[PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[trix]

From: Tom Rix <trix@redhat.com>

In ocrdma_dealloc_ucontext_pd() uctx->cntxt_pd is assigned to
the variable pd and then after uctx->cntxt_pd is freed, the
variable pd is passed to function _ocrdma_dealloc_pd() which
dereferences pd directly or through its call to
ocrdma_mbx_dealloc_pd().

Reorder the free using the variable pd.

Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
Signed-off-by: Tom Rix <trix@redhat.com>
---
 drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
index bc98bd950d99..3acb5c10b155 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c
@@ -434,9 +434,9 @@ static void ocrdma_dealloc_ucontext_pd(struct ocrdma_ucontext *uctx)
 		pr_err("%s(%d) Freeing in use pdid=0x%x.\n",
 		       __func__, dev->id, pd->id);
 	}
-	kfree(uctx->cntxt_pd);
 	uctx->cntxt_pd = NULL;
 	_ocrdma_dealloc_pd(dev, pd);
+	kfree(pd);
 }
 
 static struct ocrdma_pd *ocrdma_get_ucontext_pd(struct ocrdma_ucontext *uctx)
-- 
2.27.0

Re: [PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[Leon Romanovsky]

On Tue, Dec 29, 2020 at 06:46:53PM -0800, trix@redhat.com wrote:
> From: Tom Rix <trix@redhat.com>
>
> In ocrdma_dealloc_ucontext_pd() uctx->cntxt_pd is assigned to
> the variable pd and then after uctx->cntxt_pd is freed, the
> variable pd is passed to function _ocrdma_dealloc_pd() which
> dereferences pd directly or through its call to
> ocrdma_mbx_dealloc_pd().
>
> Reorder the free using the variable pd.
>
> Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
> Signed-off-by: Tom Rix <trix@redhat.com>
> ---
>  drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>

Thanks,
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>

Re: [PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[Jason Gunthorpe]

On Tue, Dec 29, 2020 at 06:46:53PM -0800, trix@redhat.com wrote:
> From: Tom Rix <trix@redhat.com>
> 
> In ocrdma_dealloc_ucontext_pd() uctx->cntxt_pd is assigned to
> the variable pd and then after uctx->cntxt_pd is freed, the
> variable pd is passed to function _ocrdma_dealloc_pd() which
> dereferences pd directly or through its call to
> ocrdma_mbx_dealloc_pd().
> 
> Reorder the free using the variable pd.
> 
> Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
> Signed-off-by: Tom Rix <trix@redhat.com>
>  drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Applied to for-rc

Is anyone testing ocrdma? Just doing the pyverbs rdma tests with kasn
turned on would have instantly caught this, and the change is nearly a
year old.

Is ocrdma obsolete enough we can delete the driver?

Thanks,
Jason

Re: [PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[Tom Rix]

On 1/7/21 12:41 PM, Jason Gunthorpe wrote:
> On Tue, Dec 29, 2020 at 06:46:53PM -0800, trix@redhat.com wrote:
>> From: Tom Rix <trix@redhat.com>
>>
>> In ocrdma_dealloc_ucontext_pd() uctx->cntxt_pd is assigned to
>> the variable pd and then after uctx->cntxt_pd is freed, the
>> variable pd is passed to function _ocrdma_dealloc_pd() which
>> dereferences pd directly or through its call to
>> ocrdma_mbx_dealloc_pd().
>>
>> Reorder the free using the variable pd.
>>
>> Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
>> Signed-off-by: Tom Rix <trix@redhat.com>
>>  drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
> Applied to for-rc
>
> Is anyone testing ocrdma? Just doing the pyverbs rdma tests with kasn
> turned on would have instantly caught this, and the change is nearly a
> year old.
>
> Is ocrdma obsolete enough we can delete the driver?

I am not an authority on ocrdma, i am fixing treewide, the problems clang static analysis flags.

Tom

>
> Thanks,
> Jason
>

Re: [PATCH] RDMA/ocrdma: fix use after free in ocrdma_dealloc_ucontext_pd()

[Selvin Xavier]

[-- Attachment #1: Type: text/plain, Size: 1357 bytes --]

On Fri, Jan 8, 2021 at 3:13 AM Tom Rix <trix@redhat.com> wrote:
>
>
> On 1/7/21 12:41 PM, Jason Gunthorpe wrote:
> > On Tue, Dec 29, 2020 at 06:46:53PM -0800, trix@redhat.com wrote:
> >> From: Tom Rix <trix@redhat.com>
> >>
> >> In ocrdma_dealloc_ucontext_pd() uctx->cntxt_pd is assigned to
> >> the variable pd and then after uctx->cntxt_pd is freed, the
> >> variable pd is passed to function _ocrdma_dealloc_pd() which
> >> dereferences pd directly or through its call to
> >> ocrdma_mbx_dealloc_pd().
> >>
> >> Reorder the free using the variable pd.
> >>
> >> Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
> >> Signed-off-by: Tom Rix <trix@redhat.com>
> >>  drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> > Applied to for-rc
> >
> > Is anyone testing ocrdma? Just doing the pyverbs rdma tests with kasn
> > turned on would have instantly caught this, and the change is nearly a
> > year old.
> >
> > Is ocrdma obsolete enough we can delete the driver?

Broadcom is not doing any active development/testing  with ocrdma now.
I am checking with other teams to see if this can be deleted
completely. Will get back asap.

Thanks,
Selvin

>
> I am not an authority on ocrdma, i am fixing treewide, the problems clang static analysis flags.
>
> Tom
>
> >
> > Thanks,
> > Jason
> >
>

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4181 bytes --]