Re: cosmetic udev glitches
On 03/03/2010 08:25 PM, Dan Nicholson wrote: > On Wed, Mar 3, 2010 at 4:59 PM, Thomas Jaeger wrote: >> On 03/03/2010 07:48 PM, Dan Nicholson wrote: >>> On Wed, Mar 3, 2010 at 4:00 PM, Thomas Jaeger wrote: The attribute 'name' might not exist (as in the case of serial wacom devices) and it's impossible to set it in an udev rule, I believe, so we should at least fall back to the NAME property if the attribute is not available. >>> >>> Not for the parent, either? Does hal give you anything for >>> info.product/input.product? >> >> Sorry, I don't know, I don't have a working hal installation anymore. >> Here's what udev knows about the device, though: >> >> udevadm info --path=/sys/class/tty/ttyS0/ --attribute-walk >> >> ... > > But you do get NAME in the uevent? You can find this from "udevadm > info --export-db". Well, sort of. I set NAME in my own udev rule where I filter by pnp id. > Seems reasonable, though. I guess we need an additional: > > if (!name) > name = udev_device_get_property_value(parent, "NAME"); > Yup, that's what I thought, too. Tom ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Wed, Mar 3, 2010 at 4:59 PM, Thomas Jaeger wrote:
> On 03/03/2010 07:48 PM, Dan Nicholson wrote:
>> On Wed, Mar 3, 2010 at 4:00 PM, Thomas Jaeger wrote:
>>> The attribute 'name' might not exist (as in the case of serial wacom
>>> devices) and it's impossible to set it in an udev rule, I believe, so we
>>> should at least fall back to the NAME property if the attribute is not
>>> available.
>>
>> Not for the parent, either? Does hal give you anything for
>> info.product/input.product?
>
> Sorry, I don't know, I don't have a working hal installation anymore.
> Here's what udev knows about the device, though:
>
> udevadm info --path=/sys/class/tty/ttyS0/ --attribute-walk
>
> ...
>
> looking at device '/devices/pnp0/00:0a/tty/ttyS0':
> KERNEL=="ttyS0"
> SUBSYSTEM=="tty"
> DRIVER==""
>
> looking at parent device '/devices/pnp0/00:0a':
> KERNELS=="00:0a"
> SUBSYSTEMS=="pnp"
> DRIVERS=="serial"
> ATTRS{id}=="WACf004"
>
> looking at parent device '/devices/pnp0':
> KERNELS=="pnp0"
> SUBSYSTEMS==""
> DRIVERS==""
But you do get NAME in the uevent? You can find this from "udevadm
info --export-db".
Seems reasonable, though. I guess we need an additional:
if (!name)
name = udev_device_get_property_value(parent, "NAME");
--
Dan
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On 03/03/2010 07:48 PM, Dan Nicholson wrote:
> On Wed, Mar 3, 2010 at 4:00 PM, Thomas Jaeger wrote:
>> The attribute 'name' might not exist (as in the case of serial wacom
>> devices) and it's impossible to set it in an udev rule, I believe, so we
>> should at least fall back to the NAME property if the attribute is not
>> available.
>
> Not for the parent, either? Does hal give you anything for
> info.product/input.product?
Sorry, I don't know, I don't have a working hal installation anymore.
Here's what udev knows about the device, though:
udevadm info --path=/sys/class/tty/ttyS0/ --attribute-walk
...
looking at device '/devices/pnp0/00:0a/tty/ttyS0':
KERNEL=="ttyS0"
SUBSYSTEM=="tty"
DRIVER==""
looking at parent device '/devices/pnp0/00:0a':
KERNELS=="00:0a"
SUBSYSTEMS=="pnp"
DRIVERS=="serial"
ATTRS{id}=="WACf004"
looking at parent device '/devices/pnp0':
KERNELS=="pnp0"
SUBSYSTEMS==""
DRIVERS==""
Tom
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Wed, Mar 3, 2010 at 4:00 PM, Thomas Jaeger wrote: > The attribute 'name' might not exist (as in the case of serial wacom > devices) and it's impossible to set it in an udev rule, I believe, so we > should at least fall back to the NAME property if the attribute is not > available. Not for the parent, either? Does hal give you anything for info.product/input.product? -- Dan ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
The attribute 'name' might not exist (as in the case of serial wacom
devices) and it's impossible to set it in an udev rule, I believe, so we
should at least fall back to the NAME property if the attribute is not
available.
