Re: [PATCH] powerpc/npu-dma.c: Fix crash after __mmu_notifier_register failure

[Alistair Popple]

Thanks Mark, this will also fix the lack of cleanup OPAL call in the unlikely
case the kzalloc() fails.

Acked-By: Alistair Popple 

On Friday, 9 February 2018 7:20:06 PM AEDT Mark Hairgrove wrote:
> pnv_npu2_init_context wasn't checking the return code from
> __mmu_notifier_register. If  __mmu_notifier_register failed, the
> npu_context was still assigned to the mm and the caller wasn't given any
> indication that things went wrong. Later on pnv_npu2_destroy_context would
> be called, which in turn called mmu_notifier_unregister and dropped
> mm->mm_count without having incremented it in the first place. This led to
> various forms of corruption like mm use-after-free and mm double-free.
> 
> __mmu_notifier_register can fail with EINTR if a signal is pending, so
> this case can be frequent.
> 
> This patch calls opal_npu_destroy_context on the failure paths, and makes
> sure not to assign mm->context.npu_context until past the failure points.
> 
> Signed-off-by: Mark Hairgrove 
> ---
>  arch/powerpc/platforms/powernv/npu-dma.c |   32 +++--
>  1 files changed, 21 insertions(+), 11 deletions(-)
> 
> diff --git a/arch/powerpc/platforms/powernv/npu-dma.c 
> b/arch/powerpc/platforms/powernv/npu-dma.c
> index f6cbc1a..48c73aa 100644
> --- a/arch/powerpc/platforms/powernv/npu-dma.c
> +++ b/arch/powerpc/platforms/powernv/npu-dma.c
> @@ -677,6 +677,11 @@ struct npu_context *pnv_npu2_init_context(struct pci_dev 
> *gpdev,
>   /* No nvlink associated with this GPU device */
>   return ERR_PTR(-ENODEV);
>  
> + nvlink_dn = of_parse_phandle(npdev->dev.of_node, "ibm,nvlink", 0);
> + if (WARN_ON(of_property_read_u32(nvlink_dn, "ibm,npu-link-index",
> + _index)))
> + return ERR_PTR(-ENODEV);
> +
>   if (!mm || mm->context.id == 0) {
>   /*
>* Kernel thread contexts are not supported and context id 0 is
> @@ -704,25 +709,30 @@ struct npu_context *pnv_npu2_init_context(struct 
> pci_dev *gpdev,
>*/
>   npu_context = mm->context.npu_context;
>   if (!npu_context) {
> + rc = -ENOMEM;
>   npu_context = kzalloc(sizeof(struct npu_context), GFP_KERNEL);
> - if (!npu_context)
> - return ERR_PTR(-ENOMEM);
> + if (npu_context) {
> + kref_init(_context->kref);
> + npu_context->mm = mm;
> + npu_context->mn.ops = _nmmu_notifier_ops;
> + rc = __mmu_notifier_register(_context->mn, mm);
> + }
> +
> + if (rc) {
> + kfree(npu_context);
> + opal_npu_destroy_context(nphb->opal_id, mm->context.id,
> + PCI_DEVID(gpdev->bus->number,
> + gpdev->devfn));
> + return ERR_PTR(rc);
> + }
>  
>   mm->context.npu_context = npu_context;
> - npu_context->mm = mm;
> - npu_context->mn.ops = _nmmu_notifier_ops;
> - __mmu_notifier_register(_context->mn, mm);
> - kref_init(_context->kref);
>   } else {
> - kref_get(_context->kref);
> + WARN_ON(!kref_get_unless_zero(_context->kref));
>   }
>  
>   npu_context->release_cb = cb;
>   npu_context->priv = priv;
> - nvlink_dn = of_parse_phandle(npdev->dev.of_node, "ibm,nvlink", 0);
> - if (WARN_ON(of_property_read_u32(nvlink_dn, "ibm,npu-link-index",
> - _index)))
> - return ERR_PTR(-ENODEV);
>   npu_context->npdev[npu->index][nvlink_index] = npdev;
>  
>   if (!nphb->npu.nmmu_flush) {
> 

