Re: [PATCH v10 12/14] arm_mpam: detect and enable MPAM-Fb PCC support

[Ben Horgan <ben.horgan@arm.com>]

Hi Andre,

On 21/09/2026 17:09, Andre Przywara wrote:
> Hi Ben,
> 
> On 9/11/26 17:09, Ben Horgan wrote:
>> Hi Andre,
>>
>> On 11/09/2026 12:28, Andre Przywara wrote:
>>> The Arm MPAM-Fb specification [1] describes a protocol to access MSC
>>> registers through a firmware interface. This requires a shared memory
>>> region to hold the message, and a mailbox to trigger the access.
>>> For ACPI this is wrapped as a PCC channel, described using existing
>>> ACPI abstractions.
>>>
>>> Add code to parse those PCC table descriptions associated with an MSC,
>>> and store the parsed information in the MSC struct.
>>> There can be multiple PCC channels, and each channel can serve multiple
>>> MSCs, so we need to keep track of the channel usage, using a list and
>>> a refcount.
>>>
>>> This will be used by the MPAM-Fb access wrapper code.
>>>
>>> [1] https://developer.arm.com/documentation/den0144/latest
>>>
>>> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
>>> Tested-by: Ritwick Sharma <ritwick.sharma@arm.com> # on Arm AGI CPU
>>> Reviewed-by: Jonathan Cameron <jonathan.cameron@oss.qualcomm.com>
>>> Tested-by: Srivathsa L Rao <srivathsa.rao@oss.qualcomm.com>
>>> Reviewed-by: Srivathsa L Rao <srivathsa.rao@oss.qualcomm.com>
>>> Reviewed-by: Gavin Shan <gshan@redhat.com>
>>> ---
>>>   drivers/resctrl/Kconfig         |   1 +
>>>   drivers/resctrl/mpam_devices.c  | 119 +++++++++++++++++++++++++++++++-
>>>   drivers/resctrl/mpam_fb.c       |  41 +++++++++++
>>>   drivers/resctrl/mpam_internal.h |   2 +
>>>   4 files changed, 161 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/resctrl/Kconfig b/drivers/resctrl/Kconfig
>>> index 9591d792736e5..3f81638f57be3 100644
>>> --- a/drivers/resctrl/Kconfig
>>> +++ b/drivers/resctrl/Kconfig
>>> @@ -2,6 +2,7 @@ menuconfig ARM64_MPAM_DRIVER
>>>       bool "MPAM driver"
>>>       depends on ARM64 && ARM64_MPAM
>>>       select ACPI_MPAM if ACPI
>>> +    select PCC if ACPI
>>>       select MAILBOX
>>>       help
>>>         Memory System Resource Partitioning and Monitoring (MPAM) driver for
>>> diff --git a/drivers/resctrl/mpam_devices.c b/drivers/resctrl/mpam_devices.c
>>> index 14dc8461e4b8f..150a529458388 100644
>>> --- a/drivers/resctrl/mpam_devices.c
>>> +++ b/drivers/resctrl/mpam_devices.c
>>> @@ -20,14 +20,19 @@
>>>   #include <linux/irqdesc.h>
>>>   #include <linux/list.h>
>>>   #include <linux/lockdep.h>
>>> +#include <linux/mailbox_client.h>
>>>   #include <linux/mutex.h>
>>>   #include <linux/platform_device.h>
>>>   #include <linux/printk.h>
>>> +#include <linux/property.h>
>>>   #include <linux/srcu.h>
>>>   #include <linux/spinlock.h>
>>>   #include <linux/types.h>
>>>   #include <linux/workqueue.h>
>>>   +#include <acpi/pcc.h>