Tom
On 03/02/2010 05:47 PM, Dan Nicholson wrote:
> On Tue, Mar 02, 2010 at 05:57:04PM +0100, Simon Thum wrote:
>
> Can you try the patch below that uses the name attribute instead of the
> value from the uevent? I ran it through a little test app, but not in X.
>
> Dan
>
> From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001
> From: Dan Nicholson
> Date: Tue, 2 Mar 2010 14:42:22 -0800
> Subject: [PATCH] config/udev: Get product name from attribute instead of
> uevent
>
> The input device product name in the kernel uevent has embedded quotes
> that aren't expected here. Use the sysfs name attribute instead, which
> does not suffer this problem.
>
> Signed-off-by: Dan Nicholson
> ---
> config/udev.c |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/config/udev.c b/config/udev.c
> index 69257e9..54306ef 100644
> --- a/config/udev.c
> +++ b/config/udev.c
> @@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device)
>
> parent = udev_device_get_parent(udev_device);
> if (parent)
> -name = udev_device_get_property_value(parent, "NAME");
> +name = udev_device_get_sysattr_value(parent, "name");
> if (!name)
> name = "(unnamed)";
> else
>From 8e791dbd4ad96c64802e864ff100f6342734a40c Mon Sep 17 00:00:00 2001
From: Dan Nicholson
Date: Tue, 2 Mar 2010 14:42:22 -0800
Subject: [PATCH] config/udev: Get product name from attribute instead of uevent
The input device product name in the kernel uevent has embedded quotes
that aren't expected here. Use the sysfs name attribute instead, which
does not suffer this problem. Fall back to the NAME property if the
attribute is not available.
Signed-off-by: Thomas Jaeger
---
config/udev.c |7 +--
1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/config/udev.c b/config/udev.c
index 69257e9..452fb5a 100644
--- a/config/udev.c
+++ b/config/udev.c
@@ -73,8 +73,11 @@ device_added(struct udev_device *udev_device)
goto unwind;
parent = udev_device_get_parent(udev_device);
-if (parent)
-name = udev_device_get_property_value(parent, "NAME");
+if (parent) {
+name = udev_device_get_sysattr_value(parent, "name");
+if (!name)
+name = udev_device_get_property_value(parent, "NAME");
+}
if (!name)
name = "(unnamed)";
else
--
1.7.0
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 2, 2010 at 14:47:22 -0800, Dan Nicholson wrote: > From: Dan Nicholson > Date: Tue, 2 Mar 2010 14:42:22 -0800 > Subject: [PATCH] config/udev: Get product name from attribute instead of > uevent > > The input device product name in the kernel uevent has embedded quotes > that aren't expected here. Use the sysfs name attribute instead, which > does not suffer this problem. > > Signed-off-by: Dan Nicholson Acked-by: Julien Cristau Thanks, Dan. Cheers, Julien ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
Julien Cristau wrote:
> This is also expected at the moment. The enumeration of devices gives
> us /dev/input/event* (which evdev can handle), and
> /dev/input/{mouse,kbd}* (which it doesn't). This results in a couple
> error messages for the latter. It should be possible to tone those down
> so users don't get scared…
Thanks for the clarification! I applied Peter's MatchDevicePath
"/dev/input/event*" which is probably better than making polite drivers.
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Wed, Mar 03, 2010 at 10:40:16AM +1000, Peter Hutterer wrote:
> On Tue, Mar 02, 2010 at 04:29:41PM -0800, Dan Nicholson wrote:
> > On Tue, Mar 2, 2010 at 4:16 PM, Peter Hutterer
> > wrote:
> > > On Tue, Mar 02, 2010 at 02:47:22PM -0800, Dan Nicholson wrote:
> > >> Can you try the patch below that uses the name attribute instead of the
> > >> value from the uevent? I ran it through a little test app, but not in X.
