linux-kernel - [PATCH HID v2 03/11] HID: core: remove one more kmemdup on .probe()

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20240910-hid-bpf-hid-generic-v2-3-083dfc189e97@kernel.org>
Date: Tue, 10 Sep 2024 23:43:39 +0900
From: Benjamin Tissoires <bentiss@...nel.org>
To: Jiri Kosina <jikos@...nel.org>, 
 Peter Hutterer <peter.hutterer@...-t.net>, Vicki Pfau <vi@...rift.com>, 
 Shuah Khan <shuah@...nel.org>
Cc: linux-input@...r.kernel.org, linux-kselftest@...r.kernel.org, 
 linux-kernel@...r.kernel.org, bpf@...r.kernel.org, 
 Benjamin Tissoires <bentiss@...nel.org>
Subject: [PATCH HID v2 03/11] HID: core: remove one more kmemdup on
 .probe()

That last kmemdup while opening the report descriptor was required to
have a common kfree() on it.

Move that kmemdup in the only special case it's required (if there is a
.report_fixup()), and add a more elaborated check before freeing
hdev->rdesc, to avoid a double free.

Signed-off-by: Benjamin Tissoires <bentiss@...nel.org>

---

new in v2
---
 drivers/hid/hid-core.c | 28 ++++++++++++++++++++--------
 1 file changed, 20 insertions(+), 8 deletions(-)

diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c
index 172746a082f9..2b70a12a2135 100644
--- a/drivers/hid/hid-core.c
+++ b/drivers/hid/hid-core.c
@@ -685,7 +685,14 @@ static void hid_close_report(struct hid_device *device)
 		INIT_LIST_HEAD(&report_enum->report_list);
 	}
 
-	kfree(device->rdesc);
+	/*
+	 * If the HID driver had a rdesc_fixup() callback, dev->rdesc
+	 * will be allocated by hid-core and needs to be freed.
+	 * Otherwise, it is either equal to dev_rdesc or bpf_rdesc, in
+	 * which cases it'll be freed later on device removal or destroy.
+	 */
+	if (device->rdesc != device->dev_rdesc && device->rdesc != device->bpf_rdesc)
+		kfree(device->rdesc);
 	device->rdesc = NULL;
 	device->rsize = 0;
 
@@ -1206,7 +1213,6 @@ int hid_open_report(struct hid_device *device)
 	struct hid_item item;
 	unsigned int size;
 	const __u8 *start;
-	__u8 *buf = NULL;
 	const __u8 *end;
 	const __u8 *next;
 	int ret;
@@ -1233,17 +1239,23 @@ int hid_open_report(struct hid_device *device)
 		 * on a copy of our report descriptor so it can
 		 * change it.
 		 */
-		buf = kmemdup(start, size, GFP_KERNEL);
+		__u8 *buf = kmemdup(start, size, GFP_KERNEL);
+
 		if (buf == NULL)
 			return -ENOMEM;
 
 		start = device->driver->report_fixup(device, buf, &size);
-	}
 
-	start = kmemdup(start, size, GFP_KERNEL);
-	kfree(buf);
-	if (start == NULL)
-		return -ENOMEM;
+		/*
+		 * The second kmemdup is required in case report_fixup() returns
+		 * a static read-only memory, but we have no idea if that memory
+		 * needs to be cleaned up or not at the end.
+		 */
+		start = kmemdup(start, size, GFP_KERNEL);
+		kfree(buf);
+		if (start == NULL)
+			return -ENOMEM;
+	}
 
 	device->rdesc = start;
 	device->rsize = size;

-- 
2.46.0