[PATCH resend] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[PATCH resend] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Hans de Goede]

Hi All,

This is a resend with Kai Heng Feng added to the Cc. Kai can you please
test this patch. This patch replaces yours, as many devices seem to be
affected so a whitelist is the better approach here.

Regards,

Hans

[PATCH] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Hans de Goede]

It seems that the WMI GUID used by the PEAQ 2-in-1 WMI hotkeys is not
as unique as a GUID should be and is used on some other devices too.

This is causing spurious key-press reports on these other devices.

This commits adds a DMI check to the PEAQ 2-in-1 WMI hotkeys driver to
ensure that it is actually running on a PEAQ 2-in-1, fixing the
spurious key-presses on these other devices.

BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1497861
BugLink: https://bugzilla.suse.com/attachment.cgi?id=743182
Cc: stable@vger.kernel.org
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 drivers/platform/x86/peaq-wmi.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/drivers/platform/x86/peaq-wmi.c b/drivers/platform/x86/peaq-wmi.c
index bc98ef95514a..67fa3fa32011 100644
--- a/drivers/platform/x86/peaq-wmi.c
+++ b/drivers/platform/x86/peaq-wmi.c
@@ -8,6 +8,7 @@
  */
 
 #include <linux/acpi.h>
+#include <linux/dmi.h>
 #include <linux/input-polldev.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
@@ -64,8 +65,22 @@ static void peaq_wmi_poll(struct input_polled_dev *dev)
 	}
 }
 
+/* Some other devices (Shuttle XS35) use the same WMI GUID for other purposes */
+static const struct dmi_system_id peaq_dmi_table[] = {
+	{
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "PEAQ"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "PEAQ PMM C1010 MD99187"),
+		},
+	},
+};
+
 static int __init peaq_wmi_init(void)
 {
+	/* WMI GUID is not unique, also check for a DMI match */
+	if (!dmi_check_system(peaq_dmi_table))
+		return -ENODEV;
+
 	if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
 		return -ENODEV;
 
@@ -86,6 +101,9 @@ static int __init peaq_wmi_init(void)
 
 static void __exit peaq_wmi_exit(void)
 {
+	if (!dmi_check_system(peaq_dmi_table))
+		return;
+
 	if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
 		return;
 
-- 
2.14.2

Re: [PATCH] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Andy Shevchenko]

On Thu, Oct 5, 2017 at 9:04 PM, Hans de Goede <hdegoede@redhat.com> wrote:
> It seems that the WMI GUID used by the PEAQ 2-in-1 WMI hotkeys is not
> as unique as a GUID should be and is used on some other devices too.
>
> This is causing spurious key-press reports on these other devices.
>
> This commits adds a DMI check to the PEAQ 2-in-1 WMI hotkeys driver to
> ensure that it is actually running on a PEAQ 2-in-1, fixing the
> spurious key-presses on these other devices.

Thanks!

One comment though.

>  static void __exit peaq_wmi_exit(void)
>  {
> +       if (!dmi_check_system(peaq_dmi_table))
> +               return;
> +
>         if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
>                 return;

I was thinking, after got kbuid bot complains on Kai's patch on
sections mismatch, do we need these checks at all?
How would be possible to get a module loaded in the first place if
system is not in whitelist?

If I'm right, please, drop the check and move table to __initconst.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Darren Hart]

On Thu, Oct 05, 2017 at 09:19:23PM +0300, Andy Shevchenko wrote:
> On Thu, Oct 5, 2017 at 9:04 PM, Hans de Goede <hdegoede@redhat.com> wrote:
> > It seems that the WMI GUID used by the PEAQ 2-in-1 WMI hotkeys is not
> > as unique as a GUID should be and is used on some other devices too.
> >
> > This is causing spurious key-press reports on these other devices.
> >
> > This commits adds a DMI check to the PEAQ 2-in-1 WMI hotkeys driver to
> > ensure that it is actually running on a PEAQ 2-in-1, fixing the
> > spurious key-presses on these other devices.
> 
> Thanks!
> 
> One comment though.
> 
> >  static void __exit peaq_wmi_exit(void)
> >  {
> > +       if (!dmi_check_system(peaq_dmi_table))
> > +               return;
> > +
> >         if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
> >                 return;
> 
> I was thinking, after got kbuid bot complains on Kai's patch on
> sections mismatch, do we need these checks at all?
> How would be possible to get a module loaded in the first place if
> system is not in whitelist?
> 

I was wondering this myself.