> > >>
> > >> From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001
> > >> From: Dan Nicholson
> > >> Date: Tue, 2 Mar 2010 14:42:22 -0800
> > >> Subject: [PATCH] config/udev: Get product name from attribute instead of
> > >> uevent
> > >>
> > >> The input device product name in the kernel uevent has embedded quotes
> > >> that aren't expected here. Use the sysfs name attribute instead, which
> > >> does not suffer this problem.
> > >>
> > >> Signed-off-by: Dan Nicholson
> > >> ---
> > >> config/udev.c | 2 +-
> > >> 1 files changed, 1 insertions(+), 1 deletions(-)
> > >>
> > >> diff --git a/config/udev.c b/config/udev.c
> > >> index 69257e9..54306ef 100644
> > >> --- a/config/udev.c
> > >> +++ b/config/udev.c
> > >> @@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device)
> > >>
> > >> parent = udev_device_get_parent(udev_device);
> > >> if (parent)
> > >> - name = udev_device_get_property_value(parent, "NAME");
> > >> + name = udev_device_get_sysattr_value(parent, "name");
> > >> if (!name)
> > >> name = "(unnamed)";
> > >> else
> > >> --
> > >> 1.6.6.1
> > >
> > > Reviewed-by: Peter Hutterer
> > >
> > > tested too.
> >
> > Is this on rawhide with udev-151? When I tried it against the udev on
> > f12, it was replacing the spaces in the name with underscores. I can't
> > figure out why, but when I ran it against a vanilla udev 145 or git,
> > it worked fine.
>
> This was on F12 (X server from git) and I don't see the underscore issue you
> mentioned. Updates-testing from yesterday.
> Not quite sure where that comes from tbh.
Never mind. I don't know what I doing, but it's working now. Here's my
little test app for reference. You have to pass it the sysfs device path.
--
Dan
#include
#include
#include
#define LIBUDEV_I_KNOW_THE_API_IS_SUBJECT_TO_CHANGE 1
#include
int main(int argc, char *argv[])
{
struct udev *udev;
struct udev_device *device, *parent;
const char *path, *name;
if (argc < 2) {
fprintf(stderr, "No device path specified\n");
exit(EXIT_FAILURE);
}
path = argv[1];
udev = udev_new();
device = udev_device_new_from_syspath(udev, path);
if (!device) {
fprintf(stderr, "Device not found from path \"%s\"\n", path);
exit(EXIT_FAILURE);
}
parent = udev_device_get_parent(device);
if (!parent)
parent = device;
name = udev_device_get_sysattr_value(parent, "name");
printf("%s: \"%s\"\n", path, name);
udev_device_unref(device);
udev_unref(udev);
return 0;
}
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 02, 2010 at 04:29:41PM -0800, Dan Nicholson wrote: > On Tue, Mar 2, 2010 at 4:16 PM, Peter Hutterer > wrote: > > On Tue, Mar 02, 2010 at 02:47:22PM -0800, Dan Nicholson wrote: > >> Can you try the patch below that uses the name attribute instead of the > >> value from the uevent? I ran it through a little test app, but not in X. > >> > >> From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001 > >> From: Dan Nicholson > >> Date: Tue, 2 Mar 2010 14:42:22 -0800 > >> Subject: [PATCH] config/udev: Get product name from attribute instead of > >> uevent > >> > >> The input device product name in the kernel uevent has embedded quotes > >> that aren't expected here. Use the sysfs name attribute instead, which > >> does not suffer this problem. > >> > >> Signed-off-by: Dan Nicholson > >> --- > >> config/udev.c | 2 +- > >> 1 files changed, 1 insertions(+), 1 deletions(-) > >> > >> diff --git a/config/udev.c b/config/udev.c > >> index 69257e9..54306ef 100644 > >> --- a/config/udev.c > >> +++ b/config/udev.c > >> @@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device) > >> > >> parent = udev_device_get_parent(udev_device); > >> if (parent) > >> - name = udev_device_get_property_value(parent, "NAME"); > >> + name = udev_device_get_sysattr_value(parent, "name"); > >> if (!name) > >> name = "(unnamed)"; > >> else > >> -- > >> 1.6.6.1 > > > > Reviewed-by: Peter Hutterer > > > > tested too. > > Is this on rawhide with udev-151? When I tried it against the udev on > f12, it was replacing the spaces in the name with underscores. I can't > figure out why, but when I ran it against a vanilla udev 145 or git, > it worked fine. This was on F12 (X server from git) and I don't see the underscore issue you mentioned. Updates-testing from yesterday. Not quite sure where that comes from tbh. Cheers, Peter ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 2, 2010 at 4:16 PM, Peter Hutterer wrote: > On Tue, Mar 02, 2010 at 02:47:22PM -0800, Dan Nicholson wrote: >> Can you try the patch below that uses the name attribute instead of the >> value from the uevent? I ran it through a little test app, but not in X. >> >> From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001 >> From: Dan Nicholson >> Date: Tue, 2 Mar 2010 14:42:22 -0800 >> Subject: [PATCH] config/udev: Get product name from attribute instead of >> uevent >> >> The input device product name in the kernel uevent has embedded quotes >> that aren't expected here. Use the sysfs name attribute instead, which >> does not suffer this problem. >> >> Signed-off-by: Dan Nicholson >> --- >> config/udev.c | 2 +- >> 1 files changed, 1 insertions(+), 1 deletions(-) >> >> diff --git a/config/udev.c b/config/udev.c >> index 69257e9..54306ef 100644 >> --- a/config/udev.c >> +++ b/config/udev.c >> @@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device) >> >> parent = udev_device_get_parent(udev_device); >> if (parent) >> - name = udev_device_get_property_value(parent, "NAME"); >> + name = udev_device_get_sysattr_value(parent, "name"); >> if (!name) >> name = "(unnamed)"; >> else >> -- >> 1.6.6.1 > > Reviewed-by: Peter Hutterer > > tested too. Is this on rawhide with udev-151? When I tried it against the udev on f12, it was replacing the spaces in the name with underscores. I can't figure out why, but when I ran it against a vanilla udev 145 or git, it worked fine. -- Dan ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 02, 2010 at 02:47:22PM -0800, Dan Nicholson wrote: > Can you try the patch below that uses the name attribute instead of the > value from the uevent? I ran it through a little test app, but not in X. > > From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001 > From: Dan Nicholson > Date: Tue, 2 Mar 2010 14:42:22 -0800 > Subject: [PATCH] config/udev: Get product name from attribute instead of > uevent > > The input device product name in the kernel uevent has embedded quotes > that aren't expected here. Use the sysfs name attribute instead, which > does not suffer this problem. > > Signed-off-by: Dan Nicholson > --- > config/udev.c |2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/config/udev.c b/config/udev.c > index 69257e9..54306ef 100644 > --- a/config/udev.c > +++ b/config/udev.c > @@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device) > > parent = udev_device_get_parent(udev_device); > if (parent) > -name = udev_device_get_property_value(parent, "NAME"); > +name = udev_device_get_sysattr_value(parent, "name"); > if (!name) > name = "(unnamed)"; > else > -- > 1.6.6.1 Reviewed-by: Peter Hutterer tested too. Cheers, Peter ___ xorg mailing list xorg@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 02, 2010 at 05:57:04PM +0100, Simon Thum wrote:
> Hi all,
>
> I just upgraded my X to use udev. It generally went fine, but I noticed
> two minor issues:
>
> At first, the device names have quotation marks. As in
>
> xinput list
> ...
> ⎜ "ImExPS/2 Logitech Wheel Mouse" id=7[slave pointer
> (2)]
>
> I looked up the source, but found nothing obvious. Is this expected
> behavior?
>
> Second, udev seems to add my mouse (and others) twice, resulting in some
> log spam. Excerpt:
>
> [ 5571.537] (II) config/udev: Adding input device "ImExPS/2 Logitech
> Wheel Mouse" (/dev/input/event3)
> (...)
> [ 5571.538] (II) XINPUT: Adding extended input device ""ImExPS/2
> Logitech Wheel Mouse"" (type: MOUSE)
> [ 5571.538] (II) "ImExPS/2 Logitech Wheel Mouse": initialized for
> relative axes.
> [ 5571.538] (II) config/udev: Adding input device "ImExPS/2 Logitech
> Wheel Mouse" (/dev/input/mouse0)
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Applying InputClass
> "evdev catchall"
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": always reports core
> events
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Device:
> "/dev/input/mouse0"
> [ 5571.539] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device
> [ 5571.539] (II) UnloadModule: "evdev"
> [ 5571.539] (EE) PreInit returned NULL for ""ImExPS/2 Logitech Wheel
> Mouse""
>
>
> Things work fine here, I just wanted to make sure whether my
> configuration is broken or these are known issues.
