public inbox for linux-riscv@lists.infradead.org
 help / color / mirror / Atom feed
From: Robin Murphy <robin.murphy@arm.com>
To: Zong Li <zong.li@sifive.com>,
	joro@8bytes.org, will@kernel.org, tjeznach@rivosinc.com,
	paul.walmsley@sifive.com, palmer@dabbelt.com,
	aou@eecs.berkeley.edu, luxu.kernel@bytedance.com,
	linux-kernel@vger.kernel.org, iommu@lists.linux.dev,
	linux-riscv@lists.infradead.org
Subject: Re: [PATCH 2/2] iommu/riscv: support HPM and interrupt handling
Date: Wed, 15 Jan 2025 21:56:21 +0000	[thread overview]
Message-ID: <6301d20b-d51d-4f96-94ea-134e065d398e@arm.com> (raw)
In-Reply-To: <20250115030306.29735-3-zong.li@sifive.com>

On 2025-01-15 3:03 am, Zong Li wrote:
> Initialize the PMU and uninitialize it when driver is removed.
> Interrupt handling is also implemented, and the handler needs
> to be a primary handler instead of a threaded function because
> pt_regs is empty when threading the IRQ. However, pt_regs is
> required by perf_event_overflow.
> 
> Signed-off-by: Zong Li <zong.li@sifive.com>
> Tested-by: Xu Lu <luxu.kernel@bytedance.com>
> ---
>   drivers/iommu/riscv/iommu.c | 65 +++++++++++++++++++++++++++++++++++++
>   1 file changed, 65 insertions(+)
> 
> diff --git a/drivers/iommu/riscv/iommu.c b/drivers/iommu/riscv/iommu.c
> index 8a05def774bd..20ae90471484 100644
> --- a/drivers/iommu/riscv/iommu.c
> +++ b/drivers/iommu/riscv/iommu.c
> @@ -552,6 +552,62 @@ static irqreturn_t riscv_iommu_fltq_process(int irq, void *data)
>   	return IRQ_HANDLED;
>   }
>   
> +/*
> + * IOMMU Hardware performance monitor
> + */
> +
> +/* HPM interrupt primary handler */
> +static irqreturn_t riscv_iommu_hpm_irq_handler(int irq, void *dev_id)
> +{
> +	struct riscv_iommu_device *iommu = (struct riscv_iommu_device *)dev_id;
> +
> +	/* Clear performance monitoring interrupt pending */
> +	riscv_iommu_writel(iommu, RISCV_IOMMU_REG_IPSR, RISCV_IOMMU_IPSR_PMIP);
> +
> +	/* Process pmu irq */
> +	riscv_iommu_pmu_handle_irq(&iommu->pmu);
> +
> +	return IRQ_HANDLED;
> +}
> +
> +/* HPM initialization */
> +static int riscv_iommu_hpm_enable(struct riscv_iommu_device *iommu)
> +{
> +	int rc;
> +
> +	if (!(iommu->caps & RISCV_IOMMU_CAPABILITIES_HPM))
> +		return 0;
> +
> +	/*
> +	 * pt_regs is empty when threading the IRQ, but pt_regs is necessary
> +	 * by perf_event_overflow. Use primary handler instead of thread
> +	 * function for PM IRQ.
> +	 *
> +	 * Set the IRQF_ONESHOT flag because this IRQ might be shared with
> +	 * other threaded IRQs by other queues.
> +	 */
> +	rc = devm_request_irq(iommu->dev,
> +			      iommu->irqs[riscv_iommu_queue_vec(iommu, RISCV_IOMMU_IPSR_PMIP)],
> +			      riscv_iommu_hpm_irq_handler, IRQF_ONESHOT | IRQF_SHARED, NULL, iommu);

Hmm, shared interrupts are tricky for PMUs, since perf requires any IRQ 
handler touching a PMU is running on pmu->cpu, so you have to be very 
careful about maintaining affinity and not letting anyone else change it 
behind your back.

The other thing is that if it really is shared, at this point you could 
now be in riscv_iommu_pmu_handle_irq() dereferencing NULL.

