The Linux Kernel Mailing List
 help / color / mirror / Atom feed
From: "Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>
To: Lizhi Hou <lizhi.hou@amd.com>
Cc: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>,
	ogabbay@kernel.org,  quic_jhugo@quicinc.com,
	maciej.falkowski@linux.intel.com,
	 Hans de Goede <hansg@kernel.org>,
	dri-devel@lists.freedesktop.org,
	 LKML <linux-kernel@vger.kernel.org>,
	max.zhen@amd.com,  sonal.santan@amd.com,
	mario.limonciello@amd.com,  platform-driver-x86@vger.kernel.org,
	VinitKumar.Shukla@amd.com,
	 Patil Rajesh Reddy <Patil.Reddy@amd.com>
Subject: Re: [PATCH V3 1/2] platform/x86/amd/pmf: Introduce new interface to export NPU metrics
Date: Tue, 6 Jan 2026 19:09:49 +0200 (EET)	[thread overview]
Message-ID: <b6c80015-d399-e8c1-4b8a-d2ebb4cee9da@linux.intel.com> (raw)
In-Reply-To: <4363d65e-c50b-eb1a-39d3-99a9809f3c37@amd.com>

[-- Attachment #1: Type: text/plain, Size: 8965 bytes --]

On Tue, 6 Jan 2026, Lizhi Hou wrote:

> What should be the next step? Should I generate both patches based on
> review-ilpo-next? Will the amdxdna change upstream through platform-x86 tree?

(We seemed to have crossed emails.)

That would work for me too, but I thought it was wanted that patch 2 to 
go through accel?

-- 
 i.

