Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <f066ce84e415dc2726a23fe8a8886fa157b34bcd.camel@linux.ibm.com>
Subject: Re: [PATCH v5 1/6] iommu/s390: Fix duplicate domain attachments
From:   Niklas Schnelle <schnelle@linux.ibm.com>
To:     Matthew Rosato <mjrosato@linux.ibm.com>,
        Pierre Morel <pmorel@linux.ibm.com>, iommu@lists.linux.dev
Cc:     linux-s390@vger.kernel.org, borntraeger@linux.ibm.com,
        hca@linux.ibm.com, gor@linux.ibm.com,
        gerald.schaefer@linux.ibm.com, agordeev@linux.ibm.com,
        svens@linux.ibm.com, joro@8bytes.org, will@kernel.org,
        robin.murphy@arm.com, jgg@nvidia.com, linux-kernel@vger.kernel.org
Date:   Fri, 07 Oct 2022 13:20:35 +0200
In-Reply-To: <ce0d9e782ae523c35173f053cbb3924e6880420c.camel@linux.ibm.com>
References: <20221006144700.3380098-1-schnelle@linux.ibm.com>
         <20221006144700.3380098-2-schnelle@linux.ibm.com>
         <c778fc2b-2c0d-777c-52fb-1b315cf95552@linux.ibm.com>
         <ce0d9e782ae523c35173f053cbb3924e6880420c.camel@linux.ibm.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri, 2022-10-07 at 08:55 +0200, Niklas Schnelle wrote:
> On Thu, 2022-10-06 at 17:02 -0400, Matthew Rosato wrote:
> > On 10/6/22 10:46 AM, Niklas Schnelle wrote:
> > > Since commit fa7e9ecc5e1c ("iommu/s390: Tolerate repeat attach_dev
> > > calls") we can end up with duplicates in the list of devices attached to
> > > a domain. This is inefficient and confusing since only one domain can
> > > actually be in control of the IOMMU translations for a device. Fix this
> > > by detaching the device from the previous domain, if any, on attach.
> > > Add a WARN_ON() in case we still have attached devices on freeing the
> > > domain. While here remove the re-attach on failure dance as it was
> > > determined to be unlikely to help and may confuse debug and recovery.
> > > 
> > > Fixes: fa7e9ecc5e1c ("iommu/s390: Tolerate repeat attach_dev calls")
> > > Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
> > > ---
> > > v4->v5:
> > > - Unregister IOAT and set zdev->dma_table on error (Matt)
> > > 
> > ...
> > 
> > >  static int s390_iommu_attach_device(struct iommu_domain *domain,
> > >  				    struct device *dev)
> > >  {
> > > @@ -90,7 +116,7 @@ static int s390_iommu_attach_device(struct iommu_domain *domain,
> > >  	struct zpci_dev *zdev = to_zpci_dev(dev);
> > >  	struct s390_domain_device *domain_device;
> > >  	unsigned long flags;
> > > -	int cc, rc;
> > > +	int cc, rc = 0;
> > >  
> > >  	if (!zdev)
> > >  		return -ENODEV;
> > > @@ -99,23 +125,17 @@ static int s390_iommu_attach_device(struct iommu_domain *domain,
> > >  	if (!domain_device)
> > >  		return -ENOMEM;
> > >  
> > > -	if (zdev->dma_table && !zdev->s390_domain) {
> > > -		cc = zpci_dma_exit_device(zdev);
> > > -		if (cc) {
> > > -			rc = -EIO;
> > > -			goto out_free;
> > > -		}
> > > -	}
> > > -
> > >  	if (zdev->s390_domain)
> > > -		zpci_unregister_ioat(zdev, 0);
> > > +		__s390_iommu_detach_device(zdev);
> > > +	else if (zdev->dma_table)
> > > +		zpci_dma_exit_device(zdev);
> > >  
> > >  	zdev->dma_table = s390_domain->dma_table;
> > >  	cc = zpci_register_ioat(zdev, 0, zdev->start_dma, zdev->end_dma,
> > >  				virt_to_phys(zdev->dma_table));
> > >  	if (cc) {
> > >  		rc = -EIO;
> > > -		goto out_restore;
> > > +		goto out_free;
> > >  	}
> > 
> > Hmm, with this we will leave attach_dev with a zdev->dma_table associated with this domain (not one generated via zpci_dma_init_device) and zdev->s390_domain == 0.  Won't this cause both s390_domain_free and zpci_dma_exit_device() to try and free the same dma table?
> > 
> > I think we also have to leave with a NULL zdev->dma_table in this case too (you technically could skip the zpci_unregister_ioat)
> 
> Argh you're right. This is I think a a bad rebase, in v4 I had the
> zpci_register_ioat() use s390_domain->dma_table and only set zdev-
> > dma_table after that succeeded. I seem to have lost that part
> somewhere along the way. With that we zdev->dma_table would be NULL and
> all would be good.
> 

Went back to the way I did it in v4 for v6. I think I was simply an
idiot and when comparing to the state prior to the commit forgot why I
did it this way and thought it was an unneeded change..

> > >  
> > >  	spin_lock_irqsave(&s390_domain->list_lock, flags);
> > > @@ -127,9 +147,9 @@ static int s390_iommu_attach_device(struct iommu_domain *domain,
> > >  	/* Allow only devices with identical DMA range limits */
> > >  	} else if (domain->geometry.aperture_start != zdev->start_dma ||
> > >  		   domain->geometry.aperture_end != zdev->end_dma) {
> > > -		rc = -EINVAL;
> > >  		spin_unlock_irqrestore(&s390_domain->list_lock, flags);
> > > -		goto out_restore;
> > > +		rc = -EINVAL;
> > > +		goto out_unregister;
> > >  	}
> > >  	domain_device->zdev = zdev;
> > >  	zdev->s390_domain = s390_domain;
> > > @@ -138,14 +158,9 @@ static int s390_iommu_attach_device(struct iommu_domain *domain,
> > >  
> > >  	return 0;
> > >  
> > > -out_restore:
> > > -	if (!zdev->s390_domain) {
> > > -		zpci_dma_init_device(zdev);
> > > -	} else {
> > > -		zdev->dma_table = zdev->s390_domain->dma_table;
> > > -		zpci_register_ioat(zdev, 0, zdev->start_dma, zdev->end_dma,
> > > -				   virt_to_phys(zdev->dma_table));
> > > -	}
> > > +out_unregister:
> > > +	zpci_unregister_ioat(zdev, 0);
> > > +	zdev->dma_table = NULL;
> > >  out_free:
> > >  	kfree(domain_device);
> > >  
> 
>