From: Jason Gunthorpe Date: Tue, 16 Sep 2025 15:42:56 +0000 (-0300) Subject: iommufd: Fix refcounting race during mmap X-Git-Tag: v6.17~28^2~3 X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=7a425ec75d2bb30a1c959a8676ef8c5ef285095d;p=thirdparty%2Fkernel%2Fstable.git iommufd: Fix refcounting race during mmap The owner object of the imap can be destroyed while the imap remains in the mtree. So access to the imap pointer without holding locks is racy with destruction. The imap is safe to access outside the lock once a users refcount is obtained, the owner object cannot start destruction until users is 0. Thus the users refcount should not be obtained at the end of iommufd_fops_mmap() but instead inside the mtree lock held around the mtree_load(). Move the refcount there and use refcount_inc_not_zero() as we can have a 0 refcount inside the mtree during destruction races. Link: https://patch.msgid.link/r/0-v1-e6faace50971+3cc-iommufd_mmap_fix_jgg@nvidia.com Cc: stable@vger.kernel.org Fixes: 56e9a0d8e53f ("iommufd: Add mmap interface") Reviewed-by: Kevin Tian Reviewed-by: Nicolin Chen Signed-off-by: Jason Gunthorpe --- diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c index 15af7ced0501d..a9d4decc8ba16 100644 --- a/drivers/iommu/iommufd/main.c +++ b/drivers/iommu/iommufd/main.c @@ -550,16 +550,23 @@ static int iommufd_fops_mmap(struct file *filp, struct vm_area_struct *vma) if (vma->vm_flags & VM_EXEC) return -EPERM; + mtree_lock(&ictx->mt_mmap); /* vma->vm_pgoff carries a page-shifted start position to an immap */ immap = mtree_load(&ictx->mt_mmap, vma->vm_pgoff << PAGE_SHIFT); - if (!immap) + if (!immap || !refcount_inc_not_zero(&immap->owner->users)) { + mtree_unlock(&ictx->mt_mmap); return -ENXIO; + } + mtree_unlock(&ictx->mt_mmap); + /* * mtree_load() returns the immap for any contained mmio_addr, so only * allow the exact immap thing to be mapped */ - if (vma->vm_pgoff != immap->vm_pgoff || length != immap->length) - return -ENXIO; + if (vma->vm_pgoff != immap->vm_pgoff || length != immap->length) { + rc = -ENXIO; + goto err_refcount; + } vma->vm_pgoff = 0; vma->vm_private_data = immap; @@ -570,10 +577,11 @@ static int iommufd_fops_mmap(struct file *filp, struct vm_area_struct *vma) immap->mmio_addr >> PAGE_SHIFT, length, vma->vm_page_prot); if (rc) - return rc; + goto err_refcount; + return 0; - /* vm_ops.open won't be called for mmap itself. */ - refcount_inc(&immap->owner->users); +err_refcount: + refcount_dec(&immap->owner->users); return rc; }