* [PATCH v3] serial: samsung: Checks the return value of function
[not found] <CGME20210706061710epcas5p2c11d1bf5afb14774c4d4db93f2b83b33@epcas5p2.samsung.com>
@ 2021-07-06 6:19 ` Tamseel Shams
2021-07-07 8:14 ` Jiri Slaby
2021-07-07 15:55 ` Johan Hovold
0 siblings, 2 replies; 3+ messages in thread
From: Tamseel Shams @ 2021-07-06 6:19 UTC (permalink / raw)
To: krzysztof.kozlowski, gregkh, jirislaby
Cc: linux-arm-kernel, linux-samsung-soc, linux-serial, linux-kernel,
alim.akhtar, Tamseel Shams
"uart_add_one_port" function call may fail and return
some error code, so adding a check for return value.
If it is returning some error code, then displaying the
result, unregistering the driver and then returning from
probe function with error code.
Signed-off-by: Tamseel Shams <m.shams@samsung.com>
---
Changes since v1:
1. Added support to unregister driver on failure of "uart_add_one_port"
function call.
2. Commit message updated.
Changes since v2:
1. Added support to unwind clocks on failure of "uart_add_one_port"
function call.
drivers/tty/serial/samsung_tty.c | 17 ++++++++++++++++-
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
index 9fbc61151c2e..a3f3a17fb54b 100644
--- a/drivers/tty/serial/samsung_tty.c
+++ b/drivers/tty/serial/samsung_tty.c
@@ -2253,7 +2253,11 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
}
dev_dbg(&pdev->dev, "%s: adding port\n", __func__);
- uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
+ ret = uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
+ if (ret < 0) {
+ dev_err(&pdev->dev, "Failed to add uart port, err %d\n", ret);
+ goto add_port_error;
+ }
platform_set_drvdata(pdev, &ourport->port);
/*
@@ -2272,6 +2276,17 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
probe_index++;
return 0;
+
+add_port_error:
+ ourport->port.mapbase = 0;
+ clk_disable_unprepare(ourport->clk);
+ clk_put(ourport->clk);
+ if (!IS_ERR(ourport->baudclk)) {
+ clk_disable_unprepare(ourport->baudclk);
+ clk_put(ourport->baudclk);
+ }
+ uart_unregister_driver(&s3c24xx_uart_drv);
+ return ret;
}
static int s3c24xx_serial_remove(struct platform_device *dev)
--
2.17.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v3] serial: samsung: Checks the return value of function
2021-07-06 6:19 ` [PATCH v3] serial: samsung: Checks the return value of function Tamseel Shams
@ 2021-07-07 8:14 ` Jiri Slaby
2021-07-07 15:55 ` Johan Hovold
1 sibling, 0 replies; 3+ messages in thread
From: Jiri Slaby @ 2021-07-07 8:14 UTC (permalink / raw)
To: Tamseel Shams, krzysztof.kozlowski, gregkh
Cc: linux-arm-kernel, linux-samsung-soc, linux-serial, linux-kernel,
alim.akhtar
On 06. 07. 21, 8:19, Tamseel Shams wrote:
> "uart_add_one_port" function call may fail and return
> some error code, so adding a check for return value.
> If it is returning some error code, then displaying the
> result, unregistering the driver and then returning from
> probe function with error code.
>
> Signed-off-by: Tamseel Shams <m.shams@samsung.com>
> ---
> Changes since v1:
> 1. Added support to unregister driver on failure of "uart_add_one_port"
> function call.
> 2. Commit message updated.
>
> Changes since v2:
> 1. Added support to unwind clocks on failure of "uart_add_one_port"
> function call.
>
> drivers/tty/serial/samsung_tty.c | 17 ++++++++++++++++-
> 1 file changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
> index 9fbc61151c2e..a3f3a17fb54b 100644
> --- a/drivers/tty/serial/samsung_tty.c
> +++ b/drivers/tty/serial/samsung_tty.c
> @@ -2253,7 +2253,11 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
> }
>
> dev_dbg(&pdev->dev, "%s: adding port\n", __func__);
> - uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
> + ret = uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
> + if (ret < 0) {
> + dev_err(&pdev->dev, "Failed to add uart port, err %d\n", ret);
> + goto add_port_error;
> + }
> platform_set_drvdata(pdev, &ourport->port);
>
> /*
> @@ -2272,6 +2276,17 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
> probe_index++;
>
> return 0;
> +
> +add_port_error:
> + ourport->port.mapbase = 0;
> + clk_disable_unprepare(ourport->clk);
> + clk_put(ourport->clk);
> + if (!IS_ERR(ourport->baudclk)) {
> + clk_disable_unprepare(ourport->baudclk);
> + clk_put(ourport->baudclk);
> + }
LGTM, now I only wonder why s3c24xx_serial_remove() does not put clocks?