> +	if (rc)
> +		return rc;
> +
> +	return riscv_iommu_pmu_init(&iommu->pmu, iommu->reg, dev_name(iommu->dev));
> +}
> +
> +/* HPM uninitialization */
> +static void riscv_iommu_hpm_disable(struct riscv_iommu_device *iommu)
> +{
> +	if (!(iommu->caps & RISCV_IOMMU_CAPABILITIES_HPM))
> +		return;
> +
> +	devm_free_irq(iommu->dev,
> +		      iommu->irqs[riscv_iommu_queue_vec(iommu, RISCV_IOMMU_IPSR_PMIP)],
> +		      iommu);
> +
> +	riscv_iommu_pmu_uninit(&iommu->pmu);
> +}
> +
>   /* Lookup and initialize device context info structure. */
>   static struct riscv_iommu_dc *riscv_iommu_get_dc(struct riscv_iommu_device *iommu,
>   						 unsigned int devid)
> @@ -1596,6 +1652,9 @@ void riscv_iommu_remove(struct riscv_iommu_device *iommu)
>   	riscv_iommu_iodir_set_mode(iommu, RISCV_IOMMU_DDTP_IOMMU_MODE_OFF);
>   	riscv_iommu_queue_disable(&iommu->cmdq);
>   	riscv_iommu_queue_disable(&iommu->fltq);
> +
> +	if (iommu->caps & RISCV_IOMMU_CAPABILITIES_HPM)
> +		riscv_iommu_pmu_uninit(&iommu->pmu);
>   }
>   
>   int riscv_iommu_init(struct riscv_iommu_device *iommu)
> @@ -1635,6 +1694,10 @@ int riscv_iommu_init(struct riscv_iommu_device *iommu)
>   	if (rc)
>   		goto err_queue_disable;
>   
> +	rc = riscv_iommu_hpm_enable(iommu);
> +	if (rc)
> +		goto err_hpm_disable;
> +

I would leave this until after the whole IOMMU setup has succeeded. The 
PMU is not critical to IOMMU operation, so at that point an error is not 
fatal, it just means you don't get a PMU, thus there shouldn't need to 
be any cleanup outside riscv_iommu_hpm_enable() itself.

Thanks,
Robin.

>   	rc = iommu_device_sysfs_add(&iommu->iommu, NULL, NULL, "riscv-iommu@%s",
>   				    dev_name(iommu->dev));
>   	if (rc) {
> @@ -1653,6 +1716,8 @@ int riscv_iommu_init(struct riscv_iommu_device *iommu)
>   err_remove_sysfs:
>   	iommu_device_sysfs_remove(&iommu->iommu);
>   err_iodir_off:
> +	riscv_iommu_hpm_disable(iommu);
> +err_hpm_disable:
>   	riscv_iommu_iodir_set_mode(iommu, RISCV_IOMMU_DDTP_IOMMU_MODE_OFF);
>   err_queue_disable:
>   	riscv_iommu_queue_disable(&iommu->fltq);


_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

  parent reply	other threads:[~2025-01-15 21:56 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-15  3:03 [PATCH 0/2] RISC-V IOMMU HPM support Zong Li
2025-01-15  3:03 ` [PATCH 1/2] iommu/riscv: add RISC-V IOMMU PMU support Zong Li
2025-01-15  3:45   ` [External] " Xu Lu
2025-01-15  7:48     ` Zong Li
2025-01-15  8:25       ` Xu Lu
2025-01-15 21:32   ` Robin Murphy
2025-01-23  6:56     ` Zong Li
2025-01-23  8:35   ` kernel test robot
2025-01-23 12:16   ` kernel test robot
2025-01-15  3:03 ` [PATCH 2/2] iommu/riscv: support HPM and interrupt handling Zong Li
2025-01-15  9:44   ` Andrew Jones
2025-01-17  2:33     ` Zong Li
2025-01-15 21:56   ` Robin Murphy [this message]
2025-01-17  2:46     ` Zong Li
2025-01-17 12:45       ` Robin Murphy
2025-01-21  8:37         ` Zong Li

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=6301d20b-d51d-4f96-94ea-134e065d398e@arm.com \
    --to=robin.murphy@arm.com \
    --cc=aou@eecs.berkeley.edu \
    --cc=iommu@lists.linux.dev \
    --cc=joro@8bytes.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=luxu.kernel@bytedance.com \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.com \
    --cc=tjeznach@rivosinc.com \
    --cc=will@kernel.org \
    --cc=zong.li@sifive.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