From: Ben Horgan <ben.horgan@arm.com>
To: Andre Przywara <andre.przywara@arm.com>,
Lorenzo Pieralisi <lpieralisi@kernel.org>,
Hanjun Guo <guohanjun@huawei.com>,
Sudeep Holla <sudeep.holla@kernel.org>,
Catalin Marinas <catalin.marinas@arm.com>,
Will Deacon <will@kernel.org>,
"Rafael J . Wysocki" <rafael@kernel.org>,
Len Brown <lenb@kernel.org>, James Morse <james.morse@arm.com>,
Reinette Chatre <reinette.chatre@intel.com>,
Fenghua Yu <fenghuay@nvidia.com>
Cc: Jonathan Cameron <jic23@kernel.org>,
linux-acpi@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/5] arm_mpam: Parse the rest of the ACPI table
Date: Fri, 8 May 2026 11:21:20 +0100 [thread overview]
Message-ID: <f12b8d73-7c8e-43e7-b9ce-d80a7c3f10c8@arm.com> (raw)
In-Reply-To: <20260429141339.3171205-2-andre.przywara@arm.com>
Hi Andre,
On 4/29/26 15:13, Andre Przywara wrote:
> From: James Morse <james.morse@arm.com>
>
> The MPAM ACPI table lists the MPAM MSCs and indicates which resources
> in the system they control. Not everything this table can describe is
> supported by resctrl, e.g. memory-side-caches.
>
> Add the additional table parsing to avoid reporting these as 'unknown'
> to the MPAM driver. This allows class+component hierarchys to be built.
>
> Until resctrl has support for any of these resources, users would be
> in-kernel managers of a resource/PARTID or perf to query bandwidth
> counters on a resource resctrl is unaware of.
Does this patch give us anything needed for MPAM-Fb? It seems to just add
parsing for things we don't support in the MPAM driver.
Thanks,
Ben
>
> Signed-off-by: James Morse <james.morse@arm.com>
> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
> ---
> drivers/acpi/arm64/mpam.c | 91 +++++++++++++++++++++++++++++++++++++--
> 1 file changed, 88 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/acpi/arm64/mpam.c b/drivers/acpi/arm64/mpam.c
> index 84963a20c3e7..99c2bdbb3314 100644
> --- a/drivers/acpi/arm64/mpam.c
> +++ b/drivers/acpi/arm64/mpam.c
> @@ -95,17 +95,51 @@ static void acpi_mpam_parse_irqs(struct platform_device *pdev,
> res[(*res_idx)++] = DEFINE_RES_IRQ_NAMED(irq, "error");
> }
>
> -static int acpi_mpam_parse_resource(struct mpam_msc *msc,
> +#define UUID_MPAM_INTERCONNECT_TABLE "fe2bd645-033b-49e6-9479-2e0b8b21d1cd"
> +
> +struct acpi_mpam_interconnect_descriptor_table {
> + u8 type_uuid[16];
> + u32 num_descriptors;
> +};
> +
> +struct acpi_mpam_interconnect_descriptor {
> + u32 source_id;
> + u32 destination_id;
> + u8 link_type;
> + u8 reserved[3];
> +};
> +
> +static int acpi_mpam_parse_resource(struct acpi_mpam_msc_node *tbl_msc,
> + struct mpam_msc *msc,
> struct acpi_mpam_resource_node *res)
> {
> + struct acpi_mpam_interconnect_descriptor_table *tbl_int_tbl;
> + struct acpi_mpam_interconnect_descriptor *tbl_int;
> + guid_t int_tbl_uuid, spec_uuid;
> int level, nid;
> u32 cache_id;
> + off_t offset;
>
> + /*
> + * Class IDs are somewhat arbitrary, but need to be co-ordinated.
> + * 0-N are caches,
> + * 64, 65: Interconnect, but ideally these would appear between the
> + * classes the controls are adjacent to.
> + * 128: SMMU,
> + * 192-192+level: Memory Side Caches, nothing checks that N is a
> + * small number.
> + * 255: Memory Controllers
> + *
> + * ACPI devices would need a class id allocated based on the _HID.
> + *
> + * Classes that the mpam driver can't currently plumb into resctrl
> + * are registered as UNKNOWN.