-- 
Darren Hart
VMware Open Source Technology Center

Re: [PATCH] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Andy Shevchenko]

On Thu, Oct 5, 2017 at 9:04 PM, Hans de Goede <hdegoede@redhat.com> wrote:
> It seems that the WMI GUID used by the PEAQ 2-in-1 WMI hotkeys is not
> as unique as a GUID should be and is used on some other devices too.
>
> This is causing spurious key-press reports on these other devices.
>
> This commits adds a DMI check to the PEAQ 2-in-1 WMI hotkeys driver to
> ensure that it is actually running on a PEAQ 2-in-1, fixing the
> spurious key-presses on these other devices.
>
> BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1497861
> BugLink: https://bugzilla.suse.com/attachment.cgi?id=743182
> Cc: stable@vger.kernel.org

So, I have pushed to my review queue as is, though please consider my
comments in perhaps a separate patch.
Thanks!

> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
> ---
>  drivers/platform/x86/peaq-wmi.c | 18 ++++++++++++++++++
>  1 file changed, 18 insertions(+)
>
> diff --git a/drivers/platform/x86/peaq-wmi.c b/drivers/platform/x86/peaq-wmi.c
> index bc98ef95514a..67fa3fa32011 100644
> --- a/drivers/platform/x86/peaq-wmi.c
> +++ b/drivers/platform/x86/peaq-wmi.c
> @@ -8,6 +8,7 @@
>   */
>
>  #include <linux/acpi.h>
> +#include <linux/dmi.h>
>  #include <linux/input-polldev.h>
>  #include <linux/kernel.h>
>  #include <linux/module.h>
> @@ -64,8 +65,22 @@ static void peaq_wmi_poll(struct input_polled_dev *dev)
>         }
>  }
>
> +/* Some other devices (Shuttle XS35) use the same WMI GUID for other purposes */
> +static const struct dmi_system_id peaq_dmi_table[] = {
> +       {
> +               .matches = {
> +                       DMI_MATCH(DMI_SYS_VENDOR, "PEAQ"),
> +                       DMI_MATCH(DMI_PRODUCT_NAME, "PEAQ PMM C1010 MD99187"),
> +               },
> +       },
> +};
> +
>  static int __init peaq_wmi_init(void)
>  {
> +       /* WMI GUID is not unique, also check for a DMI match */
> +       if (!dmi_check_system(peaq_dmi_table))
> +               return -ENODEV;
> +
>         if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
>                 return -ENODEV;
>
> @@ -86,6 +101,9 @@ static int __init peaq_wmi_init(void)
>
>  static void __exit peaq_wmi_exit(void)
>  {
> +       if (!dmi_check_system(peaq_dmi_table))
> +               return;
> +
>         if (!wmi_has_guid(PEAQ_DOLBY_BUTTON_GUID))
>                 return;
>
> --
> 2.14.2
>



-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH resend] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Kai-Heng Feng]

Hi Hans,

On Fri, Oct 6, 2017 at 2:04 AM, Hans de Goede <hdegoede@redhat.com> wrote:
> Hi All,
>
> This is a resend with Kai Heng Feng added to the Cc. Kai can you please
> test this patch. This patch replaces yours, as many devices seem to be
> affected so a whitelist is the better approach here.

I'll let the affected user to test the patch.

Thanks!

>
> Regards,
>
> Hans

Re: [PATCH resend] platform/x86: peaq-wmi: Add DMI check before binding to the WMI interface

[Kai-Heng Feng]

Hi,

On Fri, Oct 6, 2017 at 12:23 PM, Kai-Heng Feng
<kai.heng.feng@canonical.com> wrote:
> Hi Hans,
>
> On Fri, Oct 6, 2017 at 2:04 AM, Hans de Goede <hdegoede@redhat.com> wrote:
>> Hi All,
>>
>> This is a resend with Kai Heng Feng added to the Cc. Kai can you please
>> test this patch. This patch replaces yours, as many devices seem to be
>> affected so a whitelist is the better approach here.
>
> I'll let the affected user to test the patch.

The user reports the new patch works.

Thanks.

>
> Thanks!
>
>>
>> Regards,
>>
>> Hans