[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <78151482AFE8973A+20250801090600.544000-1-wangyuli@uniontech.com>
Date: Fri, 1 Aug 2025 17:06:00 +0800
From: WangYuli <wangyuli@...ontech.com>
To: alexander.usyskin@...el.com,
arnd@...db.de,
gregkh@...uxfoundation.org
Cc: linux-kernel@...r.kernel.org,
zhanjun@...ontech.com,
niecheng1@...ontech.com,
guanwentao@...ontech.com,
WangYuli <wangyuli@...ontech.com>
Subject: [PATCH] mei: vsc: fix potential array bounds violation in ACE address allocation
When ACE images require dynamic address allocation, the code accesses
frags[frag_index - 1] without bounds checking. This could lead to:
- Array underflow if frag_index is 0
- Use of uninitialized fragment data for address calculations
- Silent failures in address allocation
Add proper validation before accessing the previous fragment and
provide clear error messages when validation fails.
Signed-off-by: WangYuli <wangyuli@...ontech.com>
---
drivers/misc/mei/vsc-fw-loader.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/drivers/misc/mei/vsc-fw-loader.c b/drivers/misc/mei/vsc-fw-loader.c
index 43abefa806e1..e2d318ecb76a 100644
--- a/drivers/misc/mei/vsc-fw-loader.c
+++ b/drivers/misc/mei/vsc-fw-loader.c
@@ -516,7 +516,19 @@ static int vsc_identify_ace_image(struct vsc_fw_loader *fw_loader)
frag->type = ace_image_map[i].image_type;
if (!frag->location) {
+ if (frag_index == 0) {
+ dev_err(fw_loader->dev,
+ "Cannot auto-allocate address for first fragment\n");
+ ret = -EINVAL;
+ goto err_release_image;
+ }
last_frag = &fw_loader->frags[frag_index - 1];
+ if (!last_frag->location || !last_frag->size) {
+ dev_err(fw_loader->dev,
+ "Previous fragment not properly initialized for auto-allocation\n");
+ ret = -EINVAL;
+ goto err_release_image;
+ }
frag->location =
ALIGN(last_frag->location + last_frag->size, SZ_4K);
}
--
2.50.1
Powered by blists - more mailing lists