>>> +#include <acpi/acpi_io.h>
>>> +
>>>   #include "mpam_internal.h"
>>>     /* Values for the T241 errata workaround */
>>> @@ -50,6 +55,87 @@ static LIST_HEAD(mpam_all_msc);
>>>     struct srcu_struct mpam_srcu;
>>>   +/* PCC channels might be serving multiple MSCs, so keep a refcounted list. */
>>> +static DEFINE_MUTEX(pcc_chan_list_lock);
>>> +static LIST_HEAD(pcc_chan_list);
>>> +
>>> +static void mpam_pcc_chan_release(struct kref *ref)
>>> +{
>>> +    struct mpam_pcc_chan *cur = container_of(ref, struct mpam_pcc_chan,
>>> +                         refcount);
>>> +
>>> +    pcc_mbox_free_channel(cur->pcc_chan);
>>> +    list_del(&cur->pcc_chans);
>>> +    mutex_destroy(&cur->pcc_chan_lock);
>>> +    kfree(cur);
>>> +}
>>> +
>>> +static struct mpam_pcc_chan *mpam_pcc_chan_get(struct device *dev,
>>> +                           int subspace_id)
>>> +{
>>> +    struct mpam_pcc_chan *cur;
>>> +
>>> +    guard(mutex)(&pcc_chan_list_lock);
>>> +
>>> +    list_for_each_entry(cur, &pcc_chan_list, pcc_chans) {
>>> +        if (cur->subspace_id == subspace_id) {
>>> +            kref_get(&cur->refcount);
>>> +
>>> +            return cur;
>>> +        }
>>> +    }
>>> +
>>> +    cur = kzalloc_obj(*cur);
>>> +    if (!cur)
>>> +        return ERR_PTR(-ENOMEM);
>>> +
>>> +    cur->pcc_cl.tx_block = true;
>>> +
>>> +    cur->pcc_chan = pcc_mbox_request_channel(&cur->pcc_cl, subspace_id);
>>> +    if (IS_ERR(cur->pcc_chan)) {
>>> +        long err = PTR_ERR(cur->pcc_chan);
>>> +
>>> +        kfree(cur);
>>> +        return ERR_PTR(err);
>>> +    }
>>> +
>>> +    /*
>>> +     * Timeout based on the "nominal latency" in us, from the
>>> +     * PCC ACPI table. tx_tout is in ms.
>>> +     * Add some margin here to be on the safe side.
>>> +     */
>>> +    cur->pcc_cl.tx_tout = DIV_ROUND_UP(cur->pcc_chan->latency * 5, 1000);
>>> +
>>> +    mutex_init(&cur->pcc_chan_lock);
>>> +
>>> +    cur->subspace_id = subspace_id;
>>> +    kref_init(&cur->refcount);
>>> +
>>> +    list_add_tail(&cur->pcc_chans, &pcc_chan_list);
>>> +
>>> +    return cur;
>>> +}
>>> +
>>> +static int mpam_pcc_chan_put(struct mpam_pcc_chan *pcc_chan)
>>> +{
>>> +    struct mpam_pcc_chan *cur, *tmp;
>>> +
>>> +    if (!pcc_chan)
>>> +        return 0;
>>> +
>>> +    guard(mutex)(&pcc_chan_list_lock);
>>> +
>>> +    list_for_each_entry_safe(cur, tmp, &pcc_chan_list, pcc_chans) {
>>> +        if (cur == pcc_chan) {
>>> +            kref_put(&cur->refcount, mpam_pcc_chan_release);
>>> +
>>> +            return 0;
>>> +        }
>>> +    }
>>> +
>>> +    return -ENOENT;
>>> +}
>>> +
>>>   /*
>>>    * Number of MSCs that have been probed. Once all MSCs have been probed MPAM
>>>    * can be enabled.
>>> @@ -2285,6 +2371,8 @@ static void mpam_msc_drv_remove(struct platform_device *pdev)
>>>       if (!msc)
>>>           return;
