Netdev List
 help / color / mirror / Atom feed
From: Ruoyu Wang <ruoyuw560@gmail.com>
To: michael.chan@broadcom.com, pavan.chebbi@broadcom.com,
	andrew+netdev@lunn.ch, davem@davemloft.net, edumazet@google.com,
	kuba@kernel.org, pabeni@redhat.com
Cc: jacob.e.keller@intel.com, andrew.gospodarek@broadcom.com,
	vikas.gupta@broadcom.com, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org, Ruoyu Wang <ruoyuw560@gmail.com>
Subject: [PATCH net] bnxt_en: Handle partially initialized auxiliary devices
Date: Wed,  8 Jul 2026 22:34:01 +0800	[thread overview]
Message-ID: <20260708143401.3167477-1-ruoyuw560@gmail.com> (raw)

bnxt_aux_devices_init() calls auxiliary_device_init() before allocating
and attaching the bnxt_en_dev and ULP table.  After
auxiliary_device_init() succeeds, the auxiliary bus owns the embedded
device lifetime and the driver must unwind later errors with
auxiliary_device_uninit(), which runs bnxt_aux_dev_release().

The release callback currently assumes that aux_priv->id, aux_priv->edev
and edev->net were all populated.  If the bnxt_en_dev allocation fails,
release dereferences aux_priv->edev while it is still NULL.  If a later
failure happens before aux_priv->id is assigned, release uses the zeroed
id field and can clear the wrong auxiliary-device slot while unwinding a
partially initialized device.

Set aux_priv->id before auxiliary_device_init() so the release path can
identify the slot, make release tolerate missing partial state, and clear
bp->aux_priv[idx] on the unwind path where release cannot derive bp from
an edev.

This issue was found by a static analysis checker and confirmed by manual
source review.

Fixes: 194fad5b2781 ("bnxt_en: Refactor bnxt_rdma_aux_device_init/uninit functions")
Signed-off-by: Ruoyu Wang <ruoyuw560@gmail.com>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
index 5c751933da6a9..e85d1b6c9fb1f 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c
@@ -472,12 +472,17 @@ static void bnxt_aux_dev_release(struct device *dev)
 {
 	struct bnxt_aux_priv *aux_priv =
 		container_of(dev, struct bnxt_aux_priv, aux_dev.dev);
-	struct bnxt *bp = netdev_priv(aux_priv->edev->net);
+	struct bnxt_en_dev *edev = aux_priv->edev;
+	struct bnxt *bp = edev && edev->net ? netdev_priv(edev->net) : NULL;
 
-	kfree(aux_priv->edev->ulp_tbl);
-	bp->edev[aux_priv->id] = NULL;
-	kfree(aux_priv->edev);
-	bp->aux_priv[aux_priv->id] = NULL;
+	if (edev) {
+		kfree(edev->ulp_tbl);
+		if (bp)
+			bp->edev[aux_priv->id] = NULL;
+		kfree(edev);
+	}
+	if (bp)
+		bp->aux_priv[aux_priv->id] = NULL;
 	kfree(aux_priv);
 }
 
@@ -571,6 +576,7 @@ void bnxt_aux_devices_init(struct bnxt *bp)
 		aux_dev->name = bnxt_aux_devices[idx].name;
 		aux_dev->dev.parent = &bp->pdev->dev;
 		aux_dev->dev.release = bnxt_aux_dev_release;
+		aux_priv->id = idx;
 
 		rc = auxiliary_device_init(aux_dev);
 		if (rc) {
@@ -598,12 +604,12 @@ void bnxt_aux_devices_init(struct bnxt *bp)
 		bp->edev[idx] = edev;
 		if (idx == BNXT_AUXDEV_RDMA)
 			bp->ulp_num_msix_want = bnxt_set_dflt_ulp_msix(bp);
-		aux_priv->id = idx;
 		bnxt_auxdev_set_state(bp, idx, BNXT_ADEV_STATE_INIT);
 
 		continue;
 aux_dev_uninit:
 		auxiliary_device_uninit(aux_dev);
+		bp->aux_priv[idx] = NULL;
 next_auxdev:
 		if (idx == BNXT_AUXDEV_RDMA)
 			bp->flags &= ~BNXT_FLAG_ROCE_CAP;
-- 
2.51.0

             reply	other threads:[~2026-07-08 14:34 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-08 14:34 Ruoyu Wang [this message]
2026-07-09  4:25 ` [PATCH net] bnxt_en: Handle partially initialized auxiliary devices Pavan Chebbi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260708143401.3167477-1-ruoyuw560@gmail.com \
    --to=ruoyuw560@gmail.com \
    --cc=andrew+netdev@lunn.ch \
    --cc=andrew.gospodarek@broadcom.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=jacob.e.keller@intel.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=michael.chan@broadcom.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=pavan.chebbi@broadcom.com \
    --cc=vikas.gupta@broadcom.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox