linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: lorenzo.pieralisi@arm.com (Lorenzo Pieralisi)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH v2 2/2] acpi, gicv3-its, numa: Adding numa node mapping for gic-its units
Date: Tue, 20 Jun 2017 16:17:33 +0100	[thread overview]
Message-ID: <20170620151733.GC11351@red-moon> (raw)
In-Reply-To: <1497942437-1390-3-git-send-email-ganapatrao.kulkarni@cavium.com>

Hi Ganapatrao,

On Tue, Jun 20, 2017 at 12:37:17PM +0530, Ganapatrao Kulkarni wrote:
> Add code to parse SRAT ITS Affinity sub table as defined in ACPI 6.2
> Later in per device probe, ITS devices are mapped to
> numa node using ITS id to proximity domain mapping.
> 
> Signed-off-by: Ganapatrao Kulkarni <ganapatrao.kulkarni@cavium.com>
> ---
>  drivers/irqchip/irq-gic-v3-its.c | 76 +++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 75 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
> index 45ea1933..5865a75 100644
> --- a/drivers/irqchip/irq-gic-v3-its.c
> +++ b/drivers/irqchip/irq-gic-v3-its.c
> @@ -1833,6 +1833,78 @@ static int __init its_of_probe(struct device_node *node)
>  
>  #define ACPI_GICV3_ITS_MEM_SIZE (SZ_128K)
>  
> +#ifdef CONFIG_ACPI_NUMA
> +struct its_srat_map {
> +	u32			numa_node;  /* numa node id */
> +	u32			its_id;  /* GIC ITS ID */
> +	struct list_head	entry;
> +};
> +
> +static LIST_HEAD(its_srat_maps);
> +
> +static int acpi_get_its_numa_node(u32 its_id)
> +{
> +	struct its_srat_map *srat_map;
> +
> +	list_for_each_entry(srat_map, &its_srat_maps, entry) {
> +		if (its_id == srat_map->its_id)
> +			return srat_map->numa_node;
> +	}
> +	return NUMA_NO_NODE;
> +}

I do not want to come across as pedantic and I understand that the API
to parse static ACPI tables sucks (ie in theory you could avoid stashing
nodes altogether and just use a global variable to parse an ITS ID at
a time but it sucks to go through the SRAT multiple times) but would not
it be easier to just declare an __initdata static array and counter and
be done with this ? We do not need this stuff after boot so I do not see
a point in keeping it in memory and in allocating the nodes dynamically.

Apologies if I was not clear in my comments to the previous set, I did
not like stashing ITS PXM values in arch/arm64 code but I understand you
need an array (__initdata) to avoid reading the SRAT multiple times.

Thanks,
Lorenzo

> +
> +static int __init
> +gic_acpi_parse_srat_its(struct acpi_subtable_header *header,
> +			 const unsigned long end)
> +{
> +	int pxm, node;
> +	struct its_srat_map *srat_map;
> +	struct acpi_srat_its_affinity *its_affinity;
> +
> +	its_affinity = (struct acpi_srat_its_affinity *)header;
> +	if (!its_affinity)
> +		return -EINVAL;
> +
> +	if (its_affinity->header.length <
> +			sizeof(struct acpi_srat_its_affinity)) {
> +		pr_err("SRAT:ITS: Invalid SRAT header length: %d\n",
> +			its_affinity->header.length);
> +		return -EINVAL;
> +	}
> +
> +	pxm = its_affinity->proximity_domain;
> +	node = acpi_map_pxm_to_node(pxm);
> +
> +	if (node == NUMA_NO_NODE || node >= MAX_NUMNODES) {
> +		pr_err("SRAT:ITS Invalid numa node %d\n", node);
> +		return -EINVAL;
> +	}
> +
> +	srat_map = kzalloc(sizeof(*srat_map), GFP_KERNEL);
> +	if (!srat_map)
> +		return -ENOMEM;
> +
> +	srat_map->numa_node = node;
> +	srat_map->its_id = its_affinity->its_id;
> +	list_add(&srat_map->entry, &its_srat_maps);
> +	pr_info("ACPI: NUMA: SRAT: ITS: PXM %d -> ITS_ID %d -> NODE %d\n",
> +		srat_map->its_id, pxm, node);
> +
> +	return 0;
> +}
> +
> +static int __init acpi_table_parse_srat_its(void)
> +{
> +	return acpi_table_parse_entries(ACPI_SIG_SRAT,
> +					sizeof(struct acpi_table_srat),
> +					ACPI_SRAT_TYPE_GIC_ITS_AFFINITY,
> +					gic_acpi_parse_srat_its, 0);
> +}
> +#else
> +#define acpi_table_parse_srat_its(void) { }
> +#define acpi_get_its_numa_node(its_id) NUMA_NO_NODE
> +#endif
> +
>  static int __init gic_acpi_parse_madt_its(struct acpi_subtable_header *header,
>  					  const unsigned long end)
>  {
> @@ -1861,7 +1933,8 @@ static int __init gic_acpi_parse_madt_its(struct acpi_subtable_header *header,
>  		goto dom_err;
>  	}
>  
> -	err = its_probe_one(&res, dom_handle, NUMA_NO_NODE);
> +	err = its_probe_one(&res, dom_handle,
> +			acpi_get_its_numa_node(its_entry->translation_id));
>  	if (!err)
>  		return 0;
>  
> @@ -1873,6 +1946,7 @@ static int __init gic_acpi_parse_madt_its(struct acpi_subtable_header *header,
>  
>  static void __init its_acpi_probe(void)
>  {
> +	acpi_table_parse_srat_its();
>  	acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_TRANSLATOR,
>  			      gic_acpi_parse_madt_its, 0);
>  }
> -- 
> 1.8.1.4
> 

  reply	other threads:[~2017-06-20 15:17 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-06-20  7:07 [PATCH v2 0/2] acpi, gicv3-its, numa: Adding numa node mapping for ITS Ganapatrao Kulkarni
2017-06-20  7:07 ` [PATCH v2 1/2] ACPICA: ACPI 6.2: Add support for new SRAT subtable Ganapatrao Kulkarni
2017-06-20  7:07 ` [PATCH v2 2/2] acpi, gicv3-its, numa: Adding numa node mapping for gic-its units Ganapatrao Kulkarni
2017-06-20 15:17   ` Lorenzo Pieralisi [this message]
2017-06-20 17:17     ` Ganapatrao Kulkarni
2017-06-20 17:30       ` Lorenzo Pieralisi

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=20170620151733.GC11351@red-moon \
    --to=lorenzo.pieralisi@arm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    /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;
as well as URLs for NNTP newsgroup(s).