> On 1/6/26 08:16, Lizhi Hou wrote:
> > 
> > On 1/6/26 05:19, Shyam Sundar S K wrote:
> > > 
> > > On 1/5/2026 22:59, Lizhi Hou wrote:
> > > > From: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> > > > 
> > > > The PMF driver retrieves NPU metrics data from the PMFW. Introduce a new
> > > > interface to make NPU metrics accessible to other drivers like AMDXDNA
> > > > driver, which can access and utilize this information as needed.
> > > > 
> > > > Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
> > > > Co-developed-by: Patil Rajesh Reddy <Patil.Reddy@amd.com>
> > > > Signed-off-by: Patil Rajesh Reddy <Patil.Reddy@amd.com>
> > > > Signed-off-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> > > > [lizhi: save return value of is_npu_metrics_supported() and return it]
> > > > Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
> > > Lizhi, just noticied that this patch may not apply cleanly on
> > > Ilpo-next (which tree does this change land? drm-misc or
> > > platform-x86)? If platform-x86, I can share you a rebased version
> > 
> > I used drm-misc.
> > 
> > Lizhi
> > 
> > > offline so that you can wrap into XDNA series.
> > > 
> > > Thanks,
> > > Shyam
> > > 
> > > > ---
> > > >   drivers/platform/x86/amd/pmf/core.c | 76 +++++++++++++++++++++++++++++
> > > >   drivers/platform/x86/amd/pmf/pmf.h  |  2 +
> > > >   include/linux/amd-pmf-io.h          | 21 ++++++++
> > > >   3 files changed, 99 insertions(+)
> > > > 
> > > > diff --git a/drivers/platform/x86/amd/pmf/core.c
> > > > b/drivers/platform/x86/amd/pmf/core.c
> > > > index 8fc293c9c538..d180a39e03bf 100644
> > > > --- a/drivers/platform/x86/amd/pmf/core.c
> > > > +++ b/drivers/platform/x86/amd/pmf/core.c
> > > > @@ -8,12 +8,15 @@
> > > >    * Author: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
> > > >    */
> > > >   +#include <linux/array_size.h>
> > > > +#include <linux/cleanup.h>
> > > >   #include <linux/debugfs.h>
> > > >   #include <linux/iopoll.h>
> > > >   #include <linux/module.h>
> > > >   #include <linux/pci.h>
> > > >   #include <linux/platform_device.h>
> > > >   #include <linux/power_supply.h>
> > > > +#include <linux/string.h>
> > > >   #include <asm/amd/node.h>
> > > >   #include "pmf.h"
> > > >   @@ -53,6 +56,8 @@ static bool force_load;
> > > >   module_param(force_load, bool, 0444);
> > > >   MODULE_PARM_DESC(force_load, "Force load this driver on supported
> > > > older platforms (experimental)");
> > > >   +static struct device *pmf_device;
> > > > +
> > > >   static int amd_pmf_pwr_src_notify_call(struct notifier_block *nb,
> > > > unsigned long event, void *data)
> > > >   {
> > > >       struct amd_pmf_dev *pmf = container_of(nb, struct amd_pmf_dev,
> > > > pwr_src_notifier);
> > > > @@ -314,6 +319,71 @@ int amd_pmf_init_metrics_table(struct amd_pmf_dev
> > > > *dev)
> > > >       return 0;
> > > >   }
> > > >   +static int is_npu_metrics_supported(struct amd_pmf_dev *pdev)
> > > > +{
> > > > +    switch (pdev->cpu_id) {
> > > > +    case PCI_DEVICE_ID_AMD_1AH_M20H_ROOT:
> > > > +    case PCI_DEVICE_ID_AMD_1AH_M60H_ROOT:
> > > > +        return 0;
> > > > +    default:
> > > > +        return -EOPNOTSUPP;
> > > > +    }
> > > > +}
> > > > +
> > > > +static int amd_pmf_get_smu_metrics(struct amd_pmf_dev *dev, struct
> > > > amd_pmf_npu_metrics *data)
> > > > +{
> > > > +    int ret, i;
> > > > +
> > > > +    guard(mutex)(&dev->metrics_mutex);
> > > > +
> > > > +    ret = is_npu_metrics_supported(dev);
> > > > +    if (ret)
> > > > +        return ret;
> > > > +
> > > > +    ret = amd_pmf_set_dram_addr(dev, true);
> > > > +    if (ret)
> > > > +        return ret;
> > > > +
> > > > +    memset(dev->buf, 0, dev->mtable_size);
> > > > +
> > > > +    /* Send SMU command to get NPU metrics */
> > > > +    ret = amd_pmf_send_cmd(dev, SET_TRANSFER_TABLE, SET_CMD,
> > > > METRICS_TABLE_ID, NULL);
> > > > +    if (ret) {
> > > > +        dev_err(dev->dev, "SMU command failed to get NPU metrics:
> > > > %d\n", ret);
> > > > +        return ret;
> > > > +    }
> > > > +
> > > > +    memcpy(&dev->m_table_v2, dev->buf, dev->mtable_size);
> > > > +
> > > > +    data->npuclk_freq = dev->m_table_v2.npuclk_freq;
> > > > +    for (i = 0; i < ARRAY_SIZE(data->npu_busy); i++)
> > > > +        data->npu_busy[i] = dev->m_table_v2.npu_busy[i];
> > > > +    data->npu_power = dev->m_table_v2.npu_power;
> > > > +    data->mpnpuclk_freq = dev->m_table_v2.mpnpuclk_freq;
> > > > +    data->npu_reads = dev->m_table_v2.npu_reads;
> > > > +    data->npu_writes = dev->m_table_v2.npu_writes;
> > > > +
> > > > +    return 0;
> > > > +}
> > > > +
> > > > +int amd_pmf_get_npu_data(struct amd_pmf_npu_metrics *info)
> > > > +{
> > > > +    struct amd_pmf_dev *pdev;
> > > > +
> > > > +    if (!info)
> > > > +        return -EINVAL;
> > > > +
> > > > +    if (!pmf_device)
> > > > +        return -ENODEV;
> > > > +
> > > > +    pdev = dev_get_drvdata(pmf_device);
> > > > +    if (!pdev)
> > > > +        return -ENODEV;
> > > > +
> > > > +    return amd_pmf_get_smu_metrics(pdev, info);
> > > > +}
> > > > +EXPORT_SYMBOL_GPL(amd_pmf_get_npu_data);
> > > > +
> > > >   static int amd_pmf_suspend_handler(struct device *dev)
> > > >   {
> > > >       struct amd_pmf_dev *pdev = dev_get_drvdata(dev);
> > > > @@ -477,6 +547,10 @@ static int amd_pmf_probe(struct platform_device
> > > > *pdev)
> > > >       if (err)
> > > >           return err;
> > > >   +    err = devm_mutex_init(dev->dev, &dev->metrics_mutex);
> > > > +    if (err)
> > > > +        return err;
> > > > +
> > > >       apmf_acpi_init(dev);
> > > >       platform_set_drvdata(pdev, dev);
> > > >       amd_pmf_dbgfs_register(dev);
> > > > @@ -485,6 +559,8 @@ static int amd_pmf_probe(struct platform_device
> > > > *pdev)
> > > >       if (is_apmf_func_supported(dev, APMF_FUNC_SBIOS_HEARTBEAT_V2))
> > > >           amd_pmf_notify_sbios_heartbeat_event_v2(dev, ON_LOAD);
> > > >   +    pmf_device = dev->dev;
> > > > +
> > > >       dev_info(dev->dev, "registered PMF device successfully\n");
> > > >         return 0;
> > > > diff --git a/drivers/platform/x86/amd/pmf/pmf.h
> > > > b/drivers/platform/x86/amd/pmf/pmf.h
> > > > index 9144c8c3bbaf..1bf896c5b92c 100644
> > > > --- a/drivers/platform/x86/amd/pmf/pmf.h
> > > > +++ b/drivers/platform/x86/amd/pmf/pmf.h
> > > > @@ -12,6 +12,7 @@
> > > >   #define PMF_H
> > > >     #include <linux/acpi.h>
> > > > +#include <linux/amd-pmf-io.h>
> > > >   #include <linux/input.h>
> > > >   #include <linux/platform_device.h>
> > > >   #include <linux/platform_profile.h>
> > > > @@ -413,6 +414,7 @@ struct amd_pmf_dev {
> > > >       struct apmf_sbios_req_v1 req1;
> > > >       struct pmf_bios_inputs_prev cb_prev; /* To preserve custom BIOS
> > > > inputs */
> > > >       bool cb_flag;                 /* To handle first custom BIOS input
> > > > */
> > > > +    struct mutex metrics_mutex;
> > > >   };
> > > >     struct apmf_sps_prop_granular_v2 {
> > > > diff --git a/include/linux/amd-pmf-io.h b/include/linux/amd-pmf-io.h
> > > > index 6fa510f419c0..55198d2875cc 100644
> > > > --- a/include/linux/amd-pmf-io.h
> > > > +++ b/include/linux/amd-pmf-io.h
> > > > @@ -61,5 +61,26 @@ enum laptop_placement {
> > > >       LP_UNDEFINED,
> > > >   };
> > > >   +/**
> > > > + * struct amd_pmf_npu_metrics: Get NPU metrics data from PMF driver
> > > > + * @npuclk_freq: NPU clock frequency [MHz]
> > > > + * @npu_busy: NPU busy % [0-100]
> > > > + * @npu_power: NPU power [mW]
> > > > + * @mpnpuclk_freq: MPNPU [MHz]
> > > > + * @npu_reads: NPU read bandwidth [MB/sec]
> > > > + * @npu_writes: NPU write bandwidth [MB/sec]
> > > > + */
> > > > +struct amd_pmf_npu_metrics {
> > > > +    u16 npuclk_freq;
> > > > +    u16 npu_busy[8];
> > > > +    u16 npu_power;
> > > > +    u16 mpnpuclk_freq;
> > > > +    u16 npu_reads;
> > > > +    u16 npu_writes;
> > > > +};
> > > > +
> > > >   int amd_get_sfh_info(struct amd_sfh_info *sfh_info, enum
> > > > sfh_message_type op);
> > > > +
> > > > +/* AMD PMF and NPU interface */
> > > > +int amd_pmf_get_npu_data(struct amd_pmf_npu_metrics *info);
> > > >   #endif
> 

