[PATCH 6.6.y] net: ena: Add validation for completion descriptors consistency

[PATCH 6.6.y] net: ena: Add validation for completion descriptors consistency

[Artem Dinaburg]

From: David Arinzon <darinzon@amazon.com>

[ Upstream commit b37b98a3a0c1198bafe8c2d9ce0bc845b4e7a9a7 ]

Validate that `first` flag is set only for the first
descriptor in multi-buffer packets.
In case of an invalid descriptor, a reset will occur.
A new reset reason for RX data corruption has been added.

Signed-off-by: Shahar Itzko <itzko@amazon.com>
Signed-off-by: David Arinzon <darinzon@amazon.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20240512134637.25299-4-darinzon@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Assisted-by: LLM
Signed-off-by: Artem Dinaburg <artem@trailofbits.com>
---
Hi Greg, Sasha, and net maintainers,

I am working through the small CVE backports still missing from 6.6.y.
This one addresses CVE-2024-40999. It detects inconsistent ENA completion
descriptors and resets instead of consuming them.

The fix is already present in 6.12.y, 6.18.y, and 7.2.y, but not in 6.6.y.
This fix also affects 6.1.y, which will need a separate backport; this
submission contains only the 6.6.y patch.
The code change is identical to upstream.

Could you please queue it for 6.6.y?

CVE: CVE-2024-40999
Upstream: b37b98a3a0c1198bafe8c2d9ce0bc845b4e7a9a7

AI assistance: An LLM helped identify, adapt, and validate this backport; I
reviewed the resulting code and validation evidence.

Thanks,
Artem Dinaburg

 drivers/net/ethernet/amazon/ena/ena_eth_com.c | 37 ++++++++++++++-----
 drivers/net/ethernet/amazon/ena/ena_netdev.c  |  2 +
 .../net/ethernet/amazon/ena/ena_regs_defs.h   |  1 +
 3 files changed, 30 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/amazon/ena/ena_eth_com.c b/drivers/net/ethernet/amazon/ena/ena_eth_com.c
index 933e619b3a3134..4c6e07aa4bbb55 100644
--- a/drivers/net/ethernet/amazon/ena/ena_eth_com.c
+++ b/drivers/net/ethernet/amazon/ena/ena_eth_com.c
@@ -229,30 +229,43 @@ static struct ena_eth_io_rx_cdesc_base *
 		idx * io_cq->cdesc_entry_size_in_bytes);
 }
 
-static u16 ena_com_cdesc_rx_pkt_get(struct ena_com_io_cq *io_cq,
-					   u16 *first_cdesc_idx)
+static int ena_com_cdesc_rx_pkt_get(struct ena_com_io_cq *io_cq,
+				    u16 *first_cdesc_idx,
+				    u16 *num_descs)
 {
+	u16 count = io_cq->cur_rx_pkt_cdesc_count, head_masked;
 	struct ena_eth_io_rx_cdesc_base *cdesc;
-	u16 count = 0, head_masked;
 	u32 last = 0;
 
 	do {
+		u32 status;
+
 		cdesc = ena_com_get_next_rx_cdesc(io_cq);
 		if (!cdesc)
 			break;
+		status = READ_ONCE(cdesc->status);
 
 		ena_com_cq_inc_head(io_cq);
+		if (unlikely((status & ENA_ETH_IO_RX_CDESC_BASE_FIRST_MASK) >>
+		    ENA_ETH_IO_RX_CDESC_BASE_FIRST_SHIFT && count != 0)) {
+			struct ena_com_dev *dev = ena_com_io_cq_to_ena_dev(io_cq);
+
+			netdev_err(dev->net_device,
+				   "First bit is on in descriptor #%d on q_id: %d, req_id: %u\n",
+				   count, io_cq->qid, cdesc->req_id);
+			return -EFAULT;
+		}
 		count++;
-		last = (READ_ONCE(cdesc->status) & ENA_ETH_IO_RX_CDESC_BASE_LAST_MASK) >>
-		       ENA_ETH_IO_RX_CDESC_BASE_LAST_SHIFT;
+		last = (status & ENA_ETH_IO_RX_CDESC_BASE_LAST_MASK) >>
+			ENA_ETH_IO_RX_CDESC_BASE_LAST_SHIFT;
 	} while (!last);
 
 	if (last) {
 		*first_cdesc_idx = io_cq->cur_rx_pkt_cdesc_start_idx;
-		count += io_cq->cur_rx_pkt_cdesc_count;
 
 		head_masked = io_cq->head & (io_cq->q_depth - 1);
 
+		*num_descs = count;
 		io_cq->cur_rx_pkt_cdesc_count = 0;
 		io_cq->cur_rx_pkt_cdesc_start_idx = head_masked;
 
@@ -260,11 +273,11 @@ static u16 ena_com_cdesc_rx_pkt_get(struct ena_com_io_cq *io_cq,
 			   "ENA q_id: %d packets were completed. first desc idx %u descs# %d\n",
 			   io_cq->qid, *first_cdesc_idx, count);
 	} else {
-		io_cq->cur_rx_pkt_cdesc_count += count;
-		count = 0;
+		io_cq->cur_rx_pkt_cdesc_count = count;
+		*num_descs = 0;
 	}
 
-	return count;
+	return 0;
 }
 
 static int ena_com_create_meta(struct ena_com_io_sq *io_sq,
@@ -539,10 +552,14 @@ int ena_com_rx_pkt(struct ena_com_io_cq *io_cq,
 	u16 cdesc_idx = 0;
 	u16 nb_hw_desc;
 	u16 i = 0;
+	int rc;
 
 	WARN(io_cq->direction != ENA_COM_IO_QUEUE_DIRECTION_RX, "wrong Q type");
 
-	nb_hw_desc = ena_com_cdesc_rx_pkt_get(io_cq, &cdesc_idx);
+	rc = ena_com_cdesc_rx_pkt_get(io_cq, &cdesc_idx, &nb_hw_desc);
+	if (unlikely(rc != 0))
+		return -EFAULT;
+
 	if (nb_hw_desc == 0) {
 		ena_rx_ctx->descs = nb_hw_desc;
 		return 0;
diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c
index 4d7446bb66b771..60d8d2d99af14d 100644
--- a/drivers/net/ethernet/amazon/ena/ena_netdev.c
+++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c
@@ -1335,6 +1335,8 @@ static int ena_clean_rx_irq(struct ena_ring *rx_ring, struct napi_struct *napi,
 	if (rc == -ENOSPC) {
 		ena_increase_stat(&rx_ring->rx_stats.bad_desc_num, 1, &rx_ring->syncp);
 		ena_reset_device(adapter, ENA_REGS_RESET_TOO_MANY_RX_DESCS);
+	} else if (rc == -EFAULT) {
+		ena_reset_device(adapter, ENA_REGS_RESET_RX_DESCRIPTOR_MALFORMED);
 	} else {
 		ena_increase_stat(&rx_ring->rx_stats.bad_req_id, 1,
 				  &rx_ring->syncp);
diff --git a/drivers/net/ethernet/amazon/ena/ena_regs_defs.h b/drivers/net/ethernet/amazon/ena/ena_regs_defs.h
index 1e007a41a52564..23466d62fbc5c5 100644
--- a/drivers/net/ethernet/amazon/ena/ena_regs_defs.h
+++ b/drivers/net/ethernet/amazon/ena/ena_regs_defs.h
@@ -21,6 +21,7 @@ enum ena_regs_reset_reason_types {
 	ENA_REGS_RESET_USER_TRIGGER                 = 12,
 	ENA_REGS_RESET_GENERIC                      = 13,
 	ENA_REGS_RESET_MISS_INTERRUPT               = 14,
+	ENA_REGS_RESET_RX_DESCRIPTOR_MALFORMED	    = 16,
 };
 
 /* ena_registers offsets */
-- 
2.39.5

Re: [PATCH 6.6.y] net: ena: Add validation for completion descriptors consistency

[netdev-bot+sinfo]

Hi!

This is an automated message. This series looks like a fix, but its
commit messages seem to be missing some information:

 - How the issue was discovered, e.g. hit in production, hit during
   development, syzbot report, manual code inspection, LLM or static
   analysis tool scan.

 - Whether the issue was actually triggered, or is only theoretical
   (e.g. found by code inspection). If it was triggered please include
   the symptoms, like the stack trace or error messages.

 - What hardware the change was tested on. For driver fixes please
   mention the device (and if relevant firmware version) used for
   testing, or say that the change was not tested on real hardware.

Please do not repost the series just to address the above. Instead,
reply to this email with the missing information, so that reviewers
can take it into account. If the series needs another revision for
other reasons, please include the information in the commit messages
then.

The evaluation is done by an LLM so it may be wrong, if you think
that is the case please reply and explain.