linux-mips.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Marc Zyngier <maz@kernel.org>
To: Johan Hovold <johan+linaro@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>,
	x86@kernel.org, platform-driver-x86@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org, linux-mips@vger.kernel.org,
	linux-kernel@vger.kernel.org, stable@vger.kernel.org,
	Dmitry Torokhov <dtor@chromium.org>,
	Jon Hunter <jonathanh@nvidia.com>,
	Hsin-Yi Wang <hsinyi@chromium.org>,
	Mark-PK Tsai <mark-pk.tsai@mediatek.com>
Subject: Re: [PATCH v5 06/19] irqdomain: Fix mapping-creation race
Date: Thu, 09 Feb 2023 14:03:19 +0000	[thread overview]
Message-ID: <86edqzylzs.wl-maz@kernel.org> (raw)
In-Reply-To: <20230209132323.4599-7-johan+linaro@kernel.org>

On Thu, 09 Feb 2023 13:23:10 +0000,
Johan Hovold <johan+linaro@kernel.org> wrote:
> 
> Parallel probing of devices that share interrupts (e.g. when a driver
> uses asynchronous probing) can currently result in two mappings for the
> same hardware interrupt to be created due to missing serialisation.
> 
> Make sure to hold the irq_domain_mutex when creating mappings so that
> looking for an existing mapping before creating a new one is done
> atomically.
> 
> Fixes: 765230b5f084 ("driver-core: add asynchronous probing support for drivers")
> Fixes: b62b2cf5759b ("irqdomain: Fix handling of type settings for existing mappings")
> Link: https://lore.kernel.org/r/YuJXMHoT4ijUxnRb@hovoldconsulting.com
> Cc: stable@vger.kernel.org      # 4.8
> Cc: Dmitry Torokhov <dtor@chromium.org>
> Cc: Jon Hunter <jonathanh@nvidia.com>
> Tested-by: Hsin-Yi Wang <hsinyi@chromium.org>
> Tested-by: Mark-PK Tsai <mark-pk.tsai@mediatek.com>
> Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
> ---
>  kernel/irq/irqdomain.c | 55 ++++++++++++++++++++++++++++++------------
>  1 file changed, 40 insertions(+), 15 deletions(-)
> 
> diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
> index 7b57949bc79c..1ddb01bd49a4 100644
> --- a/kernel/irq/irqdomain.c
> +++ b/kernel/irq/irqdomain.c
> @@ -25,6 +25,9 @@ static DEFINE_MUTEX(irq_domain_mutex);
>  
>  static struct irq_domain *irq_default_domain;
>  
> +static int irq_domain_alloc_irqs_locked(struct irq_domain *domain, int irq_base,
> +					unsigned int nr_irqs, int node, void *arg,
> +					bool realloc, const struct irq_affinity_desc *affinity);
>  static void irq_domain_check_hierarchy(struct irq_domain *domain);
>  
>  struct irqchip_fwid {
> @@ -682,9 +685,9 @@ unsigned int irq_create_direct_mapping(struct irq_domain *domain)
>  EXPORT_SYMBOL_GPL(irq_create_direct_mapping);
>  #endif
>  
> -static unsigned int __irq_create_mapping_affinity(struct irq_domain *domain,
> -						  irq_hw_number_t hwirq,
> -						  const struct irq_affinity_desc *affinity)
> +static unsigned int irq_create_mapping_affinity_locked(struct irq_domain *domain,
> +						       irq_hw_number_t hwirq,
> +						       const struct irq_affinity_desc *affinity)
>  {
>  	struct device_node *of_node = irq_domain_get_of_node(domain);
>  	int virq;
> @@ -699,7 +702,7 @@ static unsigned int __irq_create_mapping_affinity(struct irq_domain *domain,
>  		return 0;
>  	}
>  
> -	if (irq_domain_associate(domain, virq, hwirq)) {
> +	if (irq_domain_associate_locked(domain, virq, hwirq)) {
>  		irq_free_desc(virq);
>  		return 0;
>  	}
> @@ -735,14 +738,20 @@ unsigned int irq_create_mapping_affinity(struct irq_domain *domain,
>  		return 0;
>  	}
>  
> +	mutex_lock(&irq_domain_mutex);
> +
>  	/* Check if mapping already exists */
>  	virq = irq_find_mapping(domain, hwirq);
>  	if (virq) {
>  		pr_debug("existing mapping on virq %d\n", virq);
> -		return virq;
> +		goto out;
>  	}
>  
> -	return __irq_create_mapping_affinity(domain, hwirq, affinity);
> +	virq = irq_create_mapping_affinity_locked(domain, hwirq, affinity);
> +out:
> +	mutex_unlock(&irq_domain_mutex);
> +
> +	return virq;
>  }
>  EXPORT_SYMBOL_GPL(irq_create_mapping_affinity);
>  
> @@ -809,6 +818,8 @@ unsigned int irq_create_fwspec_mapping(struct irq_fwspec *fwspec)
>  	if (WARN_ON(type & ~IRQ_TYPE_SENSE_MASK))
>  		type &= IRQ_TYPE_SENSE_MASK;
>  
> +	mutex_lock(&irq_domain_mutex);
> +
>  	/*
>  	 * If we've already configured this interrupt,
>  	 * don't do it again, or hell will break loose.
> @@ -821,7 +832,7 @@ unsigned int irq_create_fwspec_mapping(struct irq_fwspec *fwspec)
>  		 * interrupt number.
>  		 */
>  		if (type == IRQ_TYPE_NONE || type == irq_get_trigger_type(virq))
> -			return virq;
> +			goto out;
>  
>  		/*
>  		 * If the trigger type has not been set yet, then set
> @@ -830,36 +841,43 @@ unsigned int irq_create_fwspec_mapping(struct irq_fwspec *fwspec)
>  		if (irq_get_trigger_type(virq) == IRQ_TYPE_NONE) {
>  			irq_data = irq_get_irq_data(virq);
>  			if (!irq_data)
> -				return 0;
> +				goto err;
>  
>  			irqd_set_trigger_type(irq_data, type);
> -			return virq;
> +			goto out;
>  		}
>  
>  		pr_warn("type mismatch, failed to map hwirq-%lu for %s!\n",
>  			hwirq, of_node_full_name(to_of_node(fwspec->fwnode)));
> -		return 0;
> +		goto err;
>  	}
>  
>  	if (irq_domain_is_hierarchy(domain)) {
> -		virq = irq_domain_alloc_irqs(domain, 1, NUMA_NO_NODE, fwspec);
> +		virq = irq_domain_alloc_irqs_locked(domain, -1, 1, NUMA_NO_NODE,
> +						    fwspec, false, NULL);
>  		if (virq <= 0)
> -			return 0;
> +			goto err;
>  	} else {
>  		/* Create mapping */
> -		virq = __irq_create_mapping_affinity(domain, hwirq, NULL);
> +		virq = irq_create_mapping_affinity_locked(domain, hwirq, NULL);
>  		if (!virq)
> -			return virq;
> +			goto err;
>  	}
>  
>  	irq_data = irq_get_irq_data(virq);
>  	if (WARN_ON(!irq_data))
> -		return 0;
> +		goto err;
>  
>  	/* Store trigger type */
>  	irqd_set_trigger_type(irq_data, type);
> +out:
> +	mutex_unlock(&irq_domain_mutex);
>  
>  	return virq;
> +err:
> +	mutex_unlock(&irq_domain_mutex);
> +
> +	return 0;

nit: it'd look better if we had a single exit path with the unlock,
setting virq to 0 on failure. Not a big deal, as this can be tidied up
when applied.

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.

  reply	other threads:[~2023-02-09 14:03 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-09 13:23 [PATCH v5 00/19] irqdomain: fix mapping race and clean up locking Johan Hovold
2023-02-09 13:23 ` [PATCH v5 01/19] irqdomain: Fix association race Johan Hovold
2023-02-09 13:23 ` [PATCH v5 02/19] irqdomain: Fix disassociation race Johan Hovold
2023-02-09 13:23 ` [PATCH v5 03/19] irqdomain: Drop bogus fwspec-mapping error handling Johan Hovold
2023-02-09 13:23 ` [PATCH v5 04/19] irqdomain: Look for existing mapping only once Johan Hovold
2023-02-09 13:23 ` [PATCH v5 05/19] irqdomain: Refactor __irq_domain_alloc_irqs() Johan Hovold
2023-02-09 13:23 ` [PATCH v5 06/19] irqdomain: Fix mapping-creation race Johan Hovold
2023-02-09 14:03   ` Marc Zyngier [this message]
2023-02-10  9:10     ` Johan Hovold
2023-02-09 13:23 ` [PATCH v5 07/19] irqdomain: Drop revmap mutex Johan Hovold
2023-02-09 13:23 ` [PATCH v5 08/19] irqdomain: Drop dead domain-name assignment Johan Hovold
2023-02-09 13:23 ` [PATCH v5 09/19] irqdomain: Drop leftover brackets Johan Hovold
2023-02-09 13:23 ` [PATCH v5 10/19] irqdomain: Clean up irq_domain_push/pop_irq() Johan Hovold
2023-02-09 13:23 ` [PATCH v5 11/19] x86/ioapic: Use irq_domain_create_hierarchy() Johan Hovold
2023-02-09 13:23 ` [PATCH v5 12/19] x86/apic: " Johan Hovold
2023-02-09 13:23 ` [PATCH v5 13/19] irqchip/alpine-msi: Use irq_domain_add_hierarchy() Johan Hovold
2023-02-09 13:23 ` [PATCH v5 14/19] irqchip/gic-v2m: Use irq_domain_create_hierarchy() Johan Hovold
2023-02-09 13:23 ` [PATCH v5 15/19] irqchip/gic-v3-its: " Johan Hovold
2023-02-09 13:23 ` [PATCH v5 16/19] irqchip/gic-v3-mbi: " Johan Hovold
2023-02-09 13:23 ` [PATCH v5 17/19] irqchip/loongson-pch-msi: " Johan Hovold
2023-02-09 13:23 ` [PATCH v5 18/19] irqchip/mvebu-odmi: " Johan Hovold
2023-02-09 13:23 ` [PATCH v5 19/19] irqdomain: Switch to per-domain locking Johan Hovold
2023-02-09 16:00   ` Marc Zyngier
2023-02-10  9:56     ` Johan Hovold
2023-02-10 11:38       ` Marc Zyngier
2023-02-10 12:57         ` Johan Hovold
2023-02-10 15:06           ` Marc Zyngier
2023-02-11 11:35             ` Johan Hovold
2023-02-11 12:52               ` Marc Zyngier

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=86edqzylzs.wl-maz@kernel.org \
    --to=maz@kernel.org \
    --cc=dtor@chromium.org \
    --cc=hsinyi@chromium.org \
    --cc=johan+linaro@kernel.org \
    --cc=jonathanh@nvidia.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mips@vger.kernel.org \
    --cc=mark-pk.tsai@mediatek.com \
    --cc=platform-driver-x86@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=tglx@linutronix.de \
    --cc=x86@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;
as well as URLs for NNTP newsgroup(s).