mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Haifeng Xu <haifeng.xu@shopee.com>
To: rafael@kernel.org, lenb@kernel.org, dan.j.williams@intel.com,
	jonathan.cameron@huawei.com
Cc: dave@stgolabs.net, dave.jiang@intel.com,
	alison.schofield@intel.com, vishal.l.verma@intel.com,
	ira.weiny@intel.com, linux-cxl@vger.kernel.org,
	linux-acpi@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] ACPI: add a boot parameter to disable parsing CFMWS during NUMA init
Date: Wed, 4 Mar 2026 15:55:23 +0800	[thread overview]
Message-ID: <751572d2-2665-4a67-8098-2f3a69668e0f@shopee.com> (raw)
In-Reply-To: <20260304064609.75585-1-haifeng.xu@shopee.com>



On 2026/3/4 14:46, Haifeng Xu wrote:
> For the machine Intel(R) Xeon(R) 6746E that supports CXL memory,
> the possible node is 20 (0-19). However, only two numa nodes (0-1)
> have memory and the rest (2-19) nodes detected by CEDT is memoryless.
> 
> The problems is that when creating many pods, the shrinker map size
> need to be expanded for all memory cgroups in expand_shrinker_info().
> If the number of possibles nodes is too large, the holding time of
> shrinker lock grows significantly.
> 
> In this case, there is no CXL memory inserted in the machine, those
> memoryless nodes are useless for us, so there is no need set them to
> 'numa_nodes_parsed'. After disabling parsing CFMWS, the pod creation
> time is reduced from over 10 minutes to approximately 150 seconds in
> our intertel test.
> 
> Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
> ---
>  drivers/acpi/numa/srat.c | 29 ++++++++++++++++++++++-------
>  include/acpi/acpi_numa.h |  6 ++++++
>  2 files changed, 28 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/acpi/numa/srat.c b/drivers/acpi/numa/srat.c
> index aa87ee1583a4..8716d70043fe 100644
> --- a/drivers/acpi/numa/srat.c
> +++ b/drivers/acpi/numa/srat.c
> @@ -31,6 +31,7 @@ static int node_to_pxm_map[MAX_NUMNODES]
>  
>  unsigned char acpi_srat_revision __initdata;
>  static int acpi_numa __initdata;
> +static int cfmws_numa __initdata;
>  
>  static int last_real_pxm;
>  
> @@ -39,6 +40,12 @@ void __init disable_srat(void)
>  	acpi_numa = -1;
>  }
>  
> +void __init disable_cfmws(void)
> +{
> +	cfmws_numa = -1;
> +}
> +
> +
>  int pxm_to_node(int pxm)
>  {
>  	if (pxm < 0 || pxm >= MAX_PXM_DOMAINS || numa_off)
> @@ -313,6 +320,12 @@ int __init srat_disabled(void)
>  	return acpi_numa < 0;
>  }
>  
> +int __init cfmws_disabled(void)
> +{
> +	return cfmws_numa < 0;
> +}
> +
> +
>  __weak int __init numa_fill_memblks(u64 start, u64 end)
>  {
>  	return NUMA_NO_MEMBLK;
> @@ -648,14 +661,16 @@ int __init acpi_numa_init(void)
>  	 */
>  
>  	/* fake_pxm is the next unused PXM value after SRAT parsing */
> -	for (i = 0, fake_pxm = -1; i < MAX_NUMNODES; i++) {
> -		if (node_to_pxm_map[i] > fake_pxm)
> -			fake_pxm = node_to_pxm_map[i];
> +	if (!cfmws_disabled()) {
> +		for (i = 0, fake_pxm = -1; i < MAX_NUMNODES; i++) {
> +			if (node_to_pxm_map[i] > fake_pxm)
> +				fake_pxm = node_to_pxm_map[i];
> +		}
> +		last_real_pxm = fake_pxm;
> +		fake_pxm++;
> +		acpi_table_parse_cedt(ACPI_CEDT_TYPE_CFMWS, acpi_parse_cfmws,
> +				      &fake_pxm);
>  	}
> -	last_real_pxm = fake_pxm;
> -	fake_pxm++;
> -	acpi_table_parse_cedt(ACPI_CEDT_TYPE_CFMWS, acpi_parse_cfmws,
> -			      &fake_pxm);
>  
>  	if (cnt < 0)
>  		return cnt;
> diff --git a/include/acpi/acpi_numa.h b/include/acpi/acpi_numa.h
> index 99b960bd473c..2435f60e56ce 100644
> --- a/include/acpi/acpi_numa.h
> +++ b/include/acpi/acpi_numa.h
> @@ -21,6 +21,7 @@ extern int fix_pxm_node_maps(int max_nid);
>  
>  extern void bad_srat(void);
>  extern int srat_disabled(void);
> +extern void disable_cfmws(void);
>  
>  #else				/* CONFIG_ACPI_NUMA */
>  static inline int fix_pxm_node_maps(int max_nid)
> @@ -30,6 +31,11 @@ static inline int fix_pxm_node_maps(int max_nid)
>  static inline void disable_srat(void)
>  {
>  }
> +
> +static inline void disable_cfmws(void)
> +{
> +}
> +
>  static inline int pxm_to_node(int pxm)
>  {
>  	return 0;

Sorry, This patch is deprecated. I'll send a new one.

  reply	other threads:[~2026-03-04  7:55 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-04  6:46 Haifeng Xu
2026-03-04  7:55 ` Haifeng Xu [this message]
2026-03-04  8:06 Haifeng Xu
2026-03-04 14:10 ` kernel test robot
2026-03-04 15:11 ` kernel test robot
2026-03-04 17:16 ` Gregory Price
2026-03-05  4:18   ` Haifeng Xu
2026-03-05  6:30     ` Gregory Price
2026-03-05  7:43       ` Haifeng Xu
2026-03-05 19:57 ` kernel test robot
2026-03-05 20:19 ` kernel test robot
2026-03-05 23:00 ` kernel test robot

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=751572d2-2665-4a67-8098-2f3a69668e0f@shopee.com \
    --to=haifeng.xu@shopee.com \
    --cc=alison.schofield@intel.com \
    --cc=dan.j.williams@intel.com \
    --cc=dave.jiang@intel.com \
    --cc=dave@stgolabs.net \
    --cc=ira.weiny@intel.com \
    --cc=jonathan.cameron@huawei.com \
    --cc=lenb@kernel.org \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-cxl@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rafael@kernel.org \
    --cc=vishal.l.verma@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

all inboxes | Powered by JetHome®