From: Chris Lu <chris.lu@mediatek.com>
To: Marcel Holtmann <marcel@holtmann.org>,
Johan Hedberg <johan.hedberg@gmail.com>,
Luiz Von Dentz <luiz.dentz@gmail.com>
Cc: Sean Wang <sean.wang@mediatek.com>,
Will Lee <will-cy.Lee@mediatek.com>, SS Wu <ss.wu@mediatek.com>,
Steve Lee <steve.lee@mediatek.com>,
Paul Menzel <pmenzel@molgen.mpg.de>,
linux-bluetooth <linux-bluetooth@vger.kernel.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
linux-mediatek <linux-mediatek@lists.infradead.org>,
Chris Lu <chris.lu@mediatek.com>
Subject: [PATCH v13 1/7] Bluetooth: btmtksdio: Fix SKB handling issues in TX path
Date: Thu, 16 Jul 2026 19:23:46 +0800 [thread overview]
Message-ID: <20260716112353.2192865-2-chris.lu@mediatek.com> (raw)
In-Reply-To: <20260716112353.2192865-1-chris.lu@mediatek.com>
Fix two SKB handling issues in btmtksdio_tx_packet():
1. DMA out-of-bounds read: The driver aligns transfer size to 256 bytes
using round_up(), but does not ensure the skb buffer has sufficient
tailroom, causing DMA to read beyond the buffer. Fix by expanding skb
tailroom if needed and zero-filling the padding.
2. Cloned SKB corruption: The driver modifies SKB header (via skb_push)
and tail (via skb_put_zero) without checking if the SKB is cloned.
If the Bluetooth core passes a cloned SKB (e.g., from L2CAP
retransmission queues), the driver corrupts the shared data buffer
for other subsystems. Fix by calling skb_unshare() at function entry
to ensure we have a private copy before any modifications.
Fixes: 9aebfd4a2200 ("Bluetooth: mediatek: add support for MediaTek MT7663S and MT7668S SDIO devices")
Signed-off-by: Chris Lu <chris.lu@mediatek.com>
Assisted-by: Claude:Sonnet-4.5
---
drivers/bluetooth/btmtksdio.c | 42 ++++++++++++++++++++++++++++++-----
1 file changed, 37 insertions(+), 5 deletions(-)
diff --git a/drivers/bluetooth/btmtksdio.c b/drivers/bluetooth/btmtksdio.c
index 5b0fab7b89b5..8a7f2e1c4d56 100644
--- a/drivers/bluetooth/btmtksdio.c
+++ b/drivers/bluetooth/btmtksdio.c
@@ -272,9 +272,20 @@ static int btmtksdio_tx_packet(struct btmtksdio_dev *bdev,
struct sk_buff *skb)
{
struct mtkbtsdio_hdr *sdio_hdr;
+ unsigned int padded_len, pad_len;
int err;
- /* Make sure that there are enough rooms for SDIO header */
+ /* Ensure SKB is not cloned before any modification.
+ * If cloned (e.g., L2CAP retransmission queues), create a
+ * private copy to avoid corrupting shared data buffer.
+ */
+ skb = skb_unshare(skb, GFP_ATOMIC);
+ if (!skb)
+ return -ENOMEM;
+
+ /* Make sure that there are enough rooms for SDIO header.
+ * After skb_unshare(), we have exclusive ownership of the buffer.
+ */
if (unlikely(skb_headroom(skb) < sizeof(*sdio_hdr))) {
err = pskb_expand_head(skb, sizeof(*sdio_hdr), 0,
GFP_ATOMIC);
@@ -290,18 +301,39 @@ static int btmtksdio_tx_packet(struct btmtksdio_dev *bdev,
sdio_hdr->reserved = cpu_to_le16(0);
sdio_hdr->bt_type = hci_skb_pkt_type(skb);
+ /* Calculate padded length for block-aligned DMA transfer.
+ * SDIO requires transfers to be block-aligned (MTK_SDIO_BLOCK_SIZE).
+ * Pad with zeros to prevent DMA from reading beyond skb buffer.
+ */
+ padded_len = round_up(skb->len, MTK_SDIO_BLOCK_SIZE);
+ pad_len = padded_len - skb->len;
+
+ if (pad_len > 0) {
+ /* Ensure sufficient tailroom for padding */
+ if (unlikely(skb_tailroom(skb) < pad_len)) {
+ err = pskb_expand_head(skb, 0, pad_len, GFP_ATOMIC);
+ if (err < 0)
+ goto err_skb_pull;
+ /* Reassign sdio_hdr after buffer reallocation */
+ sdio_hdr = (void *)skb->data;
+ }
+
+ /* Zero-fill padding to prevent information disclosure */
+ skb_put_zero(skb, pad_len);
+ }
+
clear_bit(BTMTKSDIO_HW_TX_READY, &bdev->tx_state);
- err = sdio_writesb(bdev->func, MTK_REG_CTDR, skb->data,
- round_up(skb->len, MTK_SDIO_BLOCK_SIZE));
+ err = sdio_writesb(bdev->func, MTK_REG_CTDR, skb->data, padded_len);
if (err < 0)
- goto err_skb_pull;
+ goto err_skb_trim;
- bdev->hdev->stat.byte_tx += skb->len;
+ /* Record actual transmitted data (excluding padding) */
+ bdev->hdev->stat.byte_tx += le16_to_cpu(sdio_hdr->len);
kfree_skb(skb);
return 0;
+err_skb_trim:
+ if (pad_len > 0)
+ skb_trim(skb, skb->len - pad_len);
err_skb_pull:
skb_pull(skb, sizeof(*sdio_hdr));
+ kfree_skb(skb);
return err;
}
--
2.45.2
next prev parent reply other threads:[~2026-07-16 11:24 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-16 11:23 [PATCH v13 0/7] Bluetooth: btmtk: Add MT7928 support Chris Lu
2026-07-16 11:23 ` Chris Lu [this message]
2026-07-16 11:23 ` [PATCH v13 2/7] Bluetooth: btmtk: Add firmware size validation in btmtk_setup_firmware_79xx() Chris Lu
2026-07-16 11:23 ` [PATCH v13 3/7] Bluetooth: btmtksdio: Pass hardware dev_id to mt79xx_setup() Chris Lu
2026-07-16 11:23 ` [PATCH v13 4/7] Bluetooth: btmtksdio: Remove redundant firmware filename override Chris Lu
2026-07-16 11:23 ` [PATCH v13 5/7] Bluetooth: btmtk: Improve BT firmware logging Chris Lu
2026-07-16 11:23 ` [PATCH v13 6/7] Bluetooth: btmtk: Replace magic numbers with WMT packet flag enum Chris Lu
2026-07-16 11:23 ` [PATCH v13 7/7] Bluetooth: btmtk: Add MT7928 support Chris Lu
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=20260716112353.2192865-2-chris.lu@mediatek.com \
--to=chris.lu@mediatek.com \
--cc=johan.hedberg@gmail.com \
--cc=linux-bluetooth@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mediatek@lists.infradead.org \
--cc=luiz.dentz@gmail.com \
--cc=marcel@holtmann.org \
--cc=pmenzel@molgen.mpg.de \
--cc=sean.wang@mediatek.com \
--cc=ss.wu@mediatek.com \
--cc=steve.lee@mediatek.com \
--cc=will-cy.Lee@mediatek.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