Date:   Sun, 28 Nov 2021 19:37:45 +0000
Message-ID: <871r30rrzq.wl-maz@kernel.org>
From:   Marc Zyngier <maz@kernel.org>
To:     Thomas Gleixner <tglx@linutronix.de>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        Bjorn Helgaas <helgaas@kernel.org>,
        Alex Williamson <alex.williamson@redhat.com>,
        Kevin Tian <kevin.tian@intel.com>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Megha Dey <megha.dey@intel.com>,
        Ashok Raj <ashok.raj@intel.com>, linux-pci@vger.kernel.org,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Santosh Shilimkar <ssantosh@kernel.org>,
        iommu@lists.linux-foundation.org, dmaengine@vger.kernel.org,
        Stuart Yoder <stuyoder@gmail.com>,
        Laurentiu Tudor <laurentiu.tudor@nxp.com>,
        Nishanth Menon <nm@ti.com>, Tero Kristo <kristo@kernel.org>,
        linux-arm-kernel@lists.infradead.org, x86@kernel.org,
        Vinod Koul <vkoul@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Will Deacon <will@kernel.org>, Sinan Kaya <okaya@kernel.org>
Subject: Re: [patch 29/37] PCI/MSI: Use __msi_get_virq() in pci_get_vector()
In-Reply-To: <20211126230525.660206325@linutronix.de>
References: <20211126224100.303046749@linutronix.de>
        <20211126230525.660206325@linutronix.de>
User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI-EPG/1.14.7 (Harue)
 FLIM-LB/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL-LB/10.8 EasyPG/1.0.0 Emacs/27.1
 (x86_64-pc-linux-gnu) MULE/6.0 (HANACHIRUSATO)
MIME-Version: 1.0 (generated by SEMI-EPG 1.14.7 - "Harue")
Content-Type: text/plain; charset=US-ASCII
Precedence: bulk

On Sat, 27 Nov 2021 01:22:03 +0000,
Thomas Gleixner <tglx@linutronix.de> wrote:
> 
> Use __msi_get_vector() and handle the return values to be compatible.
> 
> No functional change intended.

You wish ;-).

[   15.779540] pcieport 0001:00:01.0: AER: request AER IRQ -22 failed

Notice how amusing the IRQ number is?

> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> ---
>  drivers/pci/msi/msi.c |   25 +++++--------------------
>  1 file changed, 5 insertions(+), 20 deletions(-)
> 
> --- a/drivers/pci/msi/msi.c
> +++ b/drivers/pci/msi/msi.c
> @@ -1023,28 +1023,13 @@ EXPORT_SYMBOL(pci_free_irq_vectors);
>   */
>  int pci_irq_vector(struct pci_dev *dev, unsigned int nr)
>  {
> -	if (dev->msix_enabled) {
> -		struct msi_desc *entry;
> +	int irq = __msi_get_virq(&dev->dev, nr);
>  
> -		for_each_pci_msi_entry(entry, dev) {
> -			if (entry->msi_index == nr)
> -				return entry->irq;
> -		}
> -		WARN_ON_ONCE(1);
> -		return -EINVAL;
> +	switch (irq) {
> +	case -ENODEV: return !nr ? dev->irq : -EINVAL;
> +	case -ENOENT: return -EINVAL;
>  	}
> -
> -	if (dev->msi_enabled) {
> -		struct msi_desc *entry = first_pci_msi_entry(dev);
> -
> -		if (WARN_ON_ONCE(nr >= entry->nvec_used))
> -			return -EINVAL;
> -	} else {
> -		if (WARN_ON_ONCE(nr > 0))
> -			return -EINVAL;
> -	}
> -
> -	return dev->irq + nr;
> +	return irq;
>  }
>  EXPORT_SYMBOL(pci_irq_vector);

I worked around it with the hack below, but I doubt this is the real
thing. portdrv_core.c does complicated things, and I don't completely
understand its logic.

	M.

diff --git a/drivers/pci/msi/msi.c b/drivers/pci/msi/msi.c
index 1f72bc734226..b15278a5fb4b 100644
--- a/drivers/pci/msi/msi.c
+++ b/drivers/pci/msi/msi.c
@@ -1092,8 +1092,9 @@ int pci_irq_vector(struct pci_dev *dev, unsigned int nr)
 	int irq = __msi_get_virq(&dev->dev, nr);
 
 	switch (irq) {
-	case -ENODEV: return !nr ? dev->irq : -EINVAL;
-	case -ENOENT: return -EINVAL;
+	case -ENOENT:
+	case -ENODEV:
+		return !nr ? dev->irq : -EINVAL;
 	}
 	return irq;
 }

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