> + uart_unregister_driver(&s3c24xx_uart_drv);
> + return ret;
> }
>
> static int s3c24xx_serial_remove(struct platform_device *dev)
>
--
js
suse labs
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3] serial: samsung: Checks the return value of function
2021-07-06 6:19 ` [PATCH v3] serial: samsung: Checks the return value of function Tamseel Shams
2021-07-07 8:14 ` Jiri Slaby
@ 2021-07-07 15:55 ` Johan Hovold
1 sibling, 0 replies; 3+ messages in thread
From: Johan Hovold @ 2021-07-07 15:55 UTC (permalink / raw)
To: Tamseel Shams
Cc: krzysztof.kozlowski, gregkh, jirislaby, linux-arm-kernel,
linux-samsung-soc, linux-serial, linux-kernel, alim.akhtar
On Tue, Jul 06, 2021 at 11:49:09AM +0530, Tamseel Shams wrote:
Please provide a better commit summary; "Checks the return value of
function" is too vague.
> "uart_add_one_port" function call may fail and return
> some error code, so adding a check for return value.
> If it is returning some error code, then displaying the
> result, unregistering the driver and then returning from
> probe function with error code.
>
> Signed-off-by: Tamseel Shams <m.shams@samsung.com>
> ---
> Changes since v1:
> 1. Added support to unregister driver on failure of "uart_add_one_port"
> function call.
> 2. Commit message updated.
>
> Changes since v2:
> 1. Added support to unwind clocks on failure of "uart_add_one_port"
> function call.
>
> drivers/tty/serial/samsung_tty.c | 17 ++++++++++++++++-
> 1 file changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
> index 9fbc61151c2e..a3f3a17fb54b 100644
> --- a/drivers/tty/serial/samsung_tty.c
> +++ b/drivers/tty/serial/samsung_tty.c
> @@ -2253,7 +2253,11 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
> }
>
> dev_dbg(&pdev->dev, "%s: adding port\n", __func__);
> - uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
> + ret = uart_add_one_port(&s3c24xx_uart_drv, &ourport->port);
> + if (ret < 0) {
> + dev_err(&pdev->dev, "Failed to add uart port, err %d\n", ret);
> + goto add_port_error;
> + }
> platform_set_drvdata(pdev, &ourport->port);
>
> /*
> @@ -2272,6 +2276,17 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
> probe_index++;
>
> return 0;
> +
> +add_port_error:
Name error labels after what they do, not where jump from (e.g.
"err_disable_clk").
> + ourport->port.mapbase = 0;
> + clk_disable_unprepare(ourport->clk);
> + clk_put(ourport->clk);
> + if (!IS_ERR(ourport->baudclk)) {
> + clk_disable_unprepare(ourport->baudclk);
> + clk_put(ourport->baudclk);
> + }
> + uart_unregister_driver(&s3c24xx_uart_drv);
You can't just deregister the serial driver if probe of a single port
fails. What if there are more than one port?
Looks like the driver has the same bug in remove(). What a mess... Added
by 6f134c3c7703 ("serial: samsung: Move uart_register_driver call to
device probe") in 2014.
And the clocks are never disabled and released in case
uart_register_driver() fails above either.
> + return ret;
> }
>
> static int s3c24xx_serial_remove(struct platform_device *dev)
Johan
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-07-07 15:55 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <CGME20210706061710epcas5p2c11d1bf5afb14774c4d4db93f2b83b33@epcas5p2.samsung.com>
2021-07-06 6:19 ` [PATCH v3] serial: samsung: Checks the return value of function Tamseel Shams
2021-07-07 8:14 ` Jiri Slaby
2021-07-07 15:55 ` Johan Hovold
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).