[PATCH] staging: wilc1000: fix double mutex_unlock on failure path in wilc_wlan_cleanup()

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

The patch adds return in case of failure.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>
---
 drivers/staging/wilc1000/wilc_wlan.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/staging/wilc1000/wilc_wlan.c b/drivers/staging/wilc1000/wilc_wlan.c
index c02665747705..cd7f52a51173 100644
--- a/drivers/staging/wilc1000/wilc_wlan.c
+++ b/drivers/staging/wilc1000/wilc_wlan.c

@@ -1703,12 +1703,14 @@ void wilc_wlan_cleanup(struct net_device *dev)
 	if (!ret) {
 		PRINT_ER("Error while reading reg\n");
 		release_bus(RELEASE_ALLOW_SLEEP);
+		return;
 	}
 	PRINT_ER("Writing ABORT reg\n");
 	ret = p->hif_func.hif_write_reg(WILC_GP_REG_0, (reg | ABORT_INT));
 	if (!ret) {
 		PRINT_ER("Error while writing reg\n");
 		release_bus(RELEASE_ALLOW_SLEEP);
+		return;
 	}
 	release_bus(RELEASE_ALLOW_SLEEP);
 	/**

-- 
1.9.1

On Sat, Dec 05, 2015 at 01:04:34AM +0300, Alexey Khoroshilov wrote:

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

The patch adds return in case of failure.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>
---
 drivers/staging/wilc1000/wilc_wlan.c | 2 ++
 1 file changed, 2 insertions(+)

No longer applies to my tree, can you rebase it against staging-testing
and resend?

thanks,

greg k-h

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>
---
 drivers/staging/wilc1000/wilc_wlan.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/staging/wilc1000/wilc_wlan.c b/drivers/staging/wilc1000/wilc_wlan.c
index a73e99f..4b7c8e9 100644
--- a/drivers/staging/wilc1000/wilc_wlan.c
+++ b/drivers/staging/wilc1000/wilc_wlan.c

@@ -1459,15 +1459,16 @@ void wilc_wlan_cleanup(struct net_device *dev)
 	ret = p->hif_func.hif_read_reg(wilc, WILC_GP_REG_0, &reg);
 	if (!ret) {
 		PRINT_ER("Error while reading reg\n");
-		release_bus(wilc, RELEASE_ALLOW_SLEEP);
+		goto _unlock;
 	}
 	PRINT_ER("Writing ABORT reg\n");
 	ret = p->hif_func.hif_write_reg(wilc, WILC_GP_REG_0,
 					(reg | ABORT_INT));
 	if (!ret) {
 		PRINT_ER("Error while writing reg\n");
-		release_bus(wilc, RELEASE_ALLOW_SLEEP);
+		goto _unlock;
 	}
+_unlock:
 	release_bus(wilc, RELEASE_ALLOW_SLEEP);
 	p->hif_func.hif_deinit(NULL);
 }

-- 
1.9.1

On Mon, Dec 21, 2015 at 12:46:51AM +0300, Alexey Khoroshilov wrote:

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>

Doesn't apply to my tree anymore, can you rebase this on the
staging-testing branch of staging.git?

thanks,

greg k-h

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>
---
 drivers/staging/wilc1000/wilc_wlan.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/staging/wilc1000/wilc_wlan.c b/drivers/staging/wilc1000/wilc_wlan.c
index 83af51b..b8c4a63 100644
--- a/drivers/staging/wilc1000/wilc_wlan.c
+++ b/drivers/staging/wilc1000/wilc_wlan.c

@@ -1381,15 +1381,16 @@ void wilc_wlan_cleanup(struct net_device *dev)
 	ret = wilc->hif_func->hif_read_reg(wilc, WILC_GP_REG_0, &reg);
 	if (!ret) {
 		PRINT_ER("Error while reading reg\n");
-		release_bus(wilc, RELEASE_ALLOW_SLEEP);
+		goto unlock;
 	}
 	PRINT_ER("Writing ABORT reg\n");
 	ret = wilc->hif_func->hif_write_reg(wilc, WILC_GP_REG_0,
 					(reg | ABORT_INT));
 	if (!ret) {
 		PRINT_ER("Error while writing reg\n");
-		release_bus(wilc, RELEASE_ALLOW_SLEEP);
+		goto unlock;
 	}
+unlock:
 	release_bus(wilc, RELEASE_ALLOW_SLEEP);
 	wilc->hif_func->hif_deinit(NULL);
 }

-- 
1.9.1

On Tue, Dec 22, 2015 at 08:39:26PM +0300, Alexey Khoroshilov wrote:

If hif_read_reg() or hif_write_reg() fail in wilc_wlan_cleanup(),
it calls release_bus() and continues execution. But it leads to double
release_bus() call that means double unlock of g_linux_wlan->hif_cs mutex.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <redacted>
---
 drivers/staging/wilc1000/wilc_wlan.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Why is this 'v3'?  What changed from the other versions?  Please always
document it below the --- line so that we have a chance when reviewing
them.

Please fix up and resend with that information.

thanks,

greg k-h

On Mon, Dec 21, 2015 at 12:46:51AM +0300, Alexey Khoroshilov wrote:

If you're redoing this anyway, could we get rid of the underscore in the
label name?  Just unlock: is fine.

regards,
dan carpenter