[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20170830152331.7cf8c30d@w520.home>
Date: Wed, 30 Aug 2017 15:23:31 -0600
From: Alex Williamson <alex.williamson@...hat.com>
To: Eric Auger <eric.auger@...hat.com>
Cc: eric.auger.pro@...il.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] vfio: fix noiommu vfio_iommu_group_get reference
count
On Fri, 11 Aug 2017 15:16:06 +0200
Eric Auger <eric.auger@...hat.com> wrote:
> In vfio_iommu_group_get() we want to increase the reference
> count of the iommu group.
>
> In noiommu case, the group does not exist and is allocated.
> iommu_group_add_device() increases the group ref count. However we
> then call iommu_group_put() which decrements it.
>
> This leads to a "refcount_t: underflow WARN_ON".
>
> Only decrement the ref count in case of iommu_group_add_device
> failure.
>
> Signed-off-by: Eric Auger <eric.auger@...hat.com>
>
> ---
>
> v1 -> v2:
> put the group in case iommu_group_add_device failed
> ---
> drivers/vfio/vfio.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
Applied to next branch for v4.14. Thanks,
Alex
> diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
> index 330d505..4ee4f36 100644
> --- a/drivers/vfio/vfio.c
> +++ b/drivers/vfio/vfio.c
> @@ -138,9 +138,10 @@ struct iommu_group *vfio_iommu_group_get(struct device *dev)
> iommu_group_set_name(group, "vfio-noiommu");
> iommu_group_set_iommudata(group, &noiommu, NULL);
> ret = iommu_group_add_device(group, dev);
> - iommu_group_put(group);
> - if (ret)
> + if (ret) {
> + iommu_group_put(group);
> return NULL;
> + }
>
> /*
> * Where to taint? At this point we've added an IOMMU group for a
Powered by blists - more mailing lists