In architecture like powerpc, we can have cpus without any local memory
attached to it. In such cases the node does not have real memory.
Use local_memory_node(), which is guaranteed to have memory.
local_memory_node is a noop in other architectures that does not support
memoryless nodes.
Signed-off-by: Xianting Tian <[email protected]>
---
net/core/dev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index 266073e30..dcb4533ef 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2590,7 +2590,7 @@ static struct xps_map *expand_xps_map(struct xps_map *map, int attr_index,
new_map = kzalloc(XPS_MAP_SIZE(alloc_len), GFP_KERNEL);
else
new_map = kzalloc_node(XPS_MAP_SIZE(alloc_len), GFP_KERNEL,
- cpu_to_node(attr_index));
+ local_memory_node(cpu_to_node(attr_index)));
if (!new_map)
return NULL;
--
2.17.1
On Sun, 11 Oct 2020 12:11:40 +0800 Xianting Tian wrote:
> In architecture like powerpc, we can have cpus without any local memory
> attached to it. In such cases the node does not have real memory.
>
> Use local_memory_node(), which is guaranteed to have memory.
> local_memory_node is a noop in other architectures that does not support
> memoryless nodes.
>
> Signed-off-by: Xianting Tian <[email protected]>
> ---
> net/core/dev.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 266073e30..dcb4533ef 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -2590,7 +2590,7 @@ static struct xps_map *expand_xps_map(struct xps_map *map, int attr_index,
> new_map = kzalloc(XPS_MAP_SIZE(alloc_len), GFP_KERNEL);
> else
> new_map = kzalloc_node(XPS_MAP_SIZE(alloc_len), GFP_KERNEL,
> - cpu_to_node(attr_index));
> + local_memory_node(cpu_to_node(attr_index)));
> if (!new_map)
> return NULL;
>
Are we going to patch all kmalloc_node() callers now to apply
local_memory_node()? Can't the allocator take care of this?
Hi Jakub,
Thanks for your suggestion,
Let me try it :-)
-----Original Message-----
From: Jakub Kicinski [mailto:[email protected]]
Sent: Monday, October 12, 2020 3:18 AM
To: tianxianting (RD) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH] net: Avoid allocing memory on memoryless numa node
On Sun, 11 Oct 2020 12:11:40 +0800 Xianting Tian wrote:
> In architecture like powerpc, we can have cpus without any local
> memory attached to it. In such cases the node does not have real memory.
>
> Use local_memory_node(), which is guaranteed to have memory.
> local_memory_node is a noop in other architectures that does not
> support memoryless nodes.
>
> Signed-off-by: Xianting Tian <[email protected]>
> ---
> net/core/dev.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/net/core/dev.c b/net/core/dev.c index
> 266073e30..dcb4533ef 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -2590,7 +2590,7 @@ static struct xps_map *expand_xps_map(struct xps_map *map, int attr_index,
> new_map = kzalloc(XPS_MAP_SIZE(alloc_len), GFP_KERNEL);
> else
> new_map = kzalloc_node(XPS_MAP_SIZE(alloc_len), GFP_KERNEL,
> - cpu_to_node(attr_index));
> + local_memory_node(cpu_to_node(attr_index)));
> if (!new_map)
> return NULL;
>
Are we going to patch all kmalloc_node() callers now to apply local_memory_node()? Can't the allocator take care of this?