[PATCH 1/1] ACPI: bus.c, fix error handling in acpi_bus_init

[PATCH 1/1] ACPI: bus.c, fix error handling in acpi_bus_init

[Jiri Slaby]

There was a misplaced status test. Move it to correct place and
rollback appropriately.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
---
 drivers/acpi/bus.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
index 765fd1c..79efef6 100644
--- a/drivers/acpi/bus.c
+++ b/drivers/acpi/bus.c
@@ -760,18 +760,17 @@ static int __init acpi_bus_init(void)
 
 
 	status = acpi_os_initialize1();
-
-	status =
-	    acpi_enable_subsystem(ACPI_NO_HARDWARE_INIT | ACPI_NO_ACPI_ENABLE);
 	if (ACPI_FAILURE(status)) {
 		printk(KERN_ERR PREFIX
-		       "Unable to start the ACPI Interpreter\n");
-		goto error1;
+		       "Unable to initialize ACPI OS objects\n");
+		goto error0;
 	}
 
+	status =
+	    acpi_enable_subsystem(ACPI_NO_HARDWARE_INIT | ACPI_NO_ACPI_ENABLE);
 	if (ACPI_FAILURE(status)) {
 		printk(KERN_ERR PREFIX
-		       "Unable to initialize ACPI OS objects\n");
+		       "Unable to start the ACPI Interpreter\n");
 		goto error1;
 	}
 
@@ -832,6 +831,7 @@ static int __init acpi_bus_init(void)
 	/* Mimic structured exception handling */
       error1:
 	acpi_terminate();
+error0:
 	return -ENODEV;
 }
 
-- 
1.6.1.3

Re: [PATCH 1/1] ACPI: bus.c, fix error handling in acpi_bus_init

[Len Brown]

On Sun, 15 Feb 2009, Jiri Slaby wrote:

> There was a misplaced status test. Move it to correct place and
> rollback appropriately.

hmm, looks like this has been broken since the day
acpi_os_initialize1() was invented in 2004.

Turns out that the bug and its fix are moot, however,
as acpi_os_initialize1() is hard-coded to return success,
opting for BUG_ON() if it sees a failure...

So unless that changes, I'd prefer to keep the code
simple and just not check its status -- which is effectively
what we're doing right now.

thanks,
-Len

> Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
> ---
>  drivers/acpi/bus.c |   12 ++++++------
>  1 files changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
> index 765fd1c..79efef6 100644
> --- a/drivers/acpi/bus.c
> +++ b/drivers/acpi/bus.c
> @@ -760,18 +760,17 @@ static int __init acpi_bus_init(void)
>  
>  
>  	status = acpi_os_initialize1();
> -
> -	status =
> -	    acpi_enable_subsystem(ACPI_NO_HARDWARE_INIT | ACPI_NO_ACPI_ENABLE);
>  	if (ACPI_FAILURE(status)) {
>  		printk(KERN_ERR PREFIX
> -		       "Unable to start the ACPI Interpreter\n");
> -		goto error1;
> +		       "Unable to initialize ACPI OS objects\n");
> +		goto error0;
>  	}
>  
> +	status =
> +	    acpi_enable_subsystem(ACPI_NO_HARDWARE_INIT | ACPI_NO_ACPI_ENABLE);
>  	if (ACPI_FAILURE(status)) {
>  		printk(KERN_ERR PREFIX
> -		       "Unable to initialize ACPI OS objects\n");
> +		       "Unable to start the ACPI Interpreter\n");
>  		goto error1;
>  	}
>  
> @@ -832,6 +831,7 @@ static int __init acpi_bus_init(void)
>  	/* Mimic structured exception handling */
>        error1:
>  	acpi_terminate();
> +error0:
>  	return -ENODEV;
>  }
>  
> -- 
> 1.6.1.3
> 

Re: [PATCH 1/1] ACPI: bus.c, fix error handling in acpi_bus_init

[Jiri Slaby]

On 21.2.2009 05:58, Len Brown wrote:
> On Sun, 15 Feb 2009, Jiri Slaby wrote:
>
>> There was a misplaced status test. Move it to correct place and
>> rollback appropriately.
>
> hmm, looks like this has been broken since the day
> acpi_os_initialize1() was invented in 2004.

It comes from bitkeeper times, I checked, it looked like a typo.

> Turns out that the bug and its fix are moot, however,
> as acpi_os_initialize1() is hard-coded to return success,
> opting for BUG_ON() if it sees a failure...

BUG_ON is not noreturn e.g. on embedded, anyway it would crash later in 
this case. I was thinking about returning a failure when one of them is 
NULL, but fell back to not do so and check only retval in the caller.

> So unless that changes, I'd prefer to keep the code
> simple and just not check its status -- which is effectively
> what we're doing right now.

Ok, no problem. Will repost.

Thanks.