Thread (18 messages) 18 messages, 2 authors, 2026-02-02

Re: [net-next,v7,01/13] octeontx2-af: npc: cn20k: Index management

From: Jakub Kicinski <kuba@kernel.org>
Date: 2026-01-31 03:33:22
Also in: lkml

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.

For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
octeontx2-af: npc: cn20k: Index management

This patch adds MCAM index management for CN20K silicon, implementing
contiguous and non-contiguous allocation schemes across the new subbank
architecture with reference, limit, and priority parameters.
quoted hunk ↗ jump to hunk
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npc.c b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npc.c
new file mode 100644
index 000000000000..de3d5ad762d1
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npc.c
[ ... ]
+static int __npc_subbank_alloc(struct rvu *rvu, struct npc_subbank *sb,
+			       int key_type, int ref, int limit, int prio,
+			       bool contig, int count, u16 *mcam_idx,
+			       int idx_sz, bool max_alloc, int *alloc_cnt)
+{
+	int cnt, t, b, i, blkaddr;
+	bool new_sub_bank = false;
+	unsigned long *bmap;
+	u16 *save = NULL;
+	int sidx, eidx;
+	bool diffbank;
+	int bw, bfree;
+	int rc = 0;
+	bool ret;
[ ... ]
+	/* Mark new subbank bank as used */
+	if (new_sub_bank) {
+		blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_NPC, 0);
+		if (blkaddr < 0) {
+			dev_err(rvu->dev,
+				"%s: NPC block not implemented\n", __func__);
+			goto err;
                        ^^^^^^^^

When rvu_get_blkaddr() fails here, the code jumps to the err label but rc
is still 0 from initialization. Every other error path in this function
sets rc before going to err. Should rc be set to blkaddr (or another error
code) before the goto?
+		}
+
+		rc =  __npc_subbank_mark_used(rvu, sb, key_type);
[ ... ]
+err:
+	kfree(save);
+	return rc;
+}
[ ... ]
+static void npc_subbank_init(struct rvu *rvu, struct npc_subbank *sb, int idx)
+{
+	mutex_init(&sb->lock);
[ ... ]
+void npc_cn20k_deinit(struct rvu *rvu)
+{
+	int i;
+
+	xa_destroy(&npc_priv.xa_sb_used);
+	xa_destroy(&npc_priv.xa_sb_free);
+	xa_destroy(&npc_priv.xa_idx2pf_map);
+	xa_destroy(&npc_priv.xa_pf_map);
+
+	for (i = 0; i < npc_priv.pf_cnt; i++)
+		xa_destroy(&npc_priv.xa_pf2idx_map[i]);
+
+	kfree(npc_priv.xa_pf2idx_map);
+	kfree(npc_priv.sb);
The subbank mutexes are initialized in npc_subbank_init() but are not
destroyed here. Is mutex_destroy() needed for each sb->lock before freeing
npc_priv.sb? The same issue applies to the fail2 error path in
npc_priv_init().
+	kfree(subbank_srch_order);
+}
-- 
pw-bot: cr
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help