Message-ID: <20190625165832.00007f61@huawei.com>
Date:   Tue, 25 Jun 2019 16:58:32 +0100
From:   Jonathan Cameron <jonathan.cameron@...wei.com>
To:     Dan Williams <dan.j.williams@...el.com>
CC:     <x86@...nel.org>, Len Brown <lenb@...nel.org>,
        Keith Busch <keith.busch@...el.com>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        "Vishal Verma" <vishal.l.verma@...el.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        <peterz@...radead.org>, <ard.biesheuvel@...aro.org>,
        <tglx@...utronix.de>, <linux-acpi@...r.kernel.org>,
        <linux-nvdimm@...ts.01.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 10/10] acpi/numa/hmat: Register "specific purpose"
 memory as an "hmem" device

On Mon, 24 Jun 2019 11:20:26 -0700
Dan Williams <dan.j.williams@...el.com> wrote:

> Memory that has been tagged EFI_MEMORY_SP, and has performance
> properties described by the ACPI HMAT is expected to have an application
> specific consumer.
> 
> Those consumers may want 100% of the memory capacity to be reserved from
> any usage by the kernel. By default, with this enabling, a platform
> device is created to represent this differentiated resource.
> 
> The device-dax "hmem" driver claims these devices by default and
> provides an mmap interface for the target application.  If the
> administrator prefers, the hmem resource range can be made available to
> the core-mm via the device-dax hotplug facility, kmem, to online the
> memory with its own numa node.
> 
> This was tested with an emulated HMAT produced by qemu (with the pending
> HMAT enabling patches), and "efi_fake_mem=8G@9G:0x40000" on the kernel
> command line to mark the memory ranges associated with node2 and node3
> as EFI_MEMORY_SP.
> 
> qemu numa configuration options:
> 
> -numa node,mem=4G,cpus=0-19,nodeid=0
> -numa node,mem=4G,cpus=20-39,nodeid=1
> -numa node,mem=4G,nodeid=2
> -numa node,mem=4G,nodeid=3
> -numa dist,src=0,dst=0,val=10
> -numa dist,src=0,dst=1,val=21
> -numa dist,src=0,dst=2,val=21
> -numa dist,src=0,dst=3,val=21
> -numa dist,src=1,dst=0,val=21
> -numa dist,src=1,dst=1,val=10
> -numa dist,src=1,dst=2,val=21
> -numa dist,src=1,dst=3,val=21
> -numa dist,src=2,dst=0,val=21
> -numa dist,src=2,dst=1,val=21
> -numa dist,src=2,dst=2,val=10
> -numa dist,src=2,dst=3,val=21
> -numa dist,src=3,dst=0,val=21
> -numa dist,src=3,dst=1,val=21
> -numa dist,src=3,dst=2,val=21
> -numa dist,src=3,dst=3,val=10
> -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-latency,base-lat=10,latency=5
> -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=5
> -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-latency,base-lat=10,latency=10
> -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=10
> -numa hmat-lb,initiator=0,target=2,hierarchy=memory,data-type=access-latency,base-lat=10,latency=15
> -numa hmat-lb,initiator=0,target=2,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=15
> -numa hmat-lb,initiator=0,target=3,hierarchy=memory,data-type=access-latency,base-lat=10,latency=20
> -numa hmat-lb,initiator=0,target=3,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=20
> -numa hmat-lb,initiator=1,target=0,hierarchy=memory,data-type=access-latency,base-lat=10,latency=10
> -numa hmat-lb,initiator=1,target=0,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=10
> -numa hmat-lb,initiator=1,target=1,hierarchy=memory,data-type=access-latency,base-lat=10,latency=5
> -numa hmat-lb,initiator=1,target=1,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=5
> -numa hmat-lb,initiator=1,target=2,hierarchy=memory,data-type=access-latency,base-lat=10,latency=15
> -numa hmat-lb,initiator=1,target=2,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=15
> -numa hmat-lb,initiator=1,target=3,hierarchy=memory,data-type=access-latency,base-lat=10,latency=20
> -numa hmat-lb,initiator=1,target=3,hierarchy=memory,data-type=access-bandwidth,base-bw=20,bandwidth=20
> 
> Result:
> 
> # daxctl list -RDu
> [
>   {
>     "path":"\/platform\/hmem.1",
>     "id":1,
>     "size":"4.00 GiB (4.29 GB)",
>     "align":2097152,
>     "devices":[
>       {
>         "chardev":"dax1.0",
>         "size":"4.00 GiB (4.29 GB)"
>       }
>     ]
>   },
>   {
>     "path":"\/platform\/hmem.0",
>     "id":0,
>     "size":"4.00 GiB (4.29 GB)",
>     "align":2097152,
>     "devices":[
>       {
>         "chardev":"dax0.0",
>         "size":"4.00 GiB (4.29 GB)"
>       }
>     ]
>   }
> ]
> 
> # cat /proc/iomem
> [..]
> 240000000-43fffffff : Application Reserved
>   240000000-33fffffff : hmem.0
>     240000000-33fffffff : dax0.0
>   340000000-43fffffff : hmem.1
>     340000000-43fffffff : dax1.0
> 
> Cc: Len Brown <lenb@...nel.org>
> Cc: Keith Busch <keith.busch@...el.com>
> Cc: "Rafael J. Wysocki" <rjw@...ysocki.net>
> Cc: Vishal Verma <vishal.l.verma@...el.com>
> Cc: Jonathan Cameron <Jonathan.Cameron@...wei.com>
> Reviewed-by: Dave Hansen <dave.hansen@...ux.intel.com>
> Signed-off-by: Dan Williams <dan.j.williams@...el.com>
Looks good to me.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@...wei.com>

