From: Hans de Goede <hdegoede@redhat.com>
To: "Filipe Laíns" <lains@riseup.net>,
"Bastien Nocera" <hadess@hadess.net>,
"Jiri Kosina" <jikos@kernel.org>,
"Benjamin Tissoires" <benjamin.tissoires@redhat.com>
Cc: Hans de Goede <hdegoede@redhat.com>, linux-input@vger.kernel.org
Subject: [PATCH v2 12/14] HID: logitech-hidpp: Fix connect event race
Date: Sun, 8 Oct 2023 11:54:55 +0200 [thread overview]
Message-ID: <20231008095458.8926-13-hdegoede@redhat.com> (raw)
In-Reply-To: <20231008095458.8926-1-hdegoede@redhat.com>
There is a connect event race in hidpp_probe() in these 2 lines:
connected = hidpp_root_get_protocol_version(hidpp) == 0;
atomic_set(&hidpp->connected, connected);
Specifically the following can happen:
1. This line from hidpp_probe() is executed:
connected = hidpp_root_get_protocol_version(hidpp) == 0;
and sets connected to false;
2. A connect-event packet is received and does:
atomic_set(&hidpp->connected, true);
3. The next line from hidpp_probe() is executed:
atomic_set(&hidpp->connected, connected);
and sets the atomic_t back to 0 again.
4. hidpp_connect_event() runs and sees the connected device
as disconnected because of this.
To fix this make hidpp_connect_event() query the connection status
of the device itself instead of having it rely on possibly stale
data cached in struct hidpp_device.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
drivers/hid/hid-logitech-hidpp.c | 29 +++++++----------------------
1 file changed, 7 insertions(+), 22 deletions(-)
diff --git a/drivers/hid/hid-logitech-hidpp.c b/drivers/hid/hid-logitech-hidpp.c
index 2b0a2ea5da22..37213dcc9d9c 100644
--- a/drivers/hid/hid-logitech-hidpp.c
+++ b/drivers/hid/hid-logitech-hidpp.c
@@ -194,7 +194,6 @@ struct hidpp_device {
struct work_struct work;
struct kfifo delayed_work_fifo;
- atomic_t connected;
struct input_dev *delayed_input;
unsigned long quirks;
@@ -3893,8 +3892,6 @@ static int hidpp_raw_hidpp_event(struct hidpp_device *hidpp, u8 *data,
}
if (unlikely(hidpp_report_is_connect_event(hidpp, report))) {
- atomic_set(&hidpp->connected,
- !(report->rap.params[0] & (1 << 6)));
if (schedule_work(&hidpp->work) == 0)
dbg_hid("%s: connect event already queued\n", __func__);
return 1;
@@ -4189,12 +4186,14 @@ static struct input_dev *hidpp_allocate_input(struct hid_device *hdev)
static void hidpp_connect_event(struct hidpp_device *hidpp)
{
struct hid_device *hdev = hidpp->hid_dev;
- int ret = 0;
- bool connected = atomic_read(&hidpp->connected);
struct input_dev *input;
char *name, *devm_name;
+ int ret;
- if (!connected) {
+ /* Get device version to check if it is connected */
+ ret = hidpp_root_get_protocol_version(hidpp);
+ if (ret) {
+ hid_info(hidpp->hid_dev, "Disconnected\n");
if (hidpp->battery.ps) {
hidpp->battery.online = false;
hidpp->battery.status = POWER_SUPPLY_STATUS_UNKNOWN;
@@ -4235,17 +4234,6 @@ static void hidpp_connect_event(struct hidpp_device *hidpp)
if (ret)
return;
}
-
- /* the device is already connected, we can ask for its name and
- * protocol */
- if (!hidpp->protocol_major) {
- ret = hidpp_root_get_protocol_version(hidpp);
- if (ret) {
- hid_err(hdev, "Can not get the protocol version.\n");
- return;
- }
- }
-
if (hidpp->protocol_major >= 2) {
u8 feature_index;
@@ -4418,7 +4406,6 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
{
struct hidpp_device *hidpp;
int ret;
- bool connected;
unsigned int connect_mask = HID_CONNECT_DEFAULT;
/* report_fixup needs drvdata to be set before we call hid_parse */
@@ -4511,11 +4498,9 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
/*
* Now that incoming packets are enabled and will not be disabled
- * again (which may cause missing packets) check the connected state
- * of the device.
