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  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Mon, 18 Mar 2019 17:49:06 +0000
From:   Marc Zyngier <marc.zyngier@....com>
To:     Guenter Roeck <linux@...ck-us.net>,
        Heikki Krogerus <heikki.krogerus@...ux.intel.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: [PATCH v2] usb: typec: fusb302: Fix debugfs mutex initialisation

Running a kernel with the fusb302 driver and lockdep enabled
leads to an unpleasant warning:

[    4.617477] INFO: trying to register non-static key.
[    4.617930] the code is fine but needs lockdep annotation.
[    4.618418] turning off the locking correctness validator.
[    4.618913] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.1.0-rc1-00007-g3542533f3fc9 #13
[    4.619620] Hardware name: rockchip evb_rk3399/evb_rk3399, BIOS 2019.04-rc3-00124-g2feec69fb1 03/15/2019
[    4.620454] Call trace:
[    4.620693]  dump_backtrace+0x0/0x138
[    4.621028]  show_stack+0x24/0x30
[    4.621336]  dump_stack+0xbc/0x104
[    4.621649]  register_lock_class+0x594/0x598
[    4.622036]  __lock_acquire+0x80/0x11b8
[    4.622384]  lock_acquire+0xdc/0x260
[    4.622711]  __mutex_lock+0x90/0x8a0
[    4.623037]  mutex_lock_nested+0x3c/0x50
[    4.623394]  _fusb302_log+0x88/0x1f0
[    4.623721]  fusb302_log+0x7c/0xa0
[    4.624033]  tcpm_init+0x5c/0x190
[    4.624336]  tcpm_init+0x3c/0x130
[    4.624640]  tcpm_register_port+0x574/0x878
[    4.625019]  fusb302_probe+0x2c8/0x590

Despite what the message says, the code isn't fine, as it tries to
make use of the fusb302_log facility pretty early. This requires the
logbuffer_lock mutex to be initialised, but that only happens much
later. Boo.

Hoist the fusb302_debugfs_init call before tcpm_register_port so that
we can enjoy a working mutex. At Guenter's request, also add teardown
of the debugfs facility on the error path.

Signed-off-by: Marc Zyngier <marc.zyngier@....com>
---
 drivers/usb/typec/tcpm/fusb302.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/typec/tcpm/fusb302.c b/drivers/usb/typec/tcpm/fusb302.c
index e9344997329c..76cb8be6f3eb 100644
--- a/drivers/usb/typec/tcpm/fusb302.c
+++ b/drivers/usb/typec/tcpm/fusb302.c
@@ -1790,12 +1790,13 @@ static int fusb302_probe(struct i2c_client *client,
 			goto destroy_workqueue;
 	}
 
+	fusb302_debugfs_init(chip);
 	chip->tcpm_port = tcpm_register_port(&client->dev, &chip->tcpc_dev);
 	if (IS_ERR(chip->tcpm_port)) {
 		ret = PTR_ERR(chip->tcpm_port);
 		if (ret != -EPROBE_DEFER)
 			dev_err(dev, "cannot register tcpm port, ret=%d", ret);
-		goto destroy_workqueue;
+		goto unregister_debugfs;
 	}
 
 	ret = devm_request_threaded_irq(chip->dev, chip->gpio_int_n_irq,
@@ -1807,13 +1808,14 @@ static int fusb302_probe(struct i2c_client *client,
 		goto tcpm_unregister_port;
 	}
 	enable_irq_wake(chip->gpio_int_n_irq);
-	fusb302_debugfs_init(chip);
 	i2c_set_clientdata(client, chip);
 
 	return ret;
 
 tcpm_unregister_port:
 	tcpm_unregister_port(chip->tcpm_port);
+unregister_debugfs:
+	fusb302_debugfs_exit(chip);
 destroy_workqueue:
 	destroy_workqueue(chip->wq);
 
-- 
2.20.1

Powered by blists - more mailing lists