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:   Alex Elder <elder@linaro.org>
To:     davem@davemloft.net, kuba@kernel.org
Cc:     mka@chromium.org, evgreen@chromium.org, bjorn.andersson@linaro.org,
        cpratapa@codeaurora.org, avuyyuru@codeaurora.org,
        jponduru@codeaurora.org, subashab@codeaurora.org, elder@kernel.org,
        netdev@vger.kernel.org, linux-arm-msm@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: [PATCH net-next] net: ipa: compute proper aggregation limit
Date:   Thu, 21 Apr 2022 13:53:33 -0500
Message-Id: <20220421185333.1371632-1-elder@linaro.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

The aggregation byte limit for an endpoint is currently computed
based on the endpoint's receive buffer size.

However, some bytes at the front of each receive buffer are reserved
on the assumption that--as with SKBs--it might be useful to insert
data (such as headers) before what lands in the buffer.

The aggregation byte limit currently doesn't take into account that
reserved space, and as a result, aggregation could require space
past that which is available in the buffer.

Fix this by reducing the size used to compute the aggregation byte
limit by the NET_SKB_PAD offset reserved for each receive buffer.

Signed-off-by: Alex Elder <elder@linaro.org>
---
Note:  This is a bug, but it won't apply cleanly to older kernels,
       so I will be posting back-ports separately.

 drivers/net/ipa/ipa_endpoint.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ipa/ipa_endpoint.c b/drivers/net/ipa/ipa_endpoint.c
index 888e94278a84f..e133eb2bebcfb 100644
--- a/drivers/net/ipa/ipa_endpoint.c
+++ b/drivers/net/ipa/ipa_endpoint.c
@@ -130,9 +130,10 @@ static bool ipa_endpoint_data_valid_one(struct ipa *ipa, u32 count,
 		 */
 		if (data->endpoint.config.aggregation) {
 			limit += SZ_1K * aggr_byte_limit_max(ipa->version);
-			if (buffer_size > limit) {
+			if (buffer_size - NET_SKB_PAD > limit) {
 				dev_err(dev, "RX buffer size too large for aggregated RX endpoint %u (%u > %u)\n",
-					data->endpoint_id, buffer_size, limit);
+					data->endpoint_id,
+					buffer_size - NET_SKB_PAD, limit);
 
 				return false;
 			}
@@ -739,6 +740,7 @@ static void ipa_endpoint_init_aggr(struct ipa_endpoint *endpoint)
 	if (endpoint->data->aggregation) {
 		if (!endpoint->toward_ipa) {
 			const struct ipa_endpoint_rx_data *rx_data;
+			u32 buffer_size;
 			bool close_eof;
 			u32 limit;
 
@@ -746,7 +748,8 @@ static void ipa_endpoint_init_aggr(struct ipa_endpoint *endpoint)
 			val |= u32_encode_bits(IPA_ENABLE_AGGR, AGGR_EN_FMASK);
 			val |= u32_encode_bits(IPA_GENERIC, AGGR_TYPE_FMASK);
 
-			limit = ipa_aggr_size_kb(rx_data->buffer_size);
+			buffer_size = rx_data->buffer_size;
+			limit = ipa_aggr_size_kb(buffer_size - NET_SKB_PAD);
 			val |= aggr_byte_limit_encoded(version, limit);
 
 			limit = IPA_AGGR_TIME_LIMIT;
-- 
2.32.0