> + */
> switch (res->locator_type) {
> case ACPI_MPAM_LOCATION_TYPE_PROCESSOR_CACHE:
> cache_id = res->locator.cache_locator.cache_reference;
> level = find_acpi_cache_level_from_id(cache_id);
> - if (level <= 0) {
> + if (level <= 0 || level >= 64) {
> pr_err_once("Bad level (%d) for cache with id %u\n", level, cache_id);
> return -EINVAL;
> }
> @@ -120,6 +154,57 @@ static int acpi_mpam_parse_resource(struct mpam_msc *msc,
> }
> return mpam_ris_create(msc, res->ris_index, MPAM_CLASS_MEMORY,
> MPAM_CLASS_ID_DEFAULT, nid);
> + case ACPI_MPAM_LOCATION_TYPE_SMMU:
> + return mpam_ris_create(msc, res->ris_index, MPAM_CLASS_UNKNOWN,
> + 128, res->locator.smmu_locator.smmu_interface);
> + case ACPI_MPAM_LOCATION_TYPE_MEMORY_CACHE:
> + cache_id = res->locator.mem_cache_locator.reference;
> + level = res->locator.mem_cache_locator.level;
> + if (192 + level >= 255) {
> + pr_err_once("Bad level for memory side cache with reference %u\n",
> + cache_id);
> + return -EINVAL;
> + }
> +
> + return mpam_ris_create(msc, res->ris_index, MPAM_CLASS_CACHE,
> + 192 + level, cache_id);
> +
> + case ACPI_MPAM_LOCATION_TYPE_INTERCONNECT:
> + /* Find the descriptor table, and check it lands in the parent msc */
> + offset = res->locator.interconnect_ifc_locator.inter_connect_desc_tbl_off;
> + if (offset >= tbl_msc->length) {
> + pr_err_once("Bad offset for interconnect descriptor on msc %u\n",
> + tbl_msc->identifier);
> + return -EINVAL;
> + }
> + tbl_int_tbl = ACPI_ADD_PTR(struct acpi_mpam_interconnect_descriptor_table,
> + tbl_msc, offset);
> + guid_parse(UUID_MPAM_INTERCONNECT_TABLE, &spec_uuid);
> + import_guid(&int_tbl_uuid, tbl_int_tbl->type_uuid);
> + if (guid_equal(&spec_uuid, &int_tbl_uuid)) {
> + pr_err_once("Bad UUID for interconnect descriptor on msc %u\n",
> + tbl_msc->identifier);
> + return -EINVAL;
> + }
> +
> + offset += sizeof(*tbl_int_tbl);
> + offset += tbl_int_tbl->num_descriptors * sizeof(*tbl_int);
> + if (offset >= tbl_msc->length) {
> + pr_err_once("Bad num_descriptors for interconnect descriptor on msc %u\n",
> + tbl_msc->identifier);
> + return -EINVAL;
> + }
> +
> + tbl_int = ACPI_ADD_PTR(struct acpi_mpam_interconnect_descriptor,
> + tbl_int_tbl, sizeof(*tbl_int_tbl));
> + cache_id = tbl_int->source_id;
> +
> + /* Unknown link type? */
> + if (tbl_int->link_type != 0 && tbl_int->link_type == 1)
> + return 0;
> +
> + return mpam_ris_create(msc, res->ris_index, MPAM_CLASS_UNKNOWN,
> + 64 + tbl_int->link_type, cache_id);
> default:
> /* These get discovered later and are treated as unknown */
> return 0;
> @@ -150,7 +235,7 @@ int acpi_mpam_parse_resources(struct mpam_msc *msc,
> return -EINVAL;
> }
>
> - err = acpi_mpam_parse_resource(msc, resource);
> + err = acpi_mpam_parse_resource(tbl_msc, msc, resource);
> if (err)
> return err;
>
next prev parent reply other threads:[~2026-05-08 10:21 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-29 14:13 [PATCH 0/5] arm_mpam: Add MPAM-Fb firmware support Andre Przywara
2026-04-29 14:13 ` [PATCH 1/5] arm_mpam: Parse the rest of the ACPI table Andre Przywara
2026-05-08 10:21 ` Ben Horgan [this message]
2026-04-29 14:13 ` [PATCH 2/5] arm_mpam: Split the locking around the mon_sel registers Andre Przywara
2026-05-08 10:23 ` Ben Horgan
2026-04-29 14:13 ` [PATCH 3/5] arm_mpam: add MPAM-Fb MSC firmware access support Andre Przywara
2026-05-08 10:33 ` Ben Horgan
2026-04-29 14:13 ` [PATCH 4/5] arm_mpam: prevent MPAM-Fb accesses inside IRQ handler Andre Przywara
2026-04-29 14:13 ` [PATCH 5/5] arm_mpam: detect and enable MPAM-Fb PCC support Andre Przywara
2026-04-30 8:35 ` Sudeep Holla
2026-04-30 9:20 ` Andre Przywara
2026-04-30 10:25 ` Sudeep Holla
2026-05-08 10:48 ` Ben Horgan
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=f12b8d73-7c8e-43e7-b9ce-d80a7c3f10c8@arm.com \
--to=ben.horgan@arm.com \
--cc=andre.przywara@arm.com \
--cc=catalin.marinas@arm.com \
--cc=fenghuay@nvidia.com \
--cc=guohanjun@huawei.com \
--cc=james.morse@arm.com \
--cc=jic23@kernel.org \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lpieralisi@kernel.org \
--cc=rafael@kernel.org \
--cc=reinette.chatre@intel.com \
--cc=sudeep.holla@kernel.org \
--cc=will@kernel.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