netdev - Re: [linux-nfc] [PATCH] NFC: cooperation with runtime PM

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 linux-hardening linux-cve-announce 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]

Message-ID: <0a8ca4c7-ce55-3c92-cc29-b383e546d563@canonical.com>
Date:   Wed, 12 May 2021 09:54:55 -0400
From:   Krzysztof Kozlowski <krzysztof.kozlowski@...onical.com>
To:     Oliver Neukum <oneukum@...e.com>, clement.perrochaud@...innov.com,
        charles.gorand@...innov.com, linux-nfc@...ts.01.org,
        netdev@...r.kernel.org
Subject: Re: [linux-nfc] [PATCH] NFC: cooperation with runtime PM

Hi,

Thanks for the patch. Few notes:

On 12/05/2021 09:44, Oliver Neukum wrote:
> We cannot rely on the underlying hardware to do correct
> runtime PM. NFC core needs to get PM reference while
> a device is operational, lest it be suspended when
> it is supposed to be waiting for a target to come
> into range.


Your word wrapping is unusually early - please wrap the commit msg as in
coding style (so around 75-character).
https://elixir.bootlin.com/linux/latest/source/Documentation/process/submitting-patches.rst#L578

> 
> Signed-off-by: Oliver Neukum <oneukum@...e.com>
> ---
>  net/nfc/core.c | 19 ++++++++++++++++++-
>  1 file changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/net/nfc/core.c b/net/nfc/core.c
> index 573c80c6ff7a..5ca4597c39c7 100644
> --- a/net/nfc/core.c
> +++ b/net/nfc/core.c
> @@ -15,6 +15,7 @@
>  #include <linux/slab.h>
>  #include <linux/rfkill.h>
>  #include <linux/nfc.h>
> +#include <linux/pm_runtime.h>
>  
>  #include <net/genetlink.h>
>  
> @@ -37,6 +38,7 @@ int nfc_fw_download(struct nfc_dev *dev, const char *firmware_name)
>  	pr_debug("%s do firmware %s\n", dev_name(&dev->dev), firmware_name);
>  
>  	device_lock(&dev->dev);
> +	pm_runtime_get_sync(&dev->dev);

This can fail. Probably you wanted pm_runtime_resume_and_get() here.

>  
>  	if (!device_is_registered(&dev->dev)) {
>  		rc = -ENODEV;
> @@ -58,7 +60,10 @@ int nfc_fw_download(struct nfc_dev *dev, const char *firmware_name)
>  	if (rc)
>  		dev->fw_download_in_progress = false;

goto error

>  
> +	device_unlock(&dev->dev);
> +	return rc;

Since last rc cannot be != 0, return 0

Blank line

>  error:
> +	pm_runtime_put(&dev->dev);
>  	device_unlock(&dev->dev);
>  	return rc;
>  }
> @@ -73,9 +78,13 @@ int nfc_fw_download(struct nfc_dev *dev, const char *firmware_name)
>  int nfc_fw_download_done(struct nfc_dev *dev, const char *firmware_name,
>  			 u32 result)
>  {
> +	int rv;

"int rc"

> +
>  	dev->fw_download_in_progress = false;
>  
> -	return nfc_genl_fw_download_done(dev, firmware_name, result);
> +	rv = nfc_genl_fw_download_done(dev, firmware_name, result);
> +	pm_runtime_put(&dev->dev);
> +	return rv;
>  }
>  EXPORT_SYMBOL(nfc_fw_download_done);
>  
> @@ -93,6 +102,7 @@ int nfc_dev_up(struct nfc_dev *dev)
>  	pr_debug("dev_name=%s\n", dev_name(&dev->dev));
>  
>  	device_lock(&dev->dev);
> +	pm_runtime_get_sync(&dev->dev);

Same comments as before.

>  
>  	if (dev->rfkill && rfkill_blocked(dev->rfkill)) {
>  		rc = -ERFKILL;
> @@ -124,7 +134,11 @@ int nfc_dev_up(struct nfc_dev *dev)
>  	if (dev->ops->discover_se && dev->ops->discover_se(dev))
>  		pr_err("SE discovery failed\n");
>  
> +	device_unlock(&dev->dev);
> +	return rc;

Probably same comments as before apply.

> +
>  error:
> +	pm_runtime_put(&dev->dev);
>  	device_unlock(&dev->dev);
>  	return rc;
>  }
> @@ -161,6 +175,9 @@ int nfc_dev_down(struct nfc_dev *dev)
>  		dev->ops->dev_down(dev);
>  
>  	dev->dev_up = false;
> +	pm_runtime_put(&dev->dev);
> +	device_unlock(&dev->dev);
> +	return rc;

return 0

>  
>  error:
>  	device_unlock(&dev->dev);
> 


Best regards,
Krzysztof