I like the approach of this patch set in general as it gives the flexibility to
the administrator of a given system.

Jonathan

> ---
>  drivers/acpi/numa/Kconfig |    1 
>  drivers/acpi/numa/hmat.c  |  132 +++++++++++++++++++++++++++++++++++++++++----
>  2 files changed, 122 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/acpi/numa/Kconfig b/drivers/acpi/numa/Kconfig
> index d14582387ed0..c1be746e111a 100644
> --- a/drivers/acpi/numa/Kconfig
> +++ b/drivers/acpi/numa/Kconfig
> @@ -8,6 +8,7 @@ config ACPI_HMAT
>  	bool "ACPI Heterogeneous Memory Attribute Table Support"
>  	depends on ACPI_NUMA
>  	select HMEM_REPORTING
> +	select MEMREGION
>  	help
>  	 If set, this option has the kernel parse and report the
>  	 platform's ACPI HMAT (Heterogeneous Memory Attributes Table),
> diff --git a/drivers/acpi/numa/hmat.c b/drivers/acpi/numa/hmat.c
> index 1d329c4af3bf..622c5feb3aa0 100644
> --- a/drivers/acpi/numa/hmat.c
> +++ b/drivers/acpi/numa/hmat.c
> @@ -8,11 +8,16 @@
>   * the applicable attributes with the node's interfaces.
>   */
>  
> +#define pr_fmt(fmt) "acpi/hmat: " fmt
> +#define dev_fmt(fmt) "acpi/hmat: " fmt
> +
>  #include <linux/acpi.h>
>  #include <linux/bitops.h>
>  #include <linux/device.h>
>  #include <linux/init.h>
>  #include <linux/list.h>
> +#include <linux/mm.h>
> +#include <linux/platform_device.h>
>  #include <linux/list_sort.h>
>  #include <linux/node.h>
>  #include <linux/sysfs.h>
> @@ -40,6 +45,7 @@ struct memory_target {
>  	struct list_head node;
>  	unsigned int memory_pxm;
>  	unsigned int processor_pxm;
> +	struct resource memregions;
>  	struct node_hmem_attrs hmem_attrs;
>  };
>  
> @@ -92,21 +98,35 @@ static __init void alloc_memory_initiator(unsigned int cpu_pxm)
>  	list_add_tail(&initiator->node, &initiators);
>  }
>  
> -static __init void alloc_memory_target(unsigned int mem_pxm)
> +static __init void alloc_memory_target(unsigned int mem_pxm,
> +		resource_size_t start, resource_size_t len)
>  {
>  	struct memory_target *target;
>  
>  	target = find_mem_target(mem_pxm);
> -	if (target)
> -		return;
> -
> -	target = kzalloc(sizeof(*target), GFP_KERNEL);
> -	if (!target)
> -		return;
> +	if (!target) {
> +		target = kzalloc(sizeof(*target), GFP_KERNEL);
> +		if (!target)
> +			return;
> +		target->memory_pxm = mem_pxm;
> +		target->processor_pxm = PXM_INVAL;
> +		target->memregions = (struct resource) {
> +			.name	= "ACPI mem",
> +			.start	= 0,
> +			.end	= -1,
> +			.flags	= IORESOURCE_MEM,
> +		};
> +		list_add_tail(&target->node, &targets);
> +	}
>  
> -	target->memory_pxm = mem_pxm;
> -	target->processor_pxm = PXM_INVAL;
> -	list_add_tail(&target->node, &targets);
> +	/*
> +	 * There are potentially multiple ranges per PXM, so record each
> +	 * in the per-target memregions resource tree.
> +	 */
> +	if (!__request_region(&target->memregions, start, len, "memory target",
> +				IORESOURCE_MEM))
> +		pr_warn("failed to reserve %#llx - %#llx in pxm: %d\n",
> +				start, start + len, mem_pxm);
>  }
>  
>  static __init const char *hmat_data_type(u8 type)
> @@ -428,7 +448,7 @@ static __init int srat_parse_mem_affinity(union acpi_subtable_headers *header,
>  		return -EINVAL;
>  	if (!(ma->flags & ACPI_SRAT_MEM_ENABLED))
>  		return 0;
> -	alloc_memory_target(ma->proximity_domain);
> +	alloc_memory_target(ma->proximity_domain, ma->base_address, ma->length);
