From: Christoph Hellwig <hch@lst.de>
To: David Miller <davem@davemloft.net>
Cc: hch@lst.de, linux-kernel@vger.kernel.org, netdev@oss.sgi.com,
linux-mm@kvack.org
Subject: Re: [PATCH 2/3] add dev_to_node()
Date: Sat, 4 Nov 2006 23:56:29 +0100 [thread overview]
Message-ID: <20061104225629.GA31437@lst.de> (raw)
In-Reply-To: <20061030.143357.130208425.davem@davemloft.net>
On Mon, Oct 30, 2006 at 02:33:57PM -0800, David Miller wrote:
> It may be a bit much to be calling all the way through up to the PCI
> layer just to pluck out a simple integer, don't you think? The PCI
> bus pointer comparison is just a symptom of how silly this is.
>
> Especially since this will be used for every packet allocation a
> device makes.
>
> So, please add some sanity to this situation and just put the node
> into the generic struct device. :-)
I was concerned about growing struct device, on smaller system it
already eats up a lot of memory. But we can make the node member
conditional on CONFIG_NUMA, as I did in the patch below.
This directly replaces PATCH 2/2 (the one we're replying to), all
others remain unmodified.
Index: linux-2.6/include/linux/device.h
===================================================================
--- linux-2.6.orig/include/linux/device.h 2006-10-29 16:02:38.000000000 +0100
+++ linux-2.6/include/linux/device.h 2006-11-02 12:47:17.000000000 +0100
@@ -347,6 +347,9 @@
BIOS data),reserved for device core*/
struct dev_pm_info power;
+#ifdef CONFIG_NUMA
+ int numa_node; /* NUMA node this device is close to */
+#endif
u64 *dma_mask; /* dma mask (if dma'able device) */
u64 coherent_dma_mask;/* Like dma_mask, but for
alloc_coherent mappings as
@@ -368,6 +371,12 @@
void (*release)(struct device * dev);
};
+#ifdef CONFIG_NUMA
+#define dev_to_node(dev) ((dev)->numa_node)
+#else
+#define dev_to_node(dev) (-1)
+#endif
+
static inline void *
dev_get_drvdata (struct device *dev)
{
Index: linux-2.6/drivers/base/core.c
===================================================================
--- linux-2.6.orig/drivers/base/core.c 2006-10-23 17:21:44.000000000 +0200
+++ linux-2.6/drivers/base/core.c 2006-11-02 12:48:12.000000000 +0100
@@ -381,6 +381,7 @@
INIT_LIST_HEAD(&dev->node);
init_MUTEX(&dev->sem);
device_init_wakeup(dev, 0);
+ dev->numa_node = -1;
}
/**
Index: linux-2.6/drivers/pci/probe.c
===================================================================
--- linux-2.6.orig/drivers/pci/probe.c 2006-10-23 17:21:46.000000000 +0200
+++ linux-2.6/drivers/pci/probe.c 2006-11-02 12:47:35.000000000 +0100
@@ -846,6 +846,7 @@
dev->dev.release = pci_release_dev;
pci_dev_get(dev);
+ dev->dev.numa_node = pcibus_to_node(bus);
dev->dev.dma_mask = &dev->dma_mask;
dev->dev.coherent_dma_mask = 0xffffffffull;
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Hellwig <hch@lst.de>
To: David Miller <davem@davemloft.net>
Cc: hch@lst.de, linux-kernel@vger.kernel.org, netdev@oss.sgi.com,
linux-mm@kvack.org
Subject: Re: [PATCH 2/3] add dev_to_node()
Date: Sat, 4 Nov 2006 23:56:29 +0100 [thread overview]
Message-ID: <20061104225629.GA31437@lst.de> (raw)
In-Reply-To: <20061030.143357.130208425.davem@davemloft.net>
On Mon, Oct 30, 2006 at 02:33:57PM -0800, David Miller wrote:
> It may be a bit much to be calling all the way through up to the PCI
> layer just to pluck out a simple integer, don't you think? The PCI
> bus pointer comparison is just a symptom of how silly this is.
>
> Especially since this will be used for every packet allocation a
> device makes.
>
> So, please add some sanity to this situation and just put the node
> into the generic struct device. :-)
I was concerned about growing struct device, on smaller system it
already eats up a lot of memory. But we can make the node member
conditional on CONFIG_NUMA, as I did in the patch below.
This directly replaces PATCH 2/2 (the one we're replying to), all
others remain unmodified.