  reply	other threads:[~2026-01-06 17:09 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-05 17:29 [PATCH V3 0/2] Get real time power input via AMD PMF Lizhi Hou
2026-01-05 17:29 ` [PATCH V3 1/2] platform/x86/amd/pmf: Introduce new interface to export NPU metrics Lizhi Hou
2026-01-06 13:19   ` Shyam Sundar S K
2026-01-06 16:16     ` Lizhi Hou
2026-01-06 17:05       ` Ilpo Järvinen
2026-01-06 17:07       ` Lizhi Hou
2026-01-06 17:09         ` Ilpo Järvinen [this message]
2026-01-06 17:35           ` Lizhi Hou
2026-01-15 15:18   ` Ilpo Järvinen
2026-01-15 16:49     ` Lizhi Hou
2026-01-05 17:29 ` [PATCH V3 2/2] accel/amdxdna: Add IOCTL to retrieve realtime NPU power estimate Lizhi Hou

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=b6c80015-d399-e8c1-4b8a-d2ebb4cee9da@linux.intel.com \
    --to=ilpo.jarvinen@linux.intel.com \
    --cc=Patil.Reddy@amd.com \
    --cc=Shyam-sundar.S-k@amd.com \
    --cc=VinitKumar.Shukla@amd.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=hansg@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lizhi.hou@amd.com \
    --cc=maciej.falkowski@linux.intel.com \
    --cc=mario.limonciello@amd.com \
    --cc=max.zhen@amd.com \
    --cc=ogabbay@kernel.org \
    --cc=platform-driver-x86@vger.kernel.org \
    --cc=quic_jhugo@quicinc.com \
    --cc=sonal.santan@amd.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