>  	return 0;
>  }
>  
> @@ -580,6 +600,81 @@ static __init void hmat_register_target_perf(struct memory_target *target)
>  	node_set_perf_attrs(mem_nid, &target->hmem_attrs, 0);
>  }
>  
> +static __init void hmat_register_target_device(struct memory_target *target,
> +		struct resource *r)
> +{
> +	/* define a clean / non-busy resource for the platform device */
> +	struct resource res = {
> +		.start = r->start,
> +		.end = r->end,
> +		.flags = IORESOURCE_MEM,
> +	};
> +	struct platform_device *pdev;
> +	struct memregion_info info;
> +	int rc, id;
> +
> +	rc = region_intersects(res.start, resource_size(&res), IORESOURCE_MEM,
> +			IORES_DESC_APPLICATION_RESERVED);
> +	if (rc != REGION_INTERSECTS)
> +		return;
> +
> +	id = memregion_alloc(GFP_KERNEL);
> +	if (id < 0) {
> +		pr_err("memregion allocation failure for %pr\n", &res);
> +		return;
> +	}
> +
> +	pdev = platform_device_alloc("hmem", id);
> +	if (!pdev) {
> +		pr_err("hmem device allocation failure for %pr\n", &res);
> +		goto out_pdev;
> +	}
> +
> +	pdev->dev.numa_node = acpi_map_pxm_to_online_node(target->memory_pxm);
> +	info = (struct memregion_info) {
> +		.target_node = acpi_map_pxm_to_node(target->memory_pxm),
> +	};
> +	rc = platform_device_add_data(pdev, &info, sizeof(info));
> +	if (rc < 0) {
> +		pr_err("hmem memregion_info allocation failure for %pr\n", &res);
> +		goto out_pdev;
> +	}
> +
> +	rc = platform_device_add_resources(pdev, &res, 1);
> +	if (rc < 0) {
> +		pr_err("hmem resource allocation failure for %pr\n", &res);
> +		goto out_resource;
> +	}
> +
> +	rc = platform_device_add(pdev);
> +	if (rc < 0) {
> +		dev_err(&pdev->dev, "device add failed for %pr\n", &res);
> +		goto out_resource;
> +	}
> +
> +	return;
> +
> +out_resource:
> +	put_device(&pdev->dev);
> +out_pdev:
> +	memregion_free(id);
> +}
> +
> +static __init void hmat_register_target_devices(struct memory_target *target)
> +{
> +	struct resource *res;
> +
> +	/*
> +	 * Do not bother creating devices if no driver is available to
> +	 * consume them.
> +	 */
> +	if (!IS_ENABLED(CONFIG_DEV_DAX_HMEM))
> +		return;
> +
> +	for (res = target->memregions.child; res; res = res->sibling)
> +		hmat_register_target_device(target, res);
> +}
> +
>  static __init void hmat_register_targets(void)
>  {
>  	struct memory_target *target;
> @@ -587,6 +682,12 @@ static __init void hmat_register_targets(void)
>  	list_for_each_entry(target, &targets, node) {
>  		int nid = pxm_to_node(target->memory_pxm);
>  
> +		/*
> +		 * Devices may belong to either an offline or online
> +		 * node, so unconditionally add them.
> +		 */
> +		hmat_register_target_devices(target);
> +
>  		/*
>  		 * Skip offline nodes. This can happen when memory
>  		 * marked EFI_MEMORY_SP, "specific purpose", is applied
> @@ -608,7 +709,16 @@ static __init void hmat_free_structures(void)
>  	struct memory_initiator *initiator, *inext;
>  
>  	list_for_each_entry_safe(target, tnext, &targets, node) {
> +		struct resource *res, *res_next;
> +
>  		list_del(&target->node);
> +		res = target->memregions.child;
> +		while (res) {
> +			res_next = res->sibling;
> +			__release_region(&target->memregions, res->start,
> +					resource_size(res));
> +			res = res_next;
> +		}
>  		kfree(target);
>  	}
>  
> 

Powered by blists - more mailing lists