>
> In case it helps, I did:
> udevadm info --attribute-walk --name /dev/input/mouse0
>
> Udevadm info starts with the device specified by the devpath and then
> walks up the chain of parent devices. It prints for every device
> found, all possible attributes in the udev rules key format.
> A rule to match, can be composed by the attributes of the device
> and the attributes from one single parent device.
>
> looking at device '/devices/platform/i8042/serio1/input/input3/mouse0':
> KERNEL=="mouse0"
> SUBSYSTEM=="input"
> DRIVER==""
>
> looking at parent device '/devices/platform/i8042/serio1/input/input3':
> KERNELS=="input3"
> SUBSYSTEMS=="input"
> DRIVERS==""
> ATTRS{name}=="ImExPS/2 Logitech Wheel Mouse"
> ATTRS{phys}=="isa0060/serio1/input0"
> ATTRS{uniq}==""
Can you try the patch below that uses the name attribute instead of the
value from the uevent? I ran it through a little test app, but not in X.
Dan
From c01a4e6dce7196a60b5247dde78054b27e004494 Mon Sep 17 00:00:00 2001
From: Dan Nicholson
Date: Tue, 2 Mar 2010 14:42:22 -0800
Subject: [PATCH] config/udev: Get product name from attribute instead of uevent
The input device product name in the kernel uevent has embedded quotes
that aren't expected here. Use the sysfs name attribute instead, which
does not suffer this problem.
Signed-off-by: Dan Nicholson
---
config/udev.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/config/udev.c b/config/udev.c
index 69257e9..54306ef 100644
--- a/config/udev.c
+++ b/config/udev.c
@@ -74,7 +74,7 @@ device_added(struct udev_device *udev_device)
parent = udev_device_get_parent(udev_device);
if (parent)
-name = udev_device_get_property_value(parent, "NAME");
+name = udev_device_get_sysattr_value(parent, "name");
if (!name)
name = "(unnamed)";
else
--
1.6.6.1
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 02, 2010 at 01:01:53PM -0800, Dan Nicholson wrote:
> On Tue, Mar 2, 2010 at 9:28 AM, Julien Cristau wrote:
> > On Tue, Mar 2, 2010 at 17:57:04 +0100, Simon Thum wrote:
> >
> >> Hi all,
> >>
> >> I just upgraded my X to use udev. It generally went fine, but I noticed
> >> two minor issues:
> >>
> >> At first, the device names have quotation marks. As in
> >>
> >> xinput list
> >> ...
> >> ⎜ "ImExPS/2 Logitech Wheel Mouse" id=7 [slave pointer
> >> (2)]
> >>
> >> I looked up the source, but found nothing obvious. Is this expected
> >> behavior?
> >>
> > Yes, the device names we get from udev have quotations marks. I don't
> > know why, and I didn't feel like mangling them.
>
> It seems to come straight from the kernel (drivers/input/input.c),
> which adds the quotes around the value in NAME to the uevent. Maybe we
> should ask them not to do that. Looks like this was changed back in
> 2005, commit a7fadbe10ccf430e7a8add8b45c561d864087343.
>
> However, I think we should try to be compatible with info.product from
> hal. This would require some amount of mangling. What hal does is:
>
> 1. Get the name directly from the sysfs name file (no quotes there)
> 2. Strip trailing whitespace
>
> >> Second, udev seems to add my mouse (and others) twice, resulting in some
> >> log spam. Excerpt:
> >>
> >> [ 5571.537] (II) config/udev: Adding input device "ImExPS/2 Logitech
> >> Wheel Mouse" (/dev/input/event3)
> >> (...)
> >> [ 5571.538] (II) XINPUT: Adding extended input device ""ImExPS/2
> >> Logitech Wheel Mouse"" (type: MOUSE)
> >> [ 5571.538] (II) "ImExPS/2 Logitech Wheel Mouse": initialized for
> >> relative axes.