Index: linux-2.6/include/linux/device.h
===================================================================
--- linux-2.6.orig/include/linux/device.h 2006-10-29 16:02:38.000000000 +0100
+++ linux-2.6/include/linux/device.h 2006-11-02 12:47:17.000000000 +0100
@@ -347,6 +347,9 @@
BIOS data),reserved for device core*/
struct dev_pm_info power;
+#ifdef CONFIG_NUMA
+ int numa_node; /* NUMA node this device is close to */
+#endif
u64 *dma_mask; /* dma mask (if dma'able device) */
u64 coherent_dma_mask;/* Like dma_mask, but for
alloc_coherent mappings as
@@ -368,6 +371,12 @@
void (*release)(struct device * dev);
};
+#ifdef CONFIG_NUMA
+#define dev_to_node(dev) ((dev)->numa_node)
+#else
+#define dev_to_node(dev) (-1)
+#endif
+
static inline void *
dev_get_drvdata (struct device *dev)
{
Index: linux-2.6/drivers/base/core.c
===================================================================
--- linux-2.6.orig/drivers/base/core.c 2006-10-23 17:21:44.000000000 +0200
+++ linux-2.6/drivers/base/core.c 2006-11-02 12:48:12.000000000 +0100
@@ -381,6 +381,7 @@
INIT_LIST_HEAD(&dev->node);
init_MUTEX(&dev->sem);
device_init_wakeup(dev, 0);
+ dev->numa_node = -1;
}
/**
Index: linux-2.6/drivers/pci/probe.c
===================================================================
--- linux-2.6.orig/drivers/pci/probe.c 2006-10-23 17:21:46.000000000 +0200
+++ linux-2.6/drivers/pci/probe.c 2006-11-02 12:47:35.000000000 +0100
@@ -846,6 +846,7 @@
dev->dev.release = pci_release_dev;
pci_dev_get(dev);
+ dev->dev.numa_node = pcibus_to_node(bus);
dev->dev.dma_mask = &dev->dma_mask;
dev->dev.coherent_dma_mask = 0xffffffffull;
--
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>
next prev parent reply other threads:[~2006-11-04 22:56 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-10-30 14:15 [PATCH 2/3] add dev_to_node() Christoph Hellwig
2006-10-30 14:15 ` Christoph Hellwig
2006-10-30 22:33 ` David Miller
2006-10-30 22:33 ` David Miller, Christoph Hellwig
2006-11-01 0:10 ` Christoph Lameter
2006-11-01 0:10 ` Christoph Lameter
2006-11-01 0:53 ` David Miller
2006-11-01 0:53 ` David Miller, Christoph Lameter
2006-11-01 1:58 ` Christoph Lameter
2006-11-01 1:58 ` Christoph Lameter
2006-11-04 22:56 ` Christoph Hellwig [this message]
2006-11-04 22:56 ` Christoph Hellwig
2006-11-04 23:06 ` Dave Jones
2006-11-04 23:06 ` Dave Jones
2006-11-04 23:09 ` Christoph Hellwig
2006-11-04 23:09 ` Christoph Hellwig
2006-11-04 23:53 ` Christoph Hellwig
2006-11-04 23:53 ` Christoph Hellwig
2006-11-05 8:22 ` David Miller
2006-11-05 8:22 ` David Miller, Christoph Hellwig
2006-11-06 23:39 ` Christoph Hellwig
2006-11-06 23:39 ` Christoph Hellwig
2006-11-07 6:25 ` Ravikiran G Thirumalai
2006-11-07 6:25 ` Ravikiran G Thirumalai
2006-11-07 10:15 ` Christoph Hellwig
2006-11-07 10:15 ` Christoph Hellwig
2006-11-08 2:40 ` KAMEZAWA Hiroyuki
2006-11-08 2:40 ` KAMEZAWA Hiroyuki
2006-11-10 18:16 ` Christoph Lameter
2006-11-10 18:16 ` Christoph Lameter
2006-11-10 18:28 ` Lee Schermerhorn
2006-11-10 18:28 ` Lee Schermerhorn
2006-11-11 0:08 ` KAMEZAWA Hiroyuki
2006-11-11 0:08 ` KAMEZAWA Hiroyuki
-- strict thread matches above, loose matches on Subject: below --
2006-10-30 14:15 Christoph Hellwig
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=20061104225629.GA31437@lst.de \
--to=hch@lst.de \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=netdev@oss.sgi.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.