>>>   +    mpam_pcc_chan_put(msc->pcc_chan);
>>> +
>>>       mutex_lock(&mpam_list_lock);
>>>       mpam_msc_destroy(msc);
>>>       mutex_unlock(&mpam_list_lock);
>>> @@ -2295,7 +2383,7 @@ static void mpam_msc_drv_remove(struct platform_device *pdev)
>>>   static struct mpam_msc *do_mpam_msc_drv_probe(struct platform_device *pdev)
>>>   {
>>>       int err;
>>> -    u32 tmp;
>>> +    u32 pcc_subspace_id;
>>>       struct mpam_msc *msc;
>>>       struct resource *msc_res;
>>>       struct device *dev = &pdev->dev;
>>> @@ -2339,7 +2427,7 @@ static struct mpam_msc *do_mpam_msc_drv_probe(struct platform_device *pdev)
>>>       if (err)
>>>           return ERR_PTR(err);
>>>   -    if (device_property_read_u32(&pdev->dev, "pcc-channel", &tmp))
>>> +    if (device_property_read_u32(dev, "pcc-channel", &pcc_subspace_id))
>>>           msc->iface = MPAM_IFACE_MMIO;
>>>       else
>>>           msc->iface = MPAM_IFACE_PCC;
>>> @@ -2360,6 +2448,33 @@ static struct mpam_msc *do_mpam_msc_drv_probe(struct platform_device *pdev)
>>>           }
>>>           msc->mapped_hwpage_sz = msc_res->end - msc_res->start;
>>>           msc->mapped_hwpage = io;
>>> +    } else if (msc->iface == MPAM_IFACE_PCC) {
>>> +        u32 mpam_fb_msc_id;
>>> +
>>> +        msc->pcc_chan = mpam_pcc_chan_get(dev, pcc_subspace_id);
>>> +        if (IS_ERR(msc->pcc_chan)) {
>>> +            pr_err("Failed to request MSC PCC channel\n");
>>> +            return ERR_CAST(msc->pcc_chan);
>>> +        }
>>> +
>>> +        err = mpam_fb_check_shared_buffer_size(msc);
>>> +        if (err) {
>>> +            mpam_pcc_chan_put(msc->pcc_chan);
>>> +
>>> +            return ERR_PTR(err);
>>> +        }
>>> +
>>> +        err = mpam_fb_check_protocol_version(msc);
>>> +        if (err) {
>>> +            mpam_pcc_chan_put(msc->pcc_chan);
>>> +
>>> +            return ERR_PTR(err);
>>> +        }
>>> +
>>> +        if (device_property_read_u32(&pdev->dev, "mpam-fb-msc-id",
>>> +                         &mpam_fb_msc_id))
>>> +            mpam_fb_msc_id = msc->id;
>>> +        msc->mpam_fb_msc_id = mpam_fb_msc_id;
>>
>> This looks like the change earlier in the series to drivers/acpi/arm64/mpam.c makes no functional
>> effect unless the proposed spec changes are accepted; the default is msc->id. I'd just move the
> 
> Yes, but it reduces churn, and more importantly, is more true to the MPAM-Fb spec, which says:
> "The MSC identifiers are sequential and start from 0 for each agent which uses this protocol."
> So this requires that we have per-channel IDs for the MSCs, that's why I think it's justified to
> introduce this already now, even when it's somewhat redundant as the code stands now, as we always
> assign the system-wide ID to it.
> 
> I can of course move that out (and have that already on top of v9, internally), but wanted to check
> this with you.

