Re: [PATCH 15/18] bus: mhi: core: Introduce sysfs entries for MHI

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Sep 21, 2020 at 09:38:12PM +0530, Manivannan Sadhasivam wrote:
> From: Bhaumik Bhatt <bbhatt@codeaurora.org>
> 
> Introduce sysfs entries to enable userspace clients the ability to read
> the serial number and the OEM PK Hash values obtained from BHI. OEMs
> need to read these device-specific hardware information values through
> userspace for factory testing purposes and cannot be exposed via degbufs
> as it may remain disabled for performance reasons. Also, update the
> documentation for ABI to include these entries.
> 
> Signed-off-by: Bhaumik Bhatt <bbhatt@codeaurora.org>
> Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
> Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
> ---
>  Documentation/ABI/stable/sysfs-bus-mhi | 21 ++++++++++
>  MAINTAINERS                            |  1 +
>  drivers/bus/mhi/core/init.c            | 53 ++++++++++++++++++++++++++
>  3 files changed, 75 insertions(+)
>  create mode 100644 Documentation/ABI/stable/sysfs-bus-mhi
> 
> diff --git a/Documentation/ABI/stable/sysfs-bus-mhi b/Documentation/ABI/stable/sysfs-bus-mhi
> new file mode 100644
> index 000000000000..ecfe7662f8d0
> --- /dev/null
> +++ b/Documentation/ABI/stable/sysfs-bus-mhi
> @@ -0,0 +1,21 @@
> +What:		/sys/bus/mhi/devices/.../serialnumber
> +Date:		Sept 2020
> +KernelVersion:	5.10
> +Contact:	Bhaumik Bhatt <bbhatt@codeaurora.org>
> +Description:	The file holds the serial number of the client device obtained
> +		using a BHI (Boot Host Interface) register read after at least
> +		one attempt to power up the device has been done. If read
> +		without having the device power on at least once, the file will
> +		read all 0's.
> +Users:		Any userspace application or clients interested in device info.
> +
> +What:		/sys/bus/mhi/devices/.../oem_pk_hash
> +Date:		Sept 2020
> +KernelVersion:	5.10
> +Contact:	Bhaumik Bhatt <bbhatt@codeaurora.org>
> +Description:	The file holds the OEM PK Hash value of the endpoint device
> +		obtained using a BHI (Boot Host Interface) register read after
> +		at least one attempt to power up the device has been done. If
> +		read without having the device power on at least once, the file
> +		will read all 0's.
> +Users:		Any userspace application or clients interested in device info.
> diff --git a/MAINTAINERS b/MAINTAINERS
> index deaafb617361..11e7be9b9163 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -11323,6 +11323,7 @@ M:	Hemant Kumar <hemantk@codeaurora.org>
>  L:	linux-arm-msm@vger.kernel.org
>  S:	Maintained
>  T:	git git://git.kernel.org/pub/scm/linux/kernel/git/mani/mhi.git
> +F:	Documentation/ABI/stable/sysfs-bus-mhi
>  F:	Documentation/mhi/
>  F:	drivers/bus/mhi/
>  F:	include/linux/mhi.h
> diff --git a/drivers/bus/mhi/core/init.c b/drivers/bus/mhi/core/init.c
> index 61e5885a331a..1b4161eaf0d8 100644
> --- a/drivers/bus/mhi/core/init.c
> +++ b/drivers/bus/mhi/core/init.c
> @@ -76,6 +76,56 @@ const char *to_mhi_pm_state_str(enum mhi_pm_state state)
>  	return mhi_pm_state_str[index];
>  }
>  
> +static ssize_t serial_number_show(struct device *dev,
> +				  struct device_attribute *attr,
> +				  char *buf)
> +{
> +	struct mhi_device *mhi_dev = to_mhi_device(dev);
> +	struct mhi_controller *mhi_cntrl = mhi_dev->mhi_cntrl;
> +
> +	return snprintf(buf, PAGE_SIZE, "Serial Number: %u\n",
> +			mhi_cntrl->serial_number);
> +}
> +static DEVICE_ATTR_RO(serial_number);
> +
> +static ssize_t oem_pk_hash_show(struct device *dev,
> +				struct device_attribute *attr,
> +				char *buf)
> +{
> +	struct mhi_device *mhi_dev = to_mhi_device(dev);
> +	struct mhi_controller *mhi_cntrl = mhi_dev->mhi_cntrl;
> +	int i, cnt = 0;
> +
> +	for (i = 0; i < ARRAY_SIZE(mhi_cntrl->oem_pk_hash); i++)
> +		cnt += snprintf(buf + cnt, PAGE_SIZE - cnt,
> +				"OEMPKHASH[%d]: 0x%x\n", i,
> +				mhi_cntrl->oem_pk_hash[i]);
> +
> +	return cnt;
> +}
> +static DEVICE_ATTR_RO(oem_pk_hash);
> +
> +static struct attribute *mhi_sysfs_attrs[] = {
> +	&dev_attr_serial_number.attr,
> +	&dev_attr_oem_pk_hash.attr,
> +	NULL,
> +};
> +
> +static const struct attribute_group mhi_sysfs_group = {
> +	.attrs = mhi_sysfs_attrs,
> +};
> +
> +static int mhi_create_sysfs(struct mhi_controller *mhi_cntrl)
> +{
> +	return sysfs_create_group(&mhi_cntrl->mhi_dev->dev.kobj,
> +				  &mhi_sysfs_group);

You should never have to call a sysfs_* function from a driver or bus
code, that implies something is wrong :)

Just set the sysfs attributes to be the default groups for your
bus/device/whatever-you-have-here and then the driver core will properly
manage the creation and removal of these files, in a race-free manner.

As it is, I think this is racy and will cause userspace confusion, but I
haven't looked for sure, as you should use the api that guarantees it is
safe...

thanks,

greg k-h