From: Basavaraj Natikar <bnatikar@amd.com>
To: Ho Jie Feng <hjf3108@gmail.com>, basavaraj.natikar@amd.com
Cc: stable@vger.kernel.org, Jiri Kosina <jikos@kernel.org>,
Benjamin Tissoires <bentiss@kernel.org>,
linux-input@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] HID: amd_sfh: Use shared interrupts
Date: Fri, 23 Oct 2026 13:56:16 +0530 [thread overview]
Message-ID: <7e2046b4-be10-4d7d-a7fb-ce56add7ea60@amd.com> (raw)
In-Reply-To: <20261023081803.21761-1-hjf3108@gmail.com>
On 10/23/2026 1:48 PM, Ho Jie Feng wrote:
> [You don't often get email from hjf3108@gmail.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>
> The AMD SFH driver requests interrupts without IRQF_SHARED.
> This can cause VFIO PCI passthrough initialization to fail
> when a passed-through device shares the interrupt with AMD MP2.
>
> Request the interrupt with IRQF_SHARED. Check for a pending MP2
> interrupt before clearing and acknowledging it, and return IRQ_NONE
> when no MP2 interrupt is pending so the interrupt can be shared with
> other devices.
>
> On an AMD Ryzen 7 7735HS host running 7.0.0-38-generic kernel
> from Ubuntu, trying to pass through an xHCI controller
> (1022:161d, 0000:e6:00.3) fails with this message:
>
> genirq: Flags mismatch irq 45.
> 00200080 (vfio-intx(0000:e6:00.3)) vs. 00200000 (pcie_mp2_amd)
>
> QEMU reports:
>
> vfio 0000:e6:00.3: Failed to set up TRIGGER eventfd signaling for
> interrupt INTX-0: VFIO_DEVICE_SET_IRQS failure: Device or resource busy
>
> With this change, the passthrough is successful through VFIO
> on the same host.
>
> Fixes: 7f016b35ca76 ("HID: amd_sfh: Add interrupt handler to process interrupts")
> Cc: stable@vger.kernel.org
> Signed-off-by: Ho Jie Feng <hjf3108@gmail.com>
> ---
> drivers/hid/amd-sfh-hid/amd_sfh_common.h | 6 +++---
> drivers/hid/amd-sfh-hid/amd_sfh_pcie.c | 18 ++++++++++++------
> 2 files changed, 15 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/hid/amd-sfh-hid/amd_sfh_common.h b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> index e8cc7f1d289cc..4481a50359de6 100644
> --- a/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> +++ b/drivers/hid/amd-sfh-hid/amd_sfh_common.h
> @@ -85,7 +85,7 @@ struct amd_mp2_ops {
> void (*stop)(struct amd_mp2_dev *privdata, u16 sensor_idx);
> void (*stop_all)(struct amd_mp2_dev *privdata);
> int (*response)(struct amd_mp2_dev *mp2, u8 sid, u32 sensor_sts);
> - void (*clear_intr)(struct amd_mp2_dev *privdata);
> + bool (*clear_intr)(struct amd_mp2_dev *privdata);
> int (*init_intr)(struct amd_mp2_dev *privdata);
> int (*discovery_status)(struct amd_mp2_dev *privdata);
> void (*suspend)(struct amd_mp2_dev *mp2);
> @@ -100,9 +100,9 @@ struct amd_mp2_ops {
>
> void amd_sfh_work(struct work_struct *work);
> void amd_sfh_work_buffer(struct work_struct *work);
> -void amd_sfh_clear_intr_v2(struct amd_mp2_dev *privdata);
> +bool amd_sfh_clear_intr_v2(struct amd_mp2_dev *privdata);
> int amd_sfh_irq_init_v2(struct amd_mp2_dev *privdata);
> -void amd_sfh_clear_intr(struct amd_mp2_dev *privdata);
> +bool amd_sfh_clear_intr(struct amd_mp2_dev *privdata);
> int amd_sfh_irq_init(struct amd_mp2_dev *privdata);
>
> static inline u64 amd_get_c2p_val(struct amd_mp2_dev *mp2, u32 idx)
> diff --git a/drivers/hid/amd-sfh-hid/amd_sfh_pcie.c b/drivers/hid/amd-sfh-hid/amd_sfh_pcie.c
> index b3cb13b8c5e87..f48a096a6d2bc 100644
> --- a/drivers/hid/amd-sfh-hid/amd_sfh_pcie.c
> +++ b/drivers/hid/amd-sfh-hid/amd_sfh_pcie.c
> @@ -103,25 +103,31 @@ static void amd_stop_all_sensor_v2(struct amd_mp2_dev *privdata)
> writel(cmd_base.ul, privdata->mmio + AMD_C2P_MSG0);
> }
>
> -void amd_sfh_clear_intr_v2(struct amd_mp2_dev *privdata)
> +bool amd_sfh_clear_intr_v2(struct amd_mp2_dev *privdata)
> {
> if (readl(privdata->mmio + amd_get_p2c_val(privdata, 4))) {
> writel(0, privdata->mmio + amd_get_p2c_val(privdata, 4));
> writel(0xf, privdata->mmio + amd_get_p2c_val(privdata, 5));
> + return true;
> }
> +
> + return false;
> }
>
> -void amd_sfh_clear_intr(struct amd_mp2_dev *privdata)
> +bool amd_sfh_clear_intr(struct amd_mp2_dev *privdata)
> {
> if (privdata->mp2_ops->clear_intr)
> - privdata->mp2_ops->clear_intr(privdata);
> + return privdata->mp2_ops->clear_intr(privdata);
> +
> + return false;
> }
>
> static irqreturn_t amd_sfh_irq_handler(int irq, void *data)
> {
> - amd_sfh_clear_intr(data);
> + if (amd_sfh_clear_intr(data))
> + return IRQ_HANDLED;
>
> - return IRQ_HANDLED;
> + return IRQ_NONE;
> }
>
> int amd_sfh_irq_init_v2(struct amd_mp2_dev *privdata)
> @@ -129,7 +135,7 @@ int amd_sfh_irq_init_v2(struct amd_mp2_dev *privdata)
> pcim_intx(privdata->pdev, true);
>
> return devm_request_irq(&privdata->pdev->dev, privdata->pdev->irq,
> - amd_sfh_irq_handler, 0, DRIVER_NAME, privdata);
> + amd_sfh_irq_handler, IRQF_SHARED, DRIVER_NAME, privdata);
> }
>
> static int amd_sfh_dis_sts_v2(struct amd_mp2_dev *privdata)
Acked-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Thanks,
--
Basavaraj
>
> base-commit: 2c2fe95ab19b09bef7391e3e5cdf8146d09792b6
> --
> 2.55.0
>
prev parent reply other threads:[~2026-10-23 8:26 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-23 8:18 Ho Jie Feng
2026-10-23 8:26 ` Basavaraj Natikar [this message]
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=7e2046b4-be10-4d7d-a7fb-ce56add7ea60@amd.com \
--to=bnatikar@amd.com \
--cc=basavaraj.natikar@amd.com \
--cc=bentiss@kernel.org \
--cc=hjf3108@gmail.com \
--cc=jikos@kernel.org \
--cc=linux-input@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=stable@vger.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
all inboxes | Powered by JetHome®