+ * again (which may cause missing packets) queue hidpp_connect_event()
+ * to check the connected state of the device.
*/
- connected = hidpp_root_get_protocol_version(hidpp) == 0;
- atomic_set(&hidpp->connected, connected);
schedule_work(&hidpp->work);
flush_work(&hidpp->work);
--
2.41.0
next prev parent reply other threads:[~2023-10-08 9:56 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-08 9:54 [PATCH v2 00/14] HID: logitech-hidpp: Avoid hidpp_connect_event() running while probe() restarts IO Hans de Goede
2023-10-08 9:54 ` [PATCH v2 01/14] HID: logitech-hidpp: Revert "Don't restart communication if not necessary" Hans de Goede
2023-10-08 9:54 ` [PATCH v2 02/14] HID: logitech-hidpp: Move hidpp_overwrite_name() to before connect check Hans de Goede
2023-10-08 9:54 ` [PATCH v2 03/14] HID: logitech-hidpp: Add hidpp_non_unifying_init() helper Hans de Goede
2023-10-08 9:54 ` [PATCH v2 04/14] HID: logitech-hidpp: Remove connected check for non-unifying devices Hans de Goede
2023-10-08 9:54 ` [PATCH v2 05/14] HID: logitech-hidpp: Move get_wireless_feature_index() check to hidpp_connect_event() Hans de Goede
2023-10-08 9:54 ` [PATCH v2 06/14] HID: logitech-hidpp: Remove wtp_get_config() call from probe() Hans de Goede
2023-10-08 9:54 ` [PATCH v2 07/14] HID: logitech-hidpp: Remove connected check for g920_get_config() call Hans de Goede
2023-10-08 9:54 ` [PATCH v2 08/14] HID: logitech-hidpp: Add a hidpp_connect_and_start() helper Hans de Goede
2023-10-08 9:54 ` [PATCH v2 09/14] HID: logitech-hidpp: Move the connected check to after restarting IO Hans de Goede
2023-10-08 9:54 ` [PATCH v2 10/14] HID: logitech-hidpp: Move g920_get_config() to just before hidpp_ff_init() Hans de Goede
2023-10-08 9:54 ` [PATCH v2 11/14] HID: logitech-hidpp: Remove unused connected param from *_connect() Hans de Goede
2023-10-08 9:54 ` Hans de Goede [this message]
2023-10-08 9:54 ` [PATCH v2 13/14] HID: logitech-hidpp: Avoid hidpp_connect_event() running while probe() restarts IO Hans de Goede
2023-10-08 9:54 ` [PATCH v2 14/14] HID: logitech-hidpp: Drop delayed_work_cb() Hans de Goede
2023-10-08 10:30 ` [PATCH v2 00/14] HID: logitech-hidpp: Avoid hidpp_connect_event() running while probe() restarts IO Hans de Goede
2023-10-09 8:13 ` Benjamin Tissoires
2023-10-09 15:00 ` Hans de Goede
2023-10-10 9:36 ` Benjamin Tissoires
2023-10-10 9:40 ` Hans de Goede
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=20231008095458.8926-13-hdegoede@redhat.com \
--to=hdegoede@redhat.com \
--cc=benjamin.tissoires@redhat.com \
--cc=hadess@hadess.net \
--cc=jikos@kernel.org \
--cc=lains@riseup.net \
--cc=linux-input@vger.kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).