From: Zou Nan hai <nanhai.zou@intel.com>
To: Andi Kleen <ak@suse.de>
Cc: Tony Luck <tony.luck@intel.com>,
LKML <linux-kernel@vger.kernel.org>,
Andrew Morton <akpm@osdl.org>,
Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Subject: Re: [Patch] Move swiotlb_init early on X86_64
Date: 02 Mar 2006 12:33:23 +0800 [thread overview]
Message-ID: <1141274003.2642.1400.camel@linux-znh> (raw)
In-Reply-To: <200603020530.09552.ak@suse.de>
On Thu, 2006-03-02 at 12:30, Andi Kleen wrote:
> On Thursday 02 March 2006 05:15, Tony Luck wrote:
> > On 01 Mar 2006 09:10:58 +0800, Zou Nan hai <nanhai.zou@intel.com> wrote:
> > > on X86_64, swiotlb buffer is allocated in mem_init, after memmap and vfs cache allocation.
> > >
> > > On platforms with huge physical memory,
> > > large memmap and vfs cache may eat up all usable system memory
> > > under 4G.
> > >
> > > Move swiotlb_init early before memmap is allocated can
> > > solve this issue.
> >
> > Shouldn't memmap be allocated from memory above 4G (if available)? Using
> > up lots of <4G memory on something that doesn't need to be below 4G
> > sounds like a poor use of resources.
>
> On the really large machines it will be distributed over the nodes anyways.
> But yes the single node SMP case should probably allocate it higher.
>
> -Andi
Really, then how about the following patch?
Let normal bootmem allocator go above 4G first.
This will save more memory with address less than 4G.
Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
--- linux-2.6.16-rc5/mm/bootmem.c 2006-03-03 08:31:52.000000000 +0800
+++ b/mm/bootmem.c 2006-03-03 09:05:17.000000000 +0800
@@ -381,16 +381,24 @@ unsigned long __init free_all_bootmem (v
return(free_all_bootmem_core(NODE_DATA(0)));
}
+#define LOW32LIMIT 0xffffffff
+
void * __init __alloc_bootmem(unsigned long size, unsigned long align, unsigned long goal)
{
pg_data_t *pgdat = pgdat_list;
void *ptr;
+ if (goal < LOW32LIMIT) {
+ for_each_pgdat(pgdat)
+ if ((ptr = __alloc_bootmem_core(pgdat->bdata, size,
+ align, LOW32LIMIT, 0)))
+ return(ptr);
+ }
+
for_each_pgdat(pgdat)
if ((ptr = __alloc_bootmem_core(pgdat->bdata, size,
align, goal, 0)))
return(ptr);
-
/*
* Whoops, we cannot satisfy the allocation request.
*/
@@ -405,6 +413,13 @@ void * __init __alloc_bootmem_node(pg_da
{
void *ptr;
+ if (goal < LOW32LIMIT) {
+ ptr = __alloc_bootmem_core(pgdat->bdata, size, align,
+ LOW32LIMIT, 0);
+ if (ptr)
+ return (ptr);
+ }
+
ptr = __alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
if (ptr)
return (ptr);
@@ -412,7 +427,6 @@ void * __init __alloc_bootmem_node(pg_da
return __alloc_bootmem(size, align, goal);
}
-#define LOW32LIMIT 0xffffffff
void * __init __alloc_bootmem_low(unsigned long size, unsigned long align, unsigned long goal)
{
next prev parent reply other threads:[~2006-03-02 6:13 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-03-01 1:10 [Patch] Move swiotlb_init early on X86_64 Zou Nan hai
2006-03-02 4:15 ` Tony Luck
2006-03-02 4:30 ` Andi Kleen
2006-03-02 4:33 ` Zou Nan hai [this message]
2006-03-07 8:39 ` Andi Kleen
2006-03-07 23:23 ` Zou Nan hai
2006-03-08 9:33 ` Andi Kleen
-- strict thread matches above, loose matches on Subject: below --
2006-03-02 9:09 Zhang, Yanmin
2006-03-02 23:35 ` Zou Nan hai
2006-03-03 1:32 ` Andi Kleen
2006-03-03 1:59 Pallipadi, Venkatesh
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=1141274003.2642.1400.camel@linux-znh \
--to=nanhai.zou@intel.com \
--cc=ak@suse.de \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tony.luck@intel.com \
--cc=venkatesh.pallipadi@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