Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org, Zheyu Ma <zheyuma97@gmail.com>,
        Douglas Miller <doug.miller@cornelisnetworks.com>,
        Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.10 316/452] RDMA/hfi1: Prevent use of lock before it is initialized
Date:   Tue,  7 Jun 2022 19:02:53 +0200
Message-Id: <20220607164917.974785343@linuxfoundation.org>
In-Reply-To: <20220607164908.521895282@linuxfoundation.org>
References: <20220607164908.521895282@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Douglas Miller <doug.miller@cornelisnetworks.com>

[ Upstream commit 05c03dfd09c069c4ffd783b47b2da5dcc9421f2c ]

If there is a failure during probe of hfi1 before the sdma_map_lock is
initialized, the call to hfi1_free_devdata() will attempt to use a lock
that has not been initialized. If the locking correctness validator is on
then an INFO message and stack trace resembling the following may be seen:

  INFO: trying to register non-static key.
  The code is fine but needs lockdep annotation, or maybe
  you didn't initialize this object before use?
  turning off the locking correctness validator.
  Call Trace:
  register_lock_class+0x11b/0x880
  __lock_acquire+0xf3/0x7930
  lock_acquire+0xff/0x2d0
  _raw_spin_lock_irq+0x46/0x60
  sdma_clean+0x42a/0x660 [hfi1]
  hfi1_free_devdata+0x3a7/0x420 [hfi1]
  init_one+0x867/0x11a0 [hfi1]
  pci_device_probe+0x40e/0x8d0

The use of sdma_map_lock in sdma_clean() is for freeing the sdma_map
memory, and sdma_map is not allocated/initialized until after
sdma_map_lock has been initialized. This code only needs to be run if
sdma_map is not NULL, and so checking for that condition will avoid trying
to use the lock before it is initialized.

Fixes: 473291b3ea0e ("IB/hfi1: Fix for early release of sdma context")
Fixes: 7724105686e7 ("IB/hfi1: add driver files")
Link: https://lore.kernel.org/r/20220520183701.48973.72434.stgit@awfm-01.cornelisnetworks.com
Reported-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: Douglas Miller <doug.miller@cornelisnetworks.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/infiniband/hw/hfi1/sdma.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/sdma.c b/drivers/infiniband/hw/hfi1/sdma.c
index 0b73dc7847aa..a044bee257f9 100644
--- a/drivers/infiniband/hw/hfi1/sdma.c
+++ b/drivers/infiniband/hw/hfi1/sdma.c
@@ -1330,11 +1330,13 @@ void sdma_clean(struct hfi1_devdata *dd, size_t num_engines)
 		kvfree(sde->tx_ring);
 		sde->tx_ring = NULL;
 	}
-	spin_lock_irq(&dd->sde_map_lock);
-	sdma_map_free(rcu_access_pointer(dd->sdma_map));
-	RCU_INIT_POINTER(dd->sdma_map, NULL);
-	spin_unlock_irq(&dd->sde_map_lock);
-	synchronize_rcu();
+	if (rcu_access_pointer(dd->sdma_map)) {
+		spin_lock_irq(&dd->sde_map_lock);
+		sdma_map_free(rcu_access_pointer(dd->sdma_map));
+		RCU_INIT_POINTER(dd->sdma_map, NULL);
+		spin_unlock_irq(&dd->sde_map_lock);
+		synchronize_rcu();
+	}
 	kfree(dd->per_sdma);
 	dd->per_sdma = NULL;
 
-- 
2.35.1