfn >= max_low_pfn)
> +		end = max_low_pfn;
> +	if (start_pfn > last)
> +		mark_nosave_page_range(last, start_pfn - 1);
> +	*(unsigned long*)arg = end;
> +	return 0;
> +}
> +
> +static void __init mark_nosave_pages(void)
> +{
> +	unsigned long pfn, pfn_end_rodata, last_pfn = 0;
> +	int i;
> +
> +	/* Mark all BIOS reserved regions as nosave */
> +	if (efi_enabled) {
> +		efi_memmap_walk(efi_mark_nosave_page, &last_pfn);
> +		return;
> +	} else {
> +		for (i = 0; i < e820.nr_map; i++) {
> +			unsigned long start_pfn, end_pfn;
> +			if (e820.map[i].type != E820_RAM)
> +				continue;
> +
> +			start_pfn = PFN_DOWN(e820.map[i].addr);
> +			if (start_pfn >= max_low_pfn)
> +				break;
> +			end_pfn = PFN_UP(e820.map[i].addr + e820.map[i].size);
> +
> +			if (end_pfn > max_low_pfn)
> +				end_pfn = max_low_pfn;
> +			if (end_pfn <= start_pfn)
> +				continue;
> +			if (start_pfn > last_pfn) {
> +				start_pfn--;
> +				mark_nosave_page_range(last_pfn, start_pfn);
> +			}
> +			last_pfn = end_pfn;
> +		}
> +	}
> +
> +	if (last_pfn < max_low_pfn)
> +		mark_nosave_page_range(last_pfn, max_low_pfn);
> +
> +	/* kernel text */
> +	pfn = PFN_UP(__pa(__KERNEL_START));
> +	pfn_end_rodata = PFN_DOWN(__pa(&__end_rodata));
> +	mark_nosave_page_range(pfn, pfn_end_rodata);
> +
> +	/* other pages kernel doesn't use (see setup_bootmem_allocator) */
> +	mark_nosave_page_range(0, 0);
> +#ifdef CONFIG_SMP
> +	mark_nosave_page_range(PFN_DOWN(PAGE_SIZE), PFN_DOWN(PAGE_SIZE));
> +#endif
> +}
> +
>  /*
>   * Determine if we were loaded by an EFI loader.  If so, then we have also been
>   * passed the efi memmap, systab, etc., so we should use these data structures
> @@ -1573,6 +1650,7 @@ void __init setup_arch(char **cmdline_p)
>  	remapped_pgdat_init();
>  	sparse_init();
>  	zone_sizes_init();
> +	mark_nosave_pages();
>  
>  	/*
>  	 * NOTE: at this point the bootmem allocator is fully available.
> diff -puN kernel/power/snapshot.c~nosave_pages kernel/power/snapshot.c
> --- linux-2.6.15/kernel/power/snapshot.c~nosave_pages	2006-03-14 14:42:29.000000000 +0800
> +++ linux-2.6.15-root/kernel/power/snapshot.c	2006-03-14 14:43:06.000000000 +0800
> @@ -173,7 +173,6 @@ static int saveable(struct zone *zone, u
>  		return 0;
>  
>  	page = pfn_to_page(pfn);
> -	BUG_ON(PageReserved(page) && PageNosave(page));
>  	if (PageNosave(page))
>  		return 0;
>  	if (PageReserved(page) && pfn_is_nosave(pfn))
> _
> 
> 
> 

-- 
Beer is proof that God loves us and wants us to be happy - Benjamin Franklin

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]



      parent reply	other threads:[~2006-03-16 16:05 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-03-16  3:53 swsusp: which page should be saved? Shaohua Li
2006-03-15 23:08 ` Pavel Machek
2006-03-17  1:12   ` Shaohua Li
2006-03-17  6:59     ` Pavel Machek
2006-03-21  2:19       ` Shaohua Li
2006-03-21  3:33         ` Nigel Cunningham
2006-03-21  9:42         ` Pavel Machek
2006-04-07  3:46           ` Shaohua Li
2006-04-11  7:39             ` Pavel Machek
2006-04-11  7:52               ` Shaohua Li
2006-03-17 10:50     ` Rafael J. Wysocki
2006-03-16 16:05 ` Rafael J. Wysocki [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200603161705.28080.rjw@sisk.pl \
    --to=rjw@sisk.pl \
    --cc=linux-pm@lists.osdl.org \
    --cc=pavel@ucw.cz \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox