Thread (10 messages) 10 messages, 3 authors, 2025-10-22

Re: [PATCH net-next] net: mana: Linearize SKB if TX SGEs exceeds hardware limit

From: Simon Horman <horms@kernel.org>
Date: 2025-10-04 09:38:12
Also in: linux-hyperv, linux-rdma, lkml

On Fri, Oct 03, 2025 at 08:47:24AM -0700, Aditya Garg wrote:
quoted hunk ↗ jump to hunk
The MANA hardware supports a maximum of 30 scatter-gather entries (SGEs)
per TX WQE. In rare configurations where MAX_SKB_FRAGS + 2 exceeds this
limit, the driver drops the skb. Add a check in mana_start_xmit() to
detect such cases and linearize the SKB before transmission.

Return NETDEV_TX_BUSY only for -ENOSPC from mana_gd_post_work_request(),
send other errors to free_sgl_ptr to free resources and record the tx
drop.

Signed-off-by: Aditya Garg <redacted>
Reviewed-by: Dipayaan Roy <redacted>
---
 drivers/net/ethernet/microsoft/mana/mana_en.c | 26 +++++++++++++++----
 include/net/mana/gdma.h                       |  8 +++++-
 include/net/mana/mana.h                       |  1 +
 3 files changed, 29 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c
index f4fc86f20213..22605753ca84 100644
--- a/drivers/net/ethernet/microsoft/mana/mana_en.c
+++ b/drivers/net/ethernet/microsoft/mana/mana_en.c
@@ -20,6 +20,7 @@
 
 #include <net/mana/mana.h>
 #include <net/mana/mana_auxiliary.h>
+#include <linux/skbuff.h>
 
 static DEFINE_IDA(mana_adev_ida);
 
@@ -289,6 +290,19 @@ netdev_tx_t mana_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 	cq = &apc->tx_qp[txq_idx].tx_cq;
 	tx_stats = &txq->stats;
 
+	BUILD_BUG_ON(MAX_TX_WQE_SGL_ENTRIES != MANA_MAX_TX_WQE_SGL_ENTRIES);
+	#if (MAX_SKB_FRAGS + 2 > MANA_MAX_TX_WQE_SGL_ENTRIES)
Hi Aditya,

I see that Eric has made a more substantial review of this patch,
so please follow his advice.

But I wanted to add something to keep in mind for the future: I if the #if
/ #else used here can be replaced by a simple if() statement, then that
would be preferable.  The advantage being that it improves compile
coverage.  And, as these are all constants, I would expect the compiler to
optimise away any unused code.

N.B: I did not check, so please consider this more of a general statement
+		if (skb_shinfo(skb)->nr_frags + 2 > MANA_MAX_TX_WQE_SGL_ENTRIES) {
+			netdev_info_once(ndev,
+					 "nr_frags %d exceeds max supported sge limit. Attempting skb_linearize\n",
+					 skb_shinfo(skb)->nr_frags);
+			if (skb_linearize(skb)) {
+				netdev_warn_once(ndev, "Failed to linearize skb\n");
+				goto tx_drop_count;
+			}
+		}
+	#endif
+
 	pkg.tx_oob.s_oob.vcq_num = cq->gdma_id;
 	pkg.tx_oob.s_oob.vsq_frame = txq->vsq_frame;
 
...
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help