Re: [PATCH v2] firmware: dmi: handle missing DMI data gracefully

[Jean Delvare]

Hi Ard,

On Tue, 17 Oct 2017 19:15:37 +0100, Ard Biesheuvel wrote:
> Currently, when booting a kernel with DMI support on a platform that has
> no DMI tables, the following output is emitted into the kernel log:
> 
>   [    0.128818] DMI not present or invalid.
>   ...
>   [    1.306659] dmi: Firmware registration failed.
>   ...
>   [    2.908681] dmi-sysfs: dmi entry is absent.
> 
> The first one is a pr_info(), but the subsequent ones are pr_err()s that
> complain about a condition that is not really an error to begin with.
> 
> So let's clean this up, and give up silently if dma_available is not set.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> ---
> v2: - don't use dmi_available in dmi-sysfs because it is not exported,
>       instead, change pr_err to pr_debug
> 
>  drivers/firmware/dmi-sysfs.c | 2 +-
>  drivers/firmware/dmi_scan.c  | 6 ++----
>  2 files changed, 3 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/firmware/dmi-sysfs.c b/drivers/firmware/dmi-sysfs.c
> index d5de6ee8466d..ecf2eeb5f6f9 100644
> --- a/drivers/firmware/dmi-sysfs.c
> +++ b/drivers/firmware/dmi-sysfs.c
> @@ -652,7 +652,7 @@ static int __init dmi_sysfs_init(void)
>  	int val;
>  
>  	if (!dmi_kobj) {
> -		pr_err("dmi-sysfs: dmi entry is absent.\n");
> +		pr_debug("dmi-sysfs: dmi entry is absent.\n");
>  		error = -ENODATA;
>  		goto err;
>  	}
> diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c
> index 783041964439..17a7425063c2 100644
> --- a/drivers/firmware/dmi_scan.c
> +++ b/drivers/firmware/dmi_scan.c
> @@ -715,10 +715,8 @@ static int __init dmi_init(void)
>  	u8 *dmi_table;
>  	int ret = -ENOMEM;
>  
> -	if (!dmi_available) {
> -		ret = -ENODATA;
> -		goto err;
> -	}
> +	if (!dmi_available)
> +		return 0;
>  
>  	/*
>  	 * Set up dmi directory at /sys/firmware/dmi. This entry should stay

Looks good. Applied, thanks!

-- 
Jean Delvare
SUSE L3 Support