> >> [ 5571.538] (II) config/udev: Adding input device "ImExPS/2 Logitech
> >> Wheel Mouse" (/dev/input/mouse0)
> >> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Applying InputClass
> >> "evdev catchall"
> >> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": always reports core
> >> events
> >> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Device:
> >> "/dev/input/mouse0"
> >> [ 5571.539] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device
> >> [ 5571.539] (II) UnloadModule: "evdev"
> >> [ 5571.539] (EE) PreInit returned NULL for ""ImExPS/2 Logitech Wheel
> >> Mouse""
> >>
> >>
> >> Things work fine here, I just wanted to make sure whether my
> >> configuration is broken or these are known issues.
> >>
> > This is also expected at the moment. The enumeration of devices gives
> > us /dev/input/event* (which evdev can handle), and
> > /dev/input/{mouse,kbd}* (which it doesn't). This results in a couple
> > error messages for the latter. It should be possible to tone those down
> > so users don't get scared…
>
> The other thing hal does here is ignore all devices except those where
> the sysfs path ends in event%d. Maybe we should do the same.
not really, this is a configration issue. by default the server doesn't do
anything unless an input class is specified. In this case, the evdev driver
shouldn't be assigned to the mouse0 device.
In Fedora we ship the evdev catchall rule with MatchDevicePatch
"/dev/input/event*" to avoid the above error messages.
Cheers,
Peter
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 2, 2010 at 9:28 AM, Julien Cristau wrote:
> On Tue, Mar 2, 2010 at 17:57:04 +0100, Simon Thum wrote:
>
>> Hi all,
>>
>> I just upgraded my X to use udev. It generally went fine, but I noticed
>> two minor issues:
>>
>> At first, the device names have quotation marks. As in
>>
>> xinput list
>> ...
>> ⎜ "ImExPS/2 Logitech Wheel Mouse" id=7 [slave pointer
>> (2)]
>>
>> I looked up the source, but found nothing obvious. Is this expected
>> behavior?
>>
> Yes, the device names we get from udev have quotations marks. I don't
> know why, and I didn't feel like mangling them.
It seems to come straight from the kernel (drivers/input/input.c),
which adds the quotes around the value in NAME to the uevent. Maybe we
should ask them not to do that. Looks like this was changed back in
2005, commit a7fadbe10ccf430e7a8add8b45c561d864087343.
However, I think we should try to be compatible with info.product from
hal. This would require some amount of mangling. What hal does is:
1. Get the name directly from the sysfs name file (no quotes there)
2. Strip trailing whitespace
>> Second, udev seems to add my mouse (and others) twice, resulting in some
>> log spam. Excerpt:
>>
>> [ 5571.537] (II) config/udev: Adding input device "ImExPS/2 Logitech
>> Wheel Mouse" (/dev/input/event3)
>> (...)
>> [ 5571.538] (II) XINPUT: Adding extended input device ""ImExPS/2
>> Logitech Wheel Mouse"" (type: MOUSE)
>> [ 5571.538] (II) "ImExPS/2 Logitech Wheel Mouse": initialized for
>> relative axes.
>> [ 5571.538] (II) config/udev: Adding input device "ImExPS/2 Logitech
>> Wheel Mouse" (/dev/input/mouse0)
>> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Applying InputClass
>> "evdev catchall"
>> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": always reports core
>> events
>> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Device:
>> "/dev/input/mouse0"
>> [ 5571.539] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device
>> [ 5571.539] (II) UnloadModule: "evdev"
>> [ 5571.539] (EE) PreInit returned NULL for ""ImExPS/2 Logitech Wheel
>> Mouse""
>>
>>
>> Things work fine here, I just wanted to make sure whether my
>> configuration is broken or these are known issues.
>>
> This is also expected at the moment. The enumeration of devices gives
> us /dev/input/event* (which evdev can handle), and
> /dev/input/{mouse,kbd}* (which it doesn't). This results in a couple
> error messages for the latter. It should be possible to tone those down
> so users don't get scared…
The other thing hal does here is ignore all devices except those where
the sysfs path ends in event%d. Maybe we should do the same.
--
Dan
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
Re: cosmetic udev glitches
On Tue, Mar 2, 2010 at 17:57:04 +0100, Simon Thum wrote:
> Hi all,
>
> I just upgraded my X to use udev. It generally went fine, but I noticed
> two minor issues:
>
> At first, the device names have quotation marks. As in
>
> xinput list
> ...