I don't really mind. If Lorenzo/Sudeep are happy with doing the preparation work in
drivers/acpi/arm64/mpam.c then I'm ok with it. It's currently a non-functional change but does make
the expected path forward slightly clearer.

> 
>> earlier drivers/acpi/arm64/mpam.c to your last patch and introduce this device_property_read_u32()
>> their too (or a separate patch if it's easier for anyone). That way all but the last two patches can
>> be considered for merging without knowing what's going to happen with the spec.
> 
> Well, sure, though I think a separation of MPAM-FB MSC IDs and system-wide MSC IDs is inevitable,
> regardless of which spec is changed and how the numbers are assigned, exactly.

It seems so.

Thanks,

Ben

> >> The struct msc property mpam_fb_msc_id is a bit long. Better as msc->fb_id?
> 
> Sure, can shorten that.
> 
> Cheers,
> Andre.
> 
>>>       } else {
>>>           return ERR_PTR(-EINVAL);
>>>       }
>>> diff --git a/drivers/resctrl/mpam_fb.c b/drivers/resctrl/mpam_fb.c
>>> index 98212f761de38..8fc1d78a7b5ab 100644
>>> --- a/drivers/resctrl/mpam_fb.c
>>> +++ b/drivers/resctrl/mpam_fb.c
>>> @@ -20,10 +20,15 @@
>>>   #define MPAM_FB_PROTOCOL_ID        0x1a
>>>     #define MPAM_PROTOCOL_VERSION_CMD    0x0
>>> +#define MPAM_FB_VERSION_MAJOR_MASK        GENMASK_U32(31, 16)
>>> +#define MPAM_FB_VERSION_MINOR_MASK        GENMASK_U32(15, 0)
>>> +
>>>   #define MPAM_MSC_READ_CMD        0x4
>>>   #define MPAM_MSC_WRITE_CMD        0x5
>>>     #define MPAM_FB_PROT_HEADER_LEN        sizeof(u32)
>>> +/* The longest message is MPAM_MSC_WRITE, with 4 parameters. */
>>> +#define MPAM_FB_MAX_MSG_SIZE        (4 * sizeof(u32))
>>>     #define MPAM_FB_SUCCESS         0
>>>   #define MPAM_FB_ERR_NOT_SUPPORTED    -1
>>> @@ -209,3 +214,39 @@ int mpam_fb_send_write_request(struct mpam_msc *msc, u16 reg, u32 value)
>>>       return mpam_fb_send_request(msc, msc->mpam_fb_msc_id, reg, &value,
>>>                       MPAM_MSC_WRITE_CMD);
>>>   }
>>> +
>>> +/* We only support MPAM-Fb protocol version 1.x */
>>> +int mpam_fb_check_protocol_version(struct mpam_msc *msc)
>>> +{
>>> +    u32 version;
>>> +    int ret;
>>> +
>>> +    ret = mpam_fb_send_request(msc, 0, 0, &version,
>>> +                   MPAM_PROTOCOL_VERSION_CMD);
>>> +    if (ret)
>>> +        return ret;
>>> +
>>> +    if (FIELD_GET(MPAM_FB_VERSION_MAJOR_MASK, version) != 1) {
>>> +        pr_err("Incompatible MPAM-Fb protocol version %d.%d\n",
>>> +               FIELD_GET(MPAM_FB_VERSION_MAJOR_MASK, version),
>>> +               FIELD_GET(MPAM_FB_VERSION_MINOR_MASK, version));
>>> +
>>> +        return -EINVAL;
>>> +    }
>>> +
>>> +    return 0;
>>> +}
>>> +
>>> +int mpam_fb_check_shared_buffer_size(struct mpam_msc *msc)
>>> +{
>>> +    int min_buffer_size = MPAM_FB_MAX_MSG_SIZE +
>>> +                  sizeof(struct acpi_pcct_ext_pcc_shared_memory);
>>> +
>>> +    if (msc->pcc_chan->pcc_chan->shmem_size < min_buffer_size) {
>>> +        pr_err("MPAM-Fb PCC channel size too small.\n");
>>> +
>>> +        return -ENOMEM;
>>> +    }
>>> +
>>> +    return 0;
>>> +}
>>> diff --git a/drivers/resctrl/mpam_internal.h b/drivers/resctrl/mpam_internal.h
>>> index e5a2804facd2d..314d3b48ca9fd 100644
>>> --- a/drivers/resctrl/mpam_internal.h
>>> +++ b/drivers/resctrl/mpam_internal.h
>>> @@ -538,6 +538,8 @@ static inline void mpam_resctrl_teardown_class(struct mpam_class *class) { }
>>>   /* MPAM-Fb Firmware-backed protocol wrappers */
>>>   int mpam_fb_send_read_request(struct mpam_msc *msc, u16 reg, u32 *result);
>>>   int mpam_fb_send_write_request(struct mpam_msc *msc, u16 reg, u32 value);
>>> +int mpam_fb_check_protocol_version(struct mpam_msc *msc);
>>> +int mpam_fb_check_shared_buffer_size(struct mpam_msc *msc);
>>>     /*
>>>    * MPAM MSCs have the following register layout. See:
>>
>