linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Yasunori Goto <y-goto@jp.fujitsu.com>
To: Andi Kleen <ak@suse.de>
Cc: Andrew Morton <akpm@osdl.org>,
	peterc@gelato.unsw.edu.au, linux-mm@kvack.org, mbligh@mbligh.org,
	linux-ia64@vger.kernel.org, kravetz@us.ibm.com,
	tony.luck@intel.com
Subject: Re: [PATCH] gurantee DMA area for alloc_bootmem_low() ver. 2.
Date: Fri, 19 Aug 2005 11:29:32 +0900	[thread overview]
Message-ID: <20050819102706.62C7.Y-GOTO@jp.fujitsu.com> (raw)
In-Reply-To: <p73y86ysz5c.fsf@verdi.suse.de>

Hi, Andi-san.

> Andrew Morton <akpm@osdl.org> writes:
> > > 
> > >  To avoid this panic, following patch confirm allocated area, and retry
> > >  if it is not in DMA.
> > >  I tested this patch on my Tiger 4 and our new server.
> > 
> > It kills my x86_64 box:
> 
> 
> Funny I ran into a similar problem recently. On a multi node x86-64
> system when swiotlb is forced (normally those are AMD systems which
> use the AMD hardware IOMMU) the bootmem_alloc in swiotlb.c would
> allocate from the last node. Why? Because alloc_bootmem just
> does for_each_pgdat() and tries each node and the pgdat list
> starts with the highest node going down to the lowest.
> 
> I just changed the ordering of the pgdat list that made bootmem 
> work again.

It was another candidate for modification indeed.
But, if new node is hot-added (I'm working for it),
pgdat list should be sorted by order of memory.
I suppose it is a bit messy.
At least, I think the order of pgdat link list must not depend on 
memory address.

To hot-add a node, it is better that pgdat link list is removed.
(Hot add code will set JUST node_online_map by it.)
I posted a patch to remove this link list 3 month ago.
http://marc.theaimsgroup.com/?l=linux-mm&m=111596924629564&w=2
http://marc.theaimsgroup.com/?l=linux-mm&m=111596953711780&w=2

Thanks for your comment. 
Bye.

> 
> Index: linux/mm/bootmem.c
> ===================================================================
> --- linux.orig/mm/bootmem.c
> +++ linux/mm/bootmem.c
> @@ -61,9 +61,17 @@ static unsigned long __init init_bootmem
>  {
>  	bootmem_data_t *bdata = pgdat->bdata;
>  	unsigned long mapsize = ((end - start)+7)/8;
> +	static struct pglist_data *pgdat_last;
>  
> -	pgdat->pgdat_next = pgdat_list;
> -	pgdat_list = pgdat;
> +	pgdat->pgdat_next = NULL;
> +	/* Add new nodes last so that bootmem always starts 
> +	   searching in the first nodes, not the last ones */
> +	if (pgdat_last)
> +		pgdat_last->pgdat_next = pgdat;
> +	else {
> +		pgdat_list = pgdat; 	
> +		pgdat_last = pgdat;
> +	}
>  
>  	mapsize = ALIGN(mapsize, sizeof(long));
>  	bdata->node_bootmem_map = phys_to_virt(mapstart << PAGE_SHIFT);

-- 
Yasunori Goto 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2005-08-19  2:29 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-08-09 11:11 Yasunori Goto
2005-08-09 15:05 ` Martin J. Bligh
2005-08-09 21:15 ` Mike Kravetz
2005-08-10  3:06   ` Dave Hansen
2005-08-10 16:23     ` Dave Hansen
2005-08-11 20:46   ` Christoph Lameter
2005-08-11 21:14     ` Mike Kravetz
2005-08-11 22:37       ` Christoph Lameter
2005-08-09 23:02 ` Peter Chubb
2005-08-10  6:10   ` Yasunori Goto
2005-08-18 19:52     ` Andrew Morton
2005-08-18 21:39       ` Andi Kleen
2005-08-19  2:29         ` Yasunori Goto [this message]
2005-08-19  3:03           ` Andi Kleen
2005-08-19  1:26       ` Yasunori Goto
2005-08-25  9:15       ` Yasunori Goto

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=20050819102706.62C7.Y-GOTO@jp.fujitsu.com \
    --to=y-goto@jp.fujitsu.com \
    --cc=ak@suse.de \
    --cc=akpm@osdl.org \
    --cc=kravetz@us.ibm.com \
    --cc=linux-ia64@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mbligh@mbligh.org \
    --cc=peterc@gelato.unsw.edu.au \
    --cc=tony.luck@intel.com \
    /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