> ⎜ "ImExPS/2 Logitech Wheel Mouse" id=7[slave pointer
> (2)]
>
> I looked up the source, but found nothing obvious. Is this expected
> behavior?
>
Yes, the device names we get from udev have quotations marks. I don't
know why, and I didn't feel like mangling them.
> Second, udev seems to add my mouse (and others) twice, resulting in some
> log spam. Excerpt:
>
> [ 5571.537] (II) config/udev: Adding input device "ImExPS/2 Logitech
> Wheel Mouse" (/dev/input/event3)
> (...)
> [ 5571.538] (II) XINPUT: Adding extended input device ""ImExPS/2
> Logitech Wheel Mouse"" (type: MOUSE)
> [ 5571.538] (II) "ImExPS/2 Logitech Wheel Mouse": initialized for
> relative axes.
> [ 5571.538] (II) config/udev: Adding input device "ImExPS/2 Logitech
> Wheel Mouse" (/dev/input/mouse0)
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Applying InputClass
> "evdev catchall"
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": always reports core
> events
> [ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Device:
> "/dev/input/mouse0"
> [ 5571.539] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device
> [ 5571.539] (II) UnloadModule: "evdev"
> [ 5571.539] (EE) PreInit returned NULL for ""ImExPS/2 Logitech Wheel
> Mouse""
>
>
> Things work fine here, I just wanted to make sure whether my
> configuration is broken or these are known issues.
>
This is also expected at the moment. The enumeration of devices gives
us /dev/input/event* (which evdev can handle), and
/dev/input/{mouse,kbd}* (which it doesn't). This results in a couple
error messages for the latter. It should be possible to tone those down
so users don't get scared…
Cheers,
Julien
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
cosmetic udev glitches
Hi all,
I just upgraded my X to use udev. It generally went fine, but I noticed
two minor issues:
At first, the device names have quotation marks. As in
xinput list
...
⎜ "ImExPS/2 Logitech Wheel Mouse" id=7[slave pointer
(2)]
I looked up the source, but found nothing obvious. Is this expected
behavior?
Second, udev seems to add my mouse (and others) twice, resulting in some
log spam. Excerpt:
[ 5571.537] (II) config/udev: Adding input device "ImExPS/2 Logitech
Wheel Mouse" (/dev/input/event3)
(...)
[ 5571.538] (II) XINPUT: Adding extended input device ""ImExPS/2
Logitech Wheel Mouse"" (type: MOUSE)
[ 5571.538] (II) "ImExPS/2 Logitech Wheel Mouse": initialized for
relative axes.
[ 5571.538] (II) config/udev: Adding input device "ImExPS/2 Logitech
Wheel Mouse" (/dev/input/mouse0)
[ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Applying InputClass
"evdev catchall"
[ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": always reports core
events
[ 5571.538] (**) "ImExPS/2 Logitech Wheel Mouse": Device:
"/dev/input/mouse0"
[ 5571.539] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device
[ 5571.539] (II) UnloadModule: "evdev"
[ 5571.539] (EE) PreInit returned NULL for ""ImExPS/2 Logitech Wheel
Mouse""
Things work fine here, I just wanted to make sure whether my
configuration is broken or these are known issues.
In case it helps, I did:
udevadm info --attribute-walk --name /dev/input/mouse0
Udevadm info starts with the device specified by the devpath and then
walks up the chain of parent devices. It prints for every device
found, all possible attributes in the udev rules key format.
A rule to match, can be composed by the attributes of the device
and the attributes from one single parent device.
looking at device '/devices/platform/i8042/serio1/input/input3/mouse0':
KERNEL=="mouse0"
SUBSYSTEM=="input"
DRIVER==""
looking at parent device '/devices/platform/i8042/serio1/input/input3':
KERNELS=="input3"
SUBSYSTEMS=="input"
DRIVERS==""
ATTRS{name}=="ImExPS/2 Logitech Wheel Mouse"
ATTRS{phys}=="isa0060/serio1/input0"
ATTRS{uniq}==""
___
xorg mailing list
xorg@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/xorg