[PATCH] powerpc/npu-dma.c: Fix crash after __mmu_notifier_register failure

[Mark Hairgrove]

pnv_npu2_init_context wasn't checking the return code from
__mmu_notifier_register. If  __mmu_notifier_register failed, the
npu_context was still assigned to the mm and the caller wasn't given any
indication that things went wrong. Later on pnv_npu2_destroy_context would
be called, which in turn called mmu_notifier_unregister and dropped
mm->mm_count without having incremented it in the first place. This led to
various forms of corruption like mm use-after-free and mm double-free.

__mmu_notifier_register can fail with EINTR if a signal is pending, so
this case can be frequent.

This patch calls opal_npu_destroy_context on the failure paths, and makes
sure not to assign mm->context.npu_context until past the failure points.

Signed-off-by: Mark Hairgrove 
---
 arch/powerpc/platforms/powernv/npu-dma.c |   32 +++--
 1 files changed, 21 insertions(+), 11 deletions(-)

diff --git a/arch/powerpc/platforms/powernv/npu-dma.c 
b/arch/powerpc/platforms/powernv/npu-dma.c
index f6cbc1a..48c73aa 100644
--- a/arch/powerpc/platforms/powernv/npu-dma.c
+++ b/arch/powerpc/platforms/powernv/npu-dma.c
@@ -677,6 +677,11 @@ struct npu_context *pnv_npu2_init_context(struct pci_dev 
*gpdev,
/* No nvlink associated with this GPU device */
return ERR_PTR(-ENODEV);
 
+   nvlink_dn = of_parse_phandle(npdev->dev.of_node, "ibm,nvlink", 0);
+   if (WARN_ON(of_property_read_u32(nvlink_dn, "ibm,npu-link-index",
+   _index)))
+   return ERR_PTR(-ENODEV);
+
if (!mm || mm->context.id == 0) {
/*
 * Kernel thread contexts are not supported and context id 0 is
@@ -704,25 +709,30 @@ struct npu_context *pnv_npu2_init_context(struct pci_dev 
*gpdev,
 */
npu_context = mm->context.npu_context;
if (!npu_context) {
+   rc = -ENOMEM;
npu_context = kzalloc(sizeof(struct npu_context), GFP_KERNEL);
-   if (!npu_context)
-   return ERR_PTR(-ENOMEM);
+   if (npu_context) {
+   kref_init(_context->kref);
+   npu_context->mm = mm;
+   npu_context->mn.ops = _nmmu_notifier_ops;
+   rc = __mmu_notifier_register(_context->mn, mm);
+   }
+
+   if (rc) {
+   kfree(npu_context);
+   opal_npu_destroy_context(nphb->opal_id, mm->context.id,
+   PCI_DEVID(gpdev->bus->number,
+   gpdev->devfn));
+   return ERR_PTR(rc);
+   }
 
mm->context.npu_context = npu_context;
-   npu_context->mm = mm;
-   npu_context->mn.ops = _nmmu_notifier_ops;
-   __mmu_notifier_register(_context->mn, mm);
-   kref_init(_context->kref);
} else {
-   kref_get(_context->kref);
+   WARN_ON(!kref_get_unless_zero(_context->kref));
}
 
npu_context->release_cb = cb;
npu_context->priv = priv;
-   nvlink_dn = of_parse_phandle(npdev->dev.of_node, "ibm,nvlink", 0);
-   if (WARN_ON(of_property_read_u32(nvlink_dn, "ibm,npu-link-index",
-   _index)))
-   return ERR_PTR(-ENODEV);
npu_context->npdev[npu->index][nvlink_index] = npdev;
 
if (!nphb->npu.nmmu_flush) {
